Adept

The Glass Transition Problem

Adept Armor — Mon, 23 Feb 2026 11:32:55 +0000

The Glass Transition Problem

An Unsolved Mystery in Physics That Could Transform Armor Design

The components of your body armor plate – from the ceramic strike face, to the adhesive layers, to the UHMWPE backing – contain a secret: We don’t fully understand what they are. Not the chemistry; that’s well-characterized. Not the manufacturing process; that’s been refined over decades. The secret is more fundamental. These materials typically incorporate amorphous phases alongside crystalline ones, and those amorphous phases exist in a state of matter that physics cannot properly explain.

The glass transition – the process by which a liquid becomes an amorphous solid – is listed among physics’ great unsolved problems, alongside questions about dark matter and quantum gravity. Yet unlike those cosmic mysteries, this one has immediate practical consequences. Every time we design transparent armor for a vehicle, optimize a ceramic composite, or evaluate the shelf life of polymer-backed plates, we’re working with materials whose fundamental nature remains mysterious.

What Is a Glass, Really?

First, let’s dispel a persistent myth: Glass is not a slow-moving liquid. You may have heard that medieval cathedral windows are thicker at the bottom because the glass has flowed downward over centuries. This is false. Those windows are thicker at the bottom because medieval glassmakers couldn’t produce perfectly uniform panes, and builders sensibly installed them heavy-side-down. While we don’t fully know how glass behaves over geological timescales, calculations indicate that at room temperature, the time required for observable flow far exceeds the age of the universe.

Glass is a solid – but a peculiar kind. When you cool most liquids slowly, they crystallize: Atoms arrange themselves into neat, repeating lattices. But cool certain liquids quickly enough, and they skip crystallization entirely. The atoms never find time to organize. They freeze in place in a disordered configuration, forming what we call a glass.

This isn’t limited to the silicate glass in windows. Polymers like the polyethylene in your UHMWPE backer are semi-crystalline: Perhaps 80-90% of the material forms ordered crystalline lamellae, but the rest – the amorphous regions between those lamellae – is amorphous and viscoelastic. The epoxy or polyurethane adhesives bonding your armor’s layers together? Those are entirely glassy. Even some high-performance ceramics contain glassy phases at their grain boundaries, depending on processing conditions. And the Amorphoid in Adept Armor’s Thunder plate is almost entirely amorphous, hence the name – it’s a very lightly recrystallized glass-ceramic, with small crystalline particles in an amorphous matrix.

The temperature at which a liquid becomes a glass is called T_g, the glass transition temperature. And here’s where things get strange: T_g isn’t a fixed property of the material. It depends on how fast you cool the liquid. Cool it faster, and the transition happens at a higher temperature. Cool it slower, and T_g drops. This is profoundly odd. The melting point of ice doesn’t change based on how quickly you freeze water. But the glass transition is different – it’s not a true phase transition at all, at least not in any conventional sense.

The Missing Framework

To understand why this matters for armor engineering, we need to contrast how we model crystals versus glasses.

For crystalline solids, physicists have developed a beautiful and powerful theoretical framework, broadly under the umbrella of Density Functional Theory. First, you start with the crystal structure, with the material’s atoms arranged in a repeating lattice. From that structure, you can derive the vibrational modes (phonons). From the phonons, you can calculate thermodynamic properties, thermal conductivity, and heat capacity. From the interatomic potentials and lattice geometry, you can compute elastic moduli, almost always to within 10%. Defects – dislocations, vacancies, grain boundaries – are well-defined objects whose behavior can be predicted. The theory is predictive: Given the atomic constituents and their interactions, you can calculate macroscopic properties without having to measure them first.

For glasses, there is absolutely no equivalent framework. This isn’t because no one has tried. Generations of physicists have worked on the problem. The difficulty is fundamental: Glasses are disordered and out of equilibrium and their properties depend on their history. The standard toolkit – equilibrium statistical mechanics, symmetry analysis, perturbation theory around ordered states – doesn’t straightforwardly apply.

Consider what we cannot currently do:

We cannot predict T_g from molecular structure. For a new polymer or metallic alloy, you have to synthesize it and measure the glass transition temperature. There’s no reliable equation that takes atomic characteristics, molecular weight, chain stiffness, and interaction strengths as inputs and outputs T_g.

We cannot predict glass-forming ability. Why does one alloy composition readily form a bulk metallic glass while a nearby composition crystallizes? Why do some polymers vitrify easily while others don’t? Semi-empirical rules exist, but they’re unreliable guides for new systems.

We cannot predict how processing affects properties from first principles. Two glass samples with identical composition but different thermal histories will have different densities, different elastic moduli, different fracture toughness. We can measure these differences. We can fit empirical models to the data. But we cannot start from the molecular structure and the processing parameters and calculate what will come out.

We cannot define “defects” in glasses the way we do in crystals. In a crystal, a dislocation is a topologically distinct object. In a glass, what’s a defect versus just… structure? Ramsey Theory teaches that there’ll be islands of order, which in this case can approximate defects, in any sufficiently large chaotic structure. Machine learning has recently identified “soft spots” in amorphous materials – regions likely to rearrange – but we don’t understand what makes them soft in terms of fundamental physics.

The models we do have are descriptive, not predictive. The Tool-Narayanaswamy-Moynihan model for physical aging uses a “fictive temperature” concept and can fit experimental data quite well. The Williams-Landel-Ferry and Vogel-Fulcher-Tammann equations describe how viscosity changes near the glass transition. These are useful engineering tools. But the parameters must be measured for each material; they don’t emerge from first principles. And extrapolating these models beyond their fitting range is unreliable – which matters when you’re trying to predict 20-year shelf life from accelerated testing.

Why the Fundamental Physics Matters

You might reasonably ask: Can’t we just build models using assumptions? Engineers don’t wait for philosophers to settle debates before building bridges.

The answer is that the unsolved physics has at least two possible solutions, and we don’t yet know which is correct.

Here’s the crux of it: In 1948, chemist Walter Kauzmann noticed that if you extrapolate the entropy of a supercooled liquid to lower temperatures, it would eventually drop below the entropy of the corresponding crystal – an apparent impossibility. This “Kauzmann paradox” suggests that something dramatic must happen before that point is reached. Either the liquid crystallizes, or there’s a true thermodynamic phase transition – an “ideal glass transition” – that we never observe because kinetic arrest always intervenes first.

This has spawned two competing families of theories:

Random First-Order Transition (RFOT) theories hold that there is a true thermodynamic transition underlying the glass transition. In this picture, the glass is a “mosaic” of different local configurations, and the key physics involves configurational entropy and the growing size of cooperatively rearranging regions. The mathematical framework involves replica theory borrowed from spin glass physics.

Dynamical facilitation theories hold that the glass transition is purely kinetic – there’s no hidden thermodynamic transition, but physical dynamics in the material become extraordinarily slow. In this picture, mobility is the key variable: Relaxation in one region enables (“facilitates”) relaxation in neighboring regions, creating complex spatiotemporal patterns. The mathematical framework involves kinetically constrained models and dynamic phase transitions.

These competing theories suggest different mathematical structures for building predictive models. RFOT points toward free energy landscapes, configurational entropy as the control variable, and mosaic length scales. Facilitation points toward dynamic heterogeneity, kinetic constraints, and mobility fields. Betting on the wrong framework means building on a foundation that will eventually crack.

The situation is analogous to trying to build electromagnetic devices in the early 1800s, when some physicists thought electricity and magnetism were separate phenomena and others suspected a deeper connection. You could build motors and generators either way; empirically, by trial and error. When Maxwell published a unified framework, he satisfied scientific curiosity, to be sure – but his discovery also enabled radio, radar, and wireless communications, whereas the wrong framework would have been a dead end.

For glasses, we’re still waiting for our Maxwell. We have competing frameworks, each with successes and failures, and no consensus on which is correct – or whether the true theory combines elements of both. Until this is resolved, our ability to build predictive (as opposed to descriptive) models for amorphous materials will remain limited.

The Strange Physics: Four Surprises

If this were merely a gap in theoretical elegance, it would matter little to engineers. But the physics of glasses produces genuinely strange effects that matter for real materials.

Dynamic Heterogeneity: Your Armor Is a Frozen Patchwork

At any given moment in a glass or supercooled liquid, some microscopic regions are relaxing – rearranging their configurations – while others remain frozen. Different regions of the same material, micrometers apart, can have relaxation rates differing by orders of magnitude. This “dynamic heterogeneity” has been directly imaged in colloidal glasses and confirmed in molecular simulations.

What this means: The amorphous regions in your armor aren’t uniform. There’s a hidden patchwork of mobility. How this affects response to ballistic impact – where strain rates are extreme and stress concentrations intense – is not well understood.

There does appear to be an amorphous-to-stishovite transition in certain silicate glasses under certain types of ballistic impact, but the nature of this phenomenon is still exceedingly mysterious, and most experts find it surprising when they first encounter it.

Physical Aging: The Plate Is Different Tomorrow

Glasses are never truly in equilibrium. They relax toward equilibrium forever, changing density, stiffness, brittleness, and permeability along the way. The armor plate in a warehouse today is measurably different from what it will be in five years.

NIST researchers studying UHMWPE fibers have documented complex aging involving both oxidative degradation and physical relaxation. The physics is entangled: crystalline and amorphous regions age differently, and they interact. Our shelf-life predictions carry uncertainties we cannot eliminate without better fundamental understanding.

The Gardner Transition: A Phase Transition Inside the Glass

Recent work has identified the Gardner transition – a phase transition within the glassy state. Below this transition, the energy landscape fractures from a single basin into a hierarchy of sub-basins. The glass enters a “marginally stable” state where small perturbations trigger avalanches of rearrangements.

The implications for armor under impact – where the material experiences extreme, rapid loading – are potentially significant but almost unexplored. We don’t even know if the Gardner transition occurs in amorphous solid materials under conditions relevant to armor.

Machine Learning Reveals Hidden Order

For decades, physicists sought a structural “order parameter” for glasses – some quantity that would predict dynamics the way crystalline order predicts behavior in solids. They mostly failed; a glass looks structurally almost identical to the liquid it came from.

Then an AI system, poring over vast amounts of data, found something. Researchers trained algorithms on simulations and discovered that local structure does predict which particles rearrange – just not in ways humans were able to perceive. The machine-learned quantity, called “softness,” captures subtle patterns invisible to traditional analysis.

This is both hopeful and sobering. Hopeful in that structure-property relationships exist; we just couldn’t see them. Sobering in that the order parameter is high-dimensional and non-intuitive. Understanding glasses may require mathematical frameworks that haven’t been developed yet.

Implications for Armor Engineering

Consider ceramic armor. High-performance ceramics like silicon carbide and boron carbide derive their ballistic effectiveness from complex microstructures. Some ceramics – particularly those processed with oxide sintering aids, and liquid-phase sintered SiC – contain glassy intergranular phases. The properties of these phases depend sensitively on processing: Sintering temperature, cooling rate, atmosphere, trace impurities. Small changes can produce meaningful differences in ballistic performance. Why? We don’t fully know, because we lack a predictive theory for how processing determines the properties of those amorphous regions.

(To be clear: Not all armor ceramics contain glassy phases. Hot-pressed boron carbide and solid-state sintered silicon carbide can have clean grain boundaries. But the general point stands, that where amorphous phases exist, our predictive capability is limited.)

Next, as mentioned previously, consider polymer backing materials. UHMWPE is semi-crystalline. Its ballistic properties emerge from highly aligned crystalline lamellae, but the amorphous regions govern critical behaviors: Impact response, creep under load, aging over time. We can characterize these behaviors empirically. We cannot predict them from the molecular structure and processing history.

Then consider adhesive layers. The epoxies and polyurethanes bonding armor systems are entirely glassy. How will a new adhesive formulation perform over 15 years of storage in variable conditions? We can run accelerated aging tests and fit empirical models. But we’re extrapolating, and the uncertainty in those extrapolations reflects our incomplete understanding of glass physics.

Armor engineers reading this know that adhesive failure is the most unpredictable, most vexing reason plates fail – and one which seems to have contributed to high-profile struggles in the sector, such as those faced by Ceradyne.

Lastly, consider transparent armor – glass by definition – and metallic glasses, which offer extraordinary hardness and elasticity but whose glass-forming ability we cannot reliably predict. In each case, development proceeds by empirical optimization: Make samples, test them, adjust. This process, one not unfamiliar to ancient alchemists, works. Armor gets fielded. But it’s slow, expensive, and definitely leaves performance on the table. In fact, I’ll say outright that metallic glasses are still very rare and exotic simply because there’s no physical theory for them.

Science as Strategic Advantage

Unlike the deep mysteries of the universe, the glass transition problem represents a missing chapter in our understanding of matter – one that directly limits our ability to design and optimize militarily-relevant materials.

This argues for sustained investment in fundamental condensed matter physics. The nation that develops a predictive framework for amorphous materials gains the ability to design what others can only discover by laborious trial and error. In an era where materials often determine the limits of military capability, that’s a decisive advantage.

The mystery of glass is hiding in plain sight – in the armor plates, vehicle windows, and helmet shells we use every day. The field of materials science is waiting for our Maxwell.

The post The Glass Transition Problem appeared first on Adept.

Universal Combat Helmet Typology

Adept Armor — Sat, 03 Jan 2026 10:55:48 +0000

Universal Combat Helmet Typology

A Morphological Classification System

I’ve been working on a book which follows the development of the combat helmet over time.

One of the very first things which stood out – and this has been remarked upon by other historians – is that the helmets of ancient Sumer, as depicted in the Stele of the Vultures and recovered in the archaeological find on the ramp of “The King’s Grave,” bear a striking resemblance to the medieval bascinet. Though separated by 3000 years, the forms are practically identical.

I’ve also noticed, to my amusement, that certain German secret (“secrete”) helms on display at the Germanisches Nationalmuseum in Nuremberg very closely resemble our own Novasteel helmet shell. Apart from differences in the type and thickness of the liner, they look like the same helmet.

Further inspection along such lines has uncovered many other similarities through the ages. This calls for a typological system which makes cross-temporal and cross-cultural comparisons legible – allowing, for instance, the morphological kinship between Sumerian helmets and medieval bascinets to be expressed in a common notation.

Like certain sword typologies, which separate blade form from pommel and guard types, this system treats shell form, face protection, neck defense, and other subsystems as independent variables that combine differently across periods and regions. Ultimately, it can be used to describe all combat helmets – over a period of more than 5000 years, stretching from at least 3000 B.C. to 2025 A.D. and beyond – by morphology and construction rather than by assumed date, place, or cultural origin.

One thing I would note in advance, which will be covered at length in the book but needs some justification here, is that helmets designed primarily to defend against melee weapons and arrows tend to have much more variance in their crown shapes. Pointed, peaked, and combed helmets were common centuries ago, though they seem strange to modern eyes as there are no modern helmets with similar forms. The Styrian Armory at Graz contains hundreds of conoidal and combed helmets, which are, from a modern perspective, much taller and heavier than they “need to be.” Suffice to say that such angled geometries make helmets considerably more effective against downward blows from sabres, maces, and arrows falling from the sky.

The typology is here in PDF format as a preprint: https://zenodo.org/records/17986416
Cite as: Ganor, J. (2025). Proposal for a Universal Combat Helmet Typology. Zenodo. https://doi.org/10.5281/zenodo.17986416

The whole thing is also below. A pictorial representation will be the subject of a future post.

Notation Format

A helmet is described as:

Sx–Ax–Fx–Bx–Nx–Mx-Cx–Lx–Rx

S = Shell crown form (primary shape) · A = Aural cut · F = Face system · B = Brim system · N = Neck system · M = Material · C = Construction · L = Liner/suspension · R = Retention

For unknown or indeterminate fields, use U in that field (SU, CU, etc.).

Where a field is absent, and evidence indicates absence, write 0. (e.g., F0 when a helmet does not incorporate a face protection system.)

S may take suffix modifiers; M/C may take lettered subtypes; F may take subtype plus optional aperture annotation; “?” may suffix any field to mark inference.

S-Series: Shell Crown Form

S encodes the shell’s core cranial form plus optional integral modifiers. Append t to indicate an integral rear tail formed continuously with the skull. Append n to indicate an integral nape flange. If the rear protection is attached as a separate component, do not use t or n; encode it in N-series (e.g., N1/N2/N3/N4). For enclosing shells (S6), append f (flat top), d (domed), or p (peaked) to record roof profile (e.g., S6f).

When more than one S-type seems to fit, use these precedence rules: S6 overrides all. Then S4 if elongation exceeds threshold. Then S3 if the side profile is predominantly conical. Then S5 for tall crowns that are not S3 or S4. Otherwise choose between S1 and S2 by rim drop.

Border cases (rim drop 40-60mm): Choose S2 if rim approaches ear height or wraps the occiput; otherwise S1; append ‘?’ if uncertain.

SU — Unknown / Too Fragmentary

Insufficient surviving material to classify shell form.

S1 — Shallow Cap

Low dome with rim typically at or above the upper-temple region. No integral rear tail; minimal side coverage from the shell itself. Diagnostic: Rim drop <40mm below crown apex at sides.

S2 — Deep Bowl

Dome or bowl with notably deeper drop at sides and/or back than S1. Shell provides “wrapped” cranial coverage with rim sitting low around the head. Diagnostic: Rim drop >60mm below crown apex, approaching or covering the ears.

S3 — Cone

Sides rise in a largely conical profile to an apex. A point, knob, or finial may exist but is not required. Diagnostic: Profile angle from rim to apex <60° from vertical.

S4 — Ovoid / Almond

Shell elongated fore–aft in plan view (egg or almond shape). Often features a medial keel or ridge, though this is not required. Diagnostic: Fore-aft dimension >1.15× lateral dimension at widest point.

S5 — High Crown

Tall shell with pronounced verticality – more “tower-like” than S2 – regardless of whether the apex is rounded or pointed. Diagnostic: Crown height >1.3× maximum shell width.

S6 — Enclosing Shell

Shell conceived as a full head enclosure rather than a cap or bowl: box, cylinder, or close-fitting container with small apertures. The enclosing geometry is primary rather than achieved through added face/neck components. Append f (flat top), d (domed), or p (peaked) to record roof profile. For e.g., a “sugar loaf” greathelm is S6p.

Note: A “kettle-hat” is typically S2 + brim type (B3/B4), not a unique shell class. A “sallet” is often S2t (if the tail is integral) plus appropriate face/visor codes.

A-Series: Aural cut

A encodes the shell’s aural cut: How the helmet rim is shaped around the ear region, independent of face hardware (F-series) and independent of brim geometry (B-series). Use A to describe whether the shell opening leaves the ears exposed, partially exposed, or covered by the shell’s rim line as worn in standard orientation.

A does not describe separate cheekpieces, ear lappets, or masks (encode those in F-series), and it does not describe brim projection that happens to shade the ears (encode in B-series). A describes the cut of the shell itself.

For unknown or indeterminate aural cut, use AU. If the shell has no meaningful ear region due to full enclosure (e.g., greathelms), A should be recorded as A0 (not applicable).

AU — Unknown/Indeterminate

Insufficient evidence to determine how the rim is cut around the ears (fragmentary rim, unclear orientation, missing side profile).

A0 — None / Not Applicable

Aural cut is not meaningfully defined because the shell is conceived as an enclosure rather than an open-rim helmet (typical of S6).

A1 — Ear-Covering Cut

The shell rim drops low enough at the ear station that the ear is covered by the shell opening as worn, or would be covered absent separate face components. Diagnostic: rim line at the ear station at or below the ear’s lower half on a headform (often near or below the tragus level). Common on full-coverage ballistic shells and many deep historical bowls.

A2 — Partial Ear Exposure

The rim line intersects the ear station such that the ear is partially exposed (upper or lower portion exposed) in the shell opening. This is the transitional state between full coverage and high-cut shells. Diagnostic: rim line passes through the ear region—neither clearly above the ear (A3) nor clearly below it (A1). Use ? if uncertain due to fit/orientation.

A3 — High-Cut / Ear-Exposed

The shell rim is cut high around the ear station, leaving the ear largely exposed in the opening as worn. Diagnostic: rim line lies clearly above the ear’s upper half (often above the helix/upper ear line on a headform).

F-Series: Face System

Rigid facial protection elements and their configuration. Cheekpiece subtypes (F2a–F2c) distinguish functionally and constructionally different approaches to lateral face protection.

Compound encoding (e.g., F2c+F3) is possible, see, e.g., the Sutton-Hoo helmet.

F0 — Open

No rigid facial defense beyond rim or brow reinforcement. Face fully exposed.

F1 — Nasal / Midline Bar

A rigid midline nasal bar or equivalent central guard is the primary facial element. May be fixed or adjustable.

F2 — Cheek/Ear Defenses

Rigid cheekpieces, ear covers, or lappets provide meaningful lateral facial protection while the face remains largely open frontally. Subtypes:

F2a — Integral cheekpieces: Cheek guards formed as part of the shell or rigidly fixed to it (e.g. Illyrian).

F2b — Hinged cheekpieces: Cheek guards attached via hinges allowing independent movement (e.g., Chalcidian, Roman cavalry).

F2c — Attached lappets: Separate cheek/ear defenses attached to rim or band, typically of different construction than the shell (e.g., Spangenhelm cheek guards, lamellar lappets).

F3 — Mask / Faceplate

A rigid, non-openable faceplate or mask provides primary frontal closure, with shaped apertures (cruciform, T-slit, ocular holes, breathing slots). Aperture geometry may be noted parenthetically after the subtype (e.g., F3a(+) cruciform, F3a(T) T-slit, F3a(O) ocular, F3b(f) frogmouth or horizontal slot, F3b(I) large opening with vertical bars). Subtypes:

F3a — Integral/permanent mask: Faceplate is formed as part of the shell or fixed in place as a non-removable component.

F3b — Detachable mask: Mask is a removable element (pins, hooks, straps), typically non-hinged; access/ventilation is achieved by removing the mask rather than lifting a visor.

F4 — Movable Visor Only

A hinged or pivoting visor (single principal movable element) provides frontal closure. The visor may lift, pivot laterally, or detach.

F5 — Visor + Lower-Face Closure

Visor combined with bevor, chin-piece, or equivalent lower-face defense. Together these elements close the face when in the “armed” configuration.

F6 — Fully Enclosing, Non-Openable

Face closure is integral and fixed, not an openable visor. Entry/exit is achieved by donning/doffing the entire helmet. (Note: This generally pairs with S6, but can apply even when the shell is not S6, thereby capturing certain masks/closures on otherwise non-enclosing shells.)

F7 — Lower-Face Guard / Mandible

A rigid lower-face element (bevor, chin-guard, mandible) provides meaningful jaw/chin protection while the upper face remains open and there is no movable visor as the primary closure. Subtypes:

F7a — Fixed: Lower-face element is integral to the shell or permanently fixed to it.

F7b — Detachable: Lower-face element is a removable accessory fitted as needed.

B-Series: Brim System

Perimeter projections beyond the vertical projection of the shell wall. A “brim” must project meaningfully (>15mm) to qualify; minor rim flanging is not a brim. For rear-only projections primarily protecting the nape, use N-series rather than B-series.

B0 — None

No brim or peak. Shell terminates at rim without significant horizontal projection.

B1 — Peak Only

Localized projection at brow/front; not circumferential. Provides shade and limited frontal rain/blow deflection.

B2 — Partial Brim

Distinct brim covering a sector (typically front and/or sides) but not full 360°. Rear-only projections that primarily protect the nape should be classified in the N-series (N1/N2). Subtypes:

B2a — Partial Brim, Planar: Brim sector largely in a single plane (within ±15° of horizontal over the brim area).

B2b — Partial Brim, Turned: Brim sector deliberately upturned/downturned (angle >25° from horizontal on ≥50% of the brim area).

B3 — Full Brim, Planar

360° brim largely in a single plane. May be slightly waved but not deliberately turned as a defining trait. Diagnostic: brim angle within ±15° of horizontal around full circumference.

B4 — Full Brim, Turned

360° brim with strong upturn or downturn as defining geometry. Diagnostic: brim angle >25° from horizontal on at least 50% of circumference.

B5 — Visor-Brim Hybrid

The “brim” function is achieved primarily by a projecting visor or peak integrated with the face system rather than a true circumferential brim.

N-Series: Neck System

Neck and nape protection elements. This series encodes protection that is attached to the helmet as a distinct component, versus integral shell geometry (which belongs in S-series).

N0 — None / Minimal Nape Drop

No distinct neck defense beyond the shell rim itself.

N1 — Attached Rigid Nape Flange

Small rigid rear flange projecting from shell rim (including rear-only brim-like projections). Non-integral – clearly a separate attached flange. For integral nape flanges, use the “n” modifier to S.

N2 — Attached Rigid Tail

Long rigid rear tail providing major nape/upper-neck coverage, attached as a separate rigid component (riveted plate, bolted extension) rather than formed as part of the shell geometry. Distinguished from S6 by construction: S6 tails are continuous with the shell; N2 tails are visibly joined.

N3 — Attached Flexible Curtain

Mail, leather, lamellar, scale, or textile defense hanging from rim or attachment band. Archaeological evidence includes regularly distributed rivet holes around helmet rims consistent with attaching organic or mail components.

N4 — Articulated Neck Defense

Lames or plates articulated to allow movement, attached to the helmet as a unit. Distinct from both a solid shell tail (t-modified S or N2) and a flexible curtain (N3).

N5 — Integrated Gorget Interface

Helmet designed to mate mechanically with a gorget or upper collar defense. The boundary between “helmet” and “gorget” is intentionally engineered as a system interface (flanges, slots, overlapping geometry).

M-Series: Material

The primary structural material of the protective shell. This series enables tracking technological evolution independently of form.

MU — Unknown

Material cannot be determined from available evidence.

M1 — Copper Alloy (Bronze, Brass)

Shell formed from copper-based alloys. Includes arsenical bronze, tin bronze, and brass.

M2 — Ferrous (Iron, Steel)

Shell formed from iron or steel. Includes wrought iron, low-carbon steel, and hardened/tempered steel.

M3 — Organic (Leather, Rawhide, Wicker)

Shell formed primarily from organic materials. Includes boiled leather (cuir bouilli), rawhide, lacquered leather, and woven plant fiber constructions.

M4 — Composite Traditional

Shell combining multiple traditional materials as structural and protective elements: metal or stone plates on leather substrate, horn and metal laminations, the boar-tusk helmets of Mycenae and the Iliad, etc.

M5 — Ballistic Composite

Modern fiber-reinforced composite shell: aramid (Kevlar), UHMWPE (Dyneema/Spectra), E-glass, or similar ballistic textiles in resin matrix.

M5a — Ceramic surface: Ballistic composite with ceramic strike face (alumina, silicon carbide, boron carbide) affixed to fiber backing.

M6 — Bulk Polymeric

Bulk polymers such as polycarbonate, PEEK, PEI, or similar, with or without short fiber or particle reinforcement.

M7 — Exotic Structural Metals

Titanium, aluminum, nickel, magnesium, zirconium, molybdenum, and alloys of those metals. High-entropy metals. Amorphous metals.

M8 — Precious Metals

Gold, silver, electrum, and other typically decorative or ceremonial precious metals.

C-Series: Construction

How the shell is made—fabrication method independent of material (M-series) or form (S-series).

CU — Unknown

Construction method cannot be determined.

C1 — One-Piece Monocoque Shell

Raised, forged, spun, or pressed as one rigid shell from sheet or billet. Includes traditional bronze raising and medieval steel skull forging.

C2 — Molded Shell

Shell formed in a negative mold with fluid/semi-fluid state or deposition.

C2a — Cast metal shell: Shell constructed via metal casting. Cast metallic shells may be finished by hammering, chasing, grinding, or polishing, but the basic form derives from a mold.

C2b — Molded polymer shell: Shell constructed via polymer casting, injection molding, thermoforming, vacuum forming, and other polymer forming methods.

C3 — Segmented Shell with Bands (Spangenhelm)

Plates riveted under or within crossing bands or a framework. The bands are structural, defining the geometry and holding plates in position.

C4 — Multi-Plate Riveted Shell

Large plates riveted edge-to-edge or overlapped to form the shell without a dominating band framework. Includes ridge-helm and “secret” constructions.

C5 — Scale/Lamellar on Substrate

Many small elements (scales, lamellar plates, rings) attached to or laced over an organic substrate that provides the shell form.

C6 — Pressed and Molded Composite Shell

Fiber/resin composite shell formed as a structural ballistic element: Aramid or UHMWPE fabric layers consolidated under heat and pressure. The composite layup is the shell structure.

C7 — Two-Shell System

Outer protective shell plus a separate rigid inner liner shell that carries suspension and fit. The U.S. “M1” helmet family exemplifies this: Steel outer shell, hard-hat-like inner liner with adjustable webbing. For C7 helmets, S/M/F/B/N/C describe the primary outer shell; L and R describe the helmet system as worn (often implemented on the inner liner shell).

L-Series: Liner / Suspension

How the head is supported inside the shell. Critical for comfort, stability, and impact protection.

LU — Unknown

No liner evidence survives or can be inferred.

L0 — None Present

No liner evidently used; the shell is designed to rest directly against the wearer’s head. Use L0 only when there is positive evidence of direct-contact intent (specification, preserved context, unambiguous sizing cues, or an interior edge treatment clearly meant for contact); otherwise use LU or L2

L1 — Attached Padding

Padding fixed to the helmet interior by stitching, rivets, or adhesive. Archaeological evidence includes rivet-hole patterns around rims consistent with attaching organic padding. The padding directly contacts the head without a suspension gap.

L2 — Separate Cap/Coif Worn Under

A separate organic cap, arming cap, or padded coif worn beneath the helmet provides fit and cushioning. Not fixed to the shell. Common where surviving helmets lack interior attachment points.

L3 — Internal Leather Band System

Internal leathers riveted to shell; linings sewn to those leathers and replaceable. This system allows liner replacement using the same attachment holes. Well-documented in late medieval armour construction.

L4 — Webbing/Basket Suspension

Webbing harness attached at multiple points forming a “basket” that suspends the head with a standoff gap from the shell. PASGT-style systems maintain approximately 12–13mm spacing for ventilation, impact deformation clearance, and blunt trauma protection.

L5 — Pad-Based Energy-Absorbing Liner

Foam or energy-absorbing pads arranged inside the shell, typically attached via Velcro or similar. Pads attenuate blunt impact and provide standoff. Standard in ACH and subsequent helmet generations.

L6 — Rigid Liner Shell with Suspension

A distinct rigid inner liner carries suspension geometry and fit bands. The liner provides uniform spacing between the outer shell and the head, distributing force via headband and nape band. Pairs with C7 (two-shell system).

R-Series: Retention

How the helmet is secured to the head. Retention evidence rarely survives in archaeological contexts, but attachment points often indicate system type. Note: If there is evidence of sewing points, lacing, or fixture geometry for integration, use R6; if there is evidence pointing to no retention and no integration mechanism, use R0; if neither can be determined, use RU.

RU — Unknown

No retention evidence survives; cannot determine whether a system existed.

R0 — Absent

Positive evidence indicates that no retention system was used, and it is not evident how the helmet is designed to be integrated into other headgear.

R1 — Simple Chin Tie/Thong

Two attachment points with a simple tie (cord, leather thong). Common where organic straps have not survived; inferred from paired attachment holes.

R2 — Chin Strap, Two-Point Adjustable

Two-point strap with buckle, hook, or other adjustment hardware. Standard through most historical periods where hardware survives.

R3 — Three-Point Harness

Harness with three independent attachment points (e.g., two lateral + one rear/nape, or a Y-type split). Seen in some industrial/work-at-height helmets and certain retrofit strap kits; uncommon in most archaeological contexts.

R4 — Four-Point Harness

Four-point retention creating a stable helmet-head interface. Standard in modern combat helmets (ACH, ECH, IHPS).

R5 — Boltless Retention

Retention system attached to helmet shell via Velcro or adhesive resin without slots or perforations in the helmet shell, uncommon.

R6 — Integrated into Secondary Headgear

No traditional harness. Sewn or laced into hat/hood.

Optional Fields

There are morphological variables that are of lower importance but can be described in the typology on an as-needed basis. Omit when not needed or not recorded.

Edge treatment and reinforcement can be described in an optional E series.

E0 none

E1 rolled/wired rim

E2 applied brow band

E3 reinforcing ribs/combs

This would allow the expression of crests or certain ridge-helm features without contorting S or C, but is rare and niche enough that it should remain optional, and used only when called for.

Practical Identification Key

Work through the following sequence to classify an unknown helmet. Some steps may require physical examination or detailed photography.
Shell form (S): Determine overall cranial geometry from silhouette and planform. Shallow cap (S1), deep bowl (S2), cone (S3), ovoid/almond (S4), high crown (S5), or enclosing shell (S6)? Apply precedence rules where multiple S-types seem to fit.
Aural cut (A): Determine how the shell rim is cut around the ears, independent of cheekpieces/lappets (F-series) and independent of brim projection (B-series). Classify in standard orientation (brow forward, seated naturally). Ear-covering cut (A1), partial ear exposure (A2), high-cut / ear-exposed (A3), unknown (AU). If the helmet is an enclosing shell where an “ear opening” is not meaningfully defined, record A0 (not applicable).
Face system (F): Identify rigid facial defense configuration. Movable visor only (F4) or visor plus lower-face closure (F5)? Fixed mask/faceplate (F3a) or detachable mask (F3b), with aperture annotation as applicable (e.g., (T), (+), (O), (f))? Fully enclosing, non-openable facial closure (F6)? Lower-face guard only (F7a/b)? Nasal bar (F1)? Cheekpieces/lappets (F2a/b/c)? Or fully open (F0)? Use compound encoding (e.g., F2c+F3) when two independent rigid subsystems are present.
Brim system (B): Determine perimeter projection beyond the vertical projection of the shell wall. No brim (B0), peak only (B1), partial brim planar (B2a) or turned (B2b), full brim planar (B3) or turned (B4), or visor-brim hybrid (B5). Treat minor rim flanging as B0 unless projection exceeds the brim threshold.
Neck system (N): Determine distinct nape/neck protection elements that are attached as separate components rather than integral to the shell. None/minimal (N0), attached rigid nape flange (N1), attached rigid tail (N2), attached flexible curtain (N3), articulated neck defense (N4), or integrated gorget interface (N5). If rear protection is integral to the shell, encode it in S using the appropriate integral modifier rather than in N.
Material (M): Identify primary shell material. Copper alloy (M1), ferrous (M2), organic (M3), traditional composite (M4), ballistic composite (M5), ceramic-faced ballistic composite (M5a), bulk polymeric (M6), exotic structural metals (M7), precious metals (M8), or unknown (MU).
Construction (C): Examine seams, rivets, bands, and surface evidence to determine fabrication method. One-piece monocoque (C1), molded (C2, with C2a cast metal or C2b molded polymer), segmented with bands / spangenhelm (C3), multi-plate riveted shell (C4), scale/lamellar on substrate (C5), pressed/molded composite shell (C6), or two-shell system (C7). Use CU if construction cannot be determined.
Liner / suspension (L): Examine interior evidence for how the head is supported. None present with positive evidence (L0), attached padding (L1), separate cap/coif worn under (L2), internal leather band system (L3), webbing/basket suspension (L4), pad-based energy-absorbing liner (L5), or rigid liner shell with suspension (L6). Use LU if liner cannot be determined.
Retention (R): Examine attachment points, slots, and fixtures for how the helmet is secured or integrated. Simple chin tie/thong (R1), two-point adjustable chin strap (R2), three-point harness (R3), four-point harness (R4), boltless retention (R5), integrated into secondary headgear (R6), absent with positive evidence (R0), or unknown (RU). Mark inferred identifications with “?” (e.g., R2?) rather than treating inference as direct evidence.

Appendix: Proposed Metrics

To make boundaries between types reproducible rather than judgment calls, the following measurements are proposed. All measurements should be taken from the helmet in standard orientation (brow forward, shell resting naturally). Thresholds are provisional and should be normalized to shell size (dimensionless ratios) where possible; absolute millimeter cutoffs are here utilized as heuristics.

Shell Form Metrics

Rim drop: Vertical distance from crown apex to rim at the lateral midpoint (ear position). S1 <40mm, S2 >60mm.
Profile angle: Angle of shell wall from vertical, measured at lateral midpoint. S3 (cone) <60° from vertical.
Elongation ratio: Fore–aft dimension ÷ lateral dimension at widest. S4 (ovoid) >1.15.
Crown height ratio: Crown height ÷ maximum shell width. S5 (high crown) >1.3.
Tail length: For t-modified S, measured from rear rim of “bowl” portion to tail terminus. Minimum 50mm to qualify for the t suffix; otherwise classify as S2n.

Brim Metrics

Brim projection: Horizontal distance from shell wall to brim edge. Minimum 15mm to qualify as brim (vs. rim flange).
Brim angle: Angle from horizontal plane. Planar types (B2a/B3) within ±15°; turned types (B2b/B4) >25° on ≥50% of brim area/circumference.

Liner Metrics

Attachment point count: Number and distribution of rim rivet holes. Regular distribution (4+ points, roughly symmetric) suggests L1 attached padding.
Standoff gap: For suspension systems (L4, L5, L6), measured distance between shell interior and head contact surface. Modern ballistic helmets typically 10–15mm.

Design Principles

This typology follows several principles that distinguish it from period- or culture-bound classification systems:

Morphology over origin: Classification is based on observable physical form and construction, not assumed date, place, or cultural affiliation. A Sumerian helmet and a medieval sallet can share a code if they share a form.
Independent subsystems: Shell form, face protection, neck defense, construction method, liner, and retention are treated as independent variables. This reflects reality: the same shell form can appear with different face systems across periods.
Archaeological compatibility: The system handles incomplete evidence. Unknown fields are marked “U” rather than forcing assumptions. The distinction between L1 (attached padding) and L2 (separate cap) is explicitly supported by archaeological analysis of rivet-hole patterns. Some fields may be inferred; inference is marked with a question mark – e.g. L3? or R2? – and never treated as evidentially equivalent to direct observation.
Extensibility: New types can be added to any series without breaking existing classifications. If a shell form emerges that doesn’t fit S1-S6 – say, for instance, a diving-bell-style combat helmet – add S7. The notation remains consistent.
Cross-temporal legibility: The system makes convergent evolution visible. A WWI Brodie and a medieval kettle hat both classify as S2–F0–B3, which makes clear the morphological kinship. Note that shared codes imply form similarity only; they do not assert cultural transmission.

Worked Examples

Ancient World (3000 BC – 500 AD)

1. Royal Cemetery of Ur Gold Helmet (c. 2500 BC)

Deep bowl form covering ears and nape, with integral tail-swept rear extension. Beaten gold sheet. Open face. No brim. One-piece construction. Liner and retention unknown (precious metal, burial context).

S2t–A1–F0–B0–N0–M8–C1–LU–RU

2. Early European Bronze Cap Helmet (Bronze Age)

Shallow cap or conical profile. Copper alloy, raised. Open face. No brim. Rivet holes around rim indicate attached padding. Paired holes suggest simple chin tie. Ears exposed by shell cut.

S1–A3–F0–B0–N0–M1–C1–L1–R1

Note: Use S3 if profile is distinctly conical.

3. Corinthian Helmet (c. 700–300 BC)

Deep bowl with integral, non-openable faceplate featuring eye and mouth apertures. Bronze. Ears fully covered by shell. No brim. Nape coverage integral. One-piece raised. Worn over padded cap.

S2–A1–F3a(O)–B0–N0–M1–C1–L2–R1

4. Illyrian Helmet (7th–4th c. BC)

Deep bowl with integral cheekpieces formed as part of the shell. Bronze. Ears covered. Open face above cheekpieces. No brim. One-piece raised.

S2–A1–F2a–B0–N0–M1–C1–L2–R1

5. Chalcidian Helmet (5th–4th c. BC)

Deep bowl with hinged cheekpieces allowing independent movement. Bronze. Ears covered by shell rim. Often has a nasal guard. No brim. One-piece raised.

S2–A1–F2b–B0–N0–M1–C1–L2–R1

Note: Add +F1 if nasal bar present.

6. Boar-Tusk Helmet (Mycenaean, c. 1600–1100 BC)

Cap form constructed from boar tusks sewn or laced to an organic substrate. Open face. Ears likely exposed. No brim. Composite traditional construction.

S1–A3–F0–B0–N0–M4–C5–L2–R1

7. Late Roman Ridge Helmet (4th–6th c. AD)

Deep bowl with medial ridge. Iron. Hinged cheekpieces. Ears covered. No brim. Mail curtain at rear. Multi-plate riveted construction along ridge. Leather band liner. Chin strap.

S2–A1–F2b–B0–N3–M2–C4–L3–R2

8. Roman Imperial Legionary Helmet (1st–3rd c. AD)

Deep bowl. Iron or bronze. Hinged cheekpieces. Ear-covering cut. No brim. Attached rear nape flange. Multi-plate or one-piece construction varies by type. Internal leathers. Chin strap.

S2–A1–F2b–B0–N1–M2–C4–L3–R2

Note: Some examples are C1 with applied details.

9. Roman Cavalry Parade Helmet with Detachable Mask (1st–2nd c. AD)

Iron helmet fitted with a removable full-face mask for parade or sport. Mask attaches by hooks/pins rather than hinged visor. Ears covered. No brim. Multi-plate construction. Leather band liner. Chin strap.

S2–A1–F3b(O)–B0–N0–M2–C4–L3–R2

10. Assyrian Conical Helmet (9th–7th c. BC)

Conical shell rising to a pointed apex. Iron or bronze. Open face. Ears exposed. No brim. No neck defense (worn over scale or lamellar). One-piece raised or multi-plate. Padded cap worn under.

S3–A3–F0–B0–N0–M2–C1–L2–R1

Note: Depicted extensively in Assyrian palace reliefs.

11. Phrygian-Style Helmet (4th–2nd c. BC)

Distinctive forward-curving apex creating fore-aft elongation in profile. Bronze. Open face or with hinged cheekpieces. Ears typically covered. No brim. One-piece raised.

S4–A1–F0–B0–N0–M1–C1–L2–R1

Note: The forward-leaning peak creates the ovoid signature. Add F2b if cheekpieces present.

12. Thracian Helmet (5th–3rd c. BC)

Tall crown with pronounced peak, often forward-swept. Bronze. Integral cheekpieces or open. Ears covered. No brim. One-piece raised. Worn over cap.

S5–A1–F2a–B0–N0–M1–C1–L2–R1

Note: Some examples are distinctly ovoid (S4) rather than just tall (S5).

Migration Period & Early Medieval (500–1100 AD)

13. Migration Period Spangenhelm (5th–7th c. AD)

Deep bowl or slightly conical. Iron bands with bronze/iron plates. Attached cheek lappets of different construction than shell. Partial ear exposure at rim. No brim. Mail curtain common. Banded construction. Internal leathers probable.

S2–A2–F2c–B0–N3–M2–C3–L3–R2

Note: Many variants add nasal (F1) or change neck defense (N0/N3).

14. Sutton Hoo Helmet (7th c. AD)

Deep bowl. Iron with decorated bronze/silver panels. Attached lappets plus fixed faceplate with eyeholes and nose/mouth guard. Partial ear exposure. No brim. Mail curtain at rear. Multi-plate construction. Internal leathers.

S2–A2–F2c+F3a–B0–N3–M2–C4–L3–R2

Note: Compound face encoding captures both lappets and faceplate.

15. Vendel/Valsgärde Helmets (6th–8th c. AD)

Deep bowl, often with pronounced crest/comb. Iron. Spectacle-type eye guards or attached face elements. Partial ear exposure. No brim. Mail curtain. Multi-plate riveted. Internal leathers.

S2–A2–F2c+F3a–B0–N3–M2–C4–L3–R2

16. Norman Nasal Helm (11th c.)

Conical shell. Steel. Fixed nasal bar as primary face defense. Ears exposed by shell cut. No brim. No integral neck defense (worn over mail coif). One-piece forged. Arming cap worn under.

S3–A3–F1–B0–N0–M2–C1–L2–R1

Note: If helmet carries attached mail, add N3.

17. Cervellière / Skullcap (12th–14th c.)

Shallow cap or borderline deep bowl. Steel. Open face. Ears exposed. No brim. No neck defense. One-piece forged. Worn over mail coif or padded cap.

S1–A3–F0–B0–N0–M2–C1–L2–R1

Note: Deep skullcaps drift into S2.

High & Late Medieval (1100–1500 AD)

18. Great Helm, Flat-Topped (13th c.)

Enclosing cylindrical shell with flat top. Steel. Fixed faceplate with cruciform aperture. Aural cut not applicable (full enclosure). No brim. No attached neck defense (worn over mail). Multi-plate riveted. Internal leather bands. Simple chin tie.

S6f–A0–F3a(+)–B0–N0–M2–C4–L3–R1

19. Great Helm, Sugar Loaf (13th–14th c.)

Enclosing shell with peaked/conical top. Steel. Fixed faceplate with cruciform or T-slit aperture. Full enclosure. No brim. Multi-plate riveted. Internal leathers.

S6p–A0–F3a(+)–B0–N0–M2–C4–L3–R1

Note: Use F3a(T) for T-slit aperture.

20. Kettle Hat / Chapel-de-Fer (13th–15th c.)

Deep bowl. Steel. Open face. Shell rim sits high, leaving ears exposed (brim provides overhead cover but not aural coverage). Full planar brim. No neck defense. One-piece raised or multi-plate. Internal leathers. Chin strap.

S2–A3–F0–B3–N0–M2–C1–L3–R2

Note: If brim is distinctly up/downturned, use B4. If multi-plate skull, use C4.

21. Bascinet, Open (14th c.)

Deep bowl, often with slight fore-aft elongation. Steel. Open face. Ears covered. No brim. Mail aventail at rear. One-piece raised. Internal leathers.

S2–A1–F0–B0–N3–M2–C1–L3–R2

Note: Use S4 if elongation ratio exceeds threshold.

22. Bascinet, Visored / Klappvisor (14th–15th c.)

Deep bowl or ovoid. Steel. Movable visor (klappvisor or hounskull). Ears covered. Visor provides peak function. Mail aventail. One-piece raised. Internal leathers. Chin strap.

S2–A1–F4–B5–N3–M2–C1–L3–R2

Note: Use S4 if distinctly ovoid. Use B0 if visor does not project meaningfully.

23. Pig-Faced Bascinet / Hounskull (late 14th–15th c.)

Ovoid shell with pronounced fore-aft elongation. Steel. Long pointed visor (“pig face” or “dog face”) creates distinctive snout. Ears covered. Visor projects forward. Mail aventail. One-piece raised skull. Internal leathers.

S4–A1–F4–B5–N3–M2–C1–L3–R2

Note: The elongated skull + projecting visor define this as S4. Classic example of ovoid shell form.

24. Tall Bascinet / Grand Bascinet (15th c.)

High-crowned shell with pronounced verticality. Steel. Full visor + bevor or wrapper closing face. Ears covered. Mail aventail or plate gorget interface. One-piece raised. Internal leathers.

S5–A1–F5–B0–N3–M2–C1–L3–R2

Note: Distinguished from standard bascinet by crown height ratio >1.3×.

25. Barbute (15th c.)

Deep bowl, often ovoid with pronounced T-shaped or Y-shaped facial opening. Steel. The opening constitutes a fixed mask geometry. Ears covered. No brim. No attached neck defense. One-piece raised. Internal leathers.

S2–A1–F3a(T)–B0–N0–M2–C1–L3–R2

Note: Use S4 if distinctly ovoid.

26. German Sallet, Visored (c. 1450–1490)

Tail-swept shell with integral long rear extension. Steel. Movable visor. Partial ear exposure for hearing. Visor provides peak function. Integral tail. One-piece raised. Internal leathers with replaceable liner. Chin strap.

S2t–A2–F4–B5–N0–M2–C1–L3–R2

27. German Sallet, Open (c. 1450–1490)

Tail-swept shell with integral rear extension. Steel. Open face. Partial ear exposure. No brim. Integral tail. One-piece raised. Internal leathers.

S2t–A2–F0–B0–N0–M2–C1–L3–R2

28. Italian Sallet with Bevor (15th c.)

Tail-swept shell. Steel. No visor; separate detachable bevor provides lower-face coverage. Partial ear exposure. No brim. Integral tail. One-piece raised. Internal leathers. Chin strap.

S2t–A2–F7b–B0–N0–M2–C1–L3–R2

29. Armet (15th c.)

High crown or deep bowl with full facial closure via visor + hinged cheek plates that meet at chin. Steel. Ears enclosed when armed. No brim. Gorget interface or articulated neck defense common. One-piece skull. Internal leathers.

S5–A1–F5–B0–N5–M2–C1–L3–R2

Note: The hinged cheek closures combine with visor for F5 classification.

Renaissance & Early Modern (1500–1700 AD)

30. Close Helm (16th c.)

High crown or deep bowl. Steel. Visor + bevor closing face completely. Ears covered when armed. No brim. Articulated neck lames or gorget interface. One-piece skull. Internal leathers.

S5–A1–F5–B0–N5–M2–C1–L3–R2

Note: Often the best exemplar of helmet+gorget interface (N5).

31. Burgonet (16th c.)

Deep bowl. Steel. Hinged cheekpieces; face otherwise open. Peak only at brow. Ear coverage by shell or cheekpieces. Small nape flange or none. One-piece skull. Internal leathers.

S2–A1–F2b–B1–N1–M2–C1–L3–R2

Note: Optional falling buffe changes face to F7b.

32. Morion (16th c.)

High crown with pronounced verticality. Steel. Open face. Ears covered. Full brim with strong upturn at front and rear (turned). No neck defense. One-piece raised. Internal leathers.

S5–A1–F0–B4–N0–M2–C1–L3–R2

Note: High crown + turned-up brim is the recognizable signature (S5+B4).

33. Combed Morion (16th c.)

Very tall crown with pronounced medial comb/ridge running fore-aft. Steel. Open face. Ears covered. Full brim with strong upturn. No neck defense. One-piece raised. Internal leathers.

S5–A1–F0–B4–N0–M2–C1–L3–R2

Note: Extreme example of S5; comb can be recorded via optional E3 (reinforcing ribs/combs).

34. Pikeman’s Pot (17th c.)

Tall conical or high-crowned shell. Steel. Open face. Ears covered or partially. Partial brim or peak only. No neck defense or small flange. One-piece or multi-plate. Internal leathers.

S5–A1–F0–B2a–N0–M2–C1–L3–R2

Note: English Civil War era. Some examples are S3 (conical) rather than S5 (high crown).

35. Cabasset (16th c.)

Conical or high crown shell. Steel. Open face. Ears covered. Full brim, planar or slightly turned. No neck defense. One-piece raised. Internal leathers.

S3–A1–F0–B3–N0–M2–C1–L3–R2

Note: Taller examples drift into S5; brim treatment varies by region (B3 vs B4).

36. Lobster-Tail Pot / Zischägge (17th c.)

Deep bowl. Steel. Hinged cheekpieces, often with nasal bar. Peak at brow. Articulated laminated neck defense (the “lobster tail”). Multi-plate construction. Internal leathers.

S2–A1–F2b+F1–B1–N4–M2–C4–L3–R2

Note: Defining feature is articulated neck defense (N4) + peak (B1).

37. German Secret Helmet / Secrete (16th–17th c.)

Thin steel skullcap concealed under a hat. Open face. Ears exposed. No brim. No neck defense. Often multi-plate riveted, some one-piece. Internal leather band. Typically integrated into hat (no traditional harness).

S1–A3–F0–B0–N0–M2–C4–L3–R6

Note: If truly one-piece, use C1.

Islamic & Eurasian Helmets

38. Mongol/Turko-Mongol Helmet (13th–15th c.)

Conical shell rising to pointed apex, often with spike or plume holder. Iron or steel. Open face or with nasal. Ears exposed or partial coverage. No brim. Mail or lamellar curtain at rear. Multi-plate or one-piece construction.

S3–A3–F0–B0–N3–M2–C4–L2–R1

Note: Add F1 if nasal bar present.

39. Ottoman Chichak / Sipahi Helmet (15th–17th c.)

Tall conical shell, often fluted or with medial ridge. Steel. Nasal bar with sliding adjustment. Partial ear coverage. No brim. Mail curtain. One-piece forged. Internal padding.

S3–A2–F1–B0–N3–M2–C1–L1–R2

40. Persian Kulah Khud (16th–19th c.)

Tall conical or ovoid shell with spike finial. Steel, often decorated. Nasal bar with sliding guard. Ears partially exposed. No brim. Mail curtain. One-piece forged.

S3–A2–F1–B0–N3–M2–C1–L1–R2

Note: Use S4 if shell shows pronounced fore-aft elongation; S5 if crown height exceeds threshold.

41. Russian Shishak (16th–17th c.)

Tall conical shell with pronounced spike. Steel. Open face or with nasal. Partial ear coverage. No brim. Mail curtain or articulated neck defense. One-piece forged.

S3–A2–F0–B0–N3–M2–C1–L1–R2

Note: Add F1 if nasal present. Some examples have N4 articulated defense.

42. Mamluk Helmet (14th–16th c.)

High crown with pronounced verticality, often with medial ridge. Steel. Open face or with nasal. Ear coverage varies. No brim. Mail curtain. One-piece forged.

S5–A2–F0–B0–N3–M2–C1–L1–R2

Note: The tall crown distinguishes from standard conical (S3).

43. Indo-Persian Helmet / Top (16th–19th c.)

Ovoid shell with distinct fore-aft elongation, often with medial rib. Steel. Nasal bar and/or face guard. Partial ear coverage. No brim. Mail curtain. One-piece forged or raised.

S4–A2–F1–B0–N3–M2–C1–L1–R2

Note: The fore-aft elongation (>1.15× ratio) defines S4 classification.

44. Mughal Helmet (16th–18th c.)

High-crowned shell, often bulbous or with pronounced dome. Steel, frequently gilt or decorated. Nasal guard. Partial ear coverage. No brim. Mail or plated curtain. One-piece forged.

S5–A2–F1–B0–N3–M2–C1–L1–R2

Note: Distinguished by tall crown (height >1.3× width).

Japanese Helmets

45. Kabuto, Suji-Bachi (Multi-Plate) (15th–17th c.)

High crown or conical shell constructed from many vertical plates. Iron/steel. Face often open or fitted with separate menpō (detachable face guard). Peak common. Ears partially exposed. Articulated neck defense (shikoro) standard. Multi-plate riveted. Worn over padded cap.

S5–A2–F0–B1–N4–M2–C4–L2–R1

Note: Add F7b if menpō present; F3b for full masks.

46. Kabuto, Zunari (Few-Plate Bowl) (16th–17th c.)

Deep bowl or high crown from fewer, larger plates. Iron/steel. Face typically open or with separate menpō. Peak or none. Partial ear exposure. Articulated shikoro. Multi-plate construction. Padded cap worn under.

S2–A2–F0–B1–N4–M2–C4–L2–R1

Note: Simpler plate architecture than suji-bachi but same subsystem logic.

47. Jingasa (Ashigaru Hat-Helmet)

Conical hat form. Iron, bronze, or lacquered organic materials. Open face. Ears exposed or partially covered depending on depth. Partial brim. No articulated neck defense. One-piece or molded construction. Worn over cap.

S3–A3–F0–B2a–N0–M2–C1–L2–R1

Note: Construction varies: C2a for cast, C2b for molded organic.

World War I Era

48. British Brodie / Mk I (1915+)

Deep bowl. Steel. Open face. Shell rim sits high, leaving ears exposed (brim provides overhead cover but not aural coverage). Full planar brim. No neck defense. One-piece pressed. Webbing suspension. Chin strap.

S2–A3–F0–B3–N0–M2–C1–L4–R2

Note: Morphologically a kettle hat with modern liner/retention. Brim does not count toward A-series.

49. French Adrian M15 (1915+)

Deep bowl. Steel. Open face. Shell rim sits high, leaving ears exposed. Peak/visor at front. Attached rigid rear nape plate. Multi-plate construction (riveted crest and components). Internal leather band system. Chin strap.

S2–A3–F0–B1–N2–M2–C4–L3–R2

50. German Stahlhelm M16/M17/M18 (1916+)

Deep bowl with pronounced temporal/nape flare designed for ear and temple protection. Steel. Open face. Ears covered or partially covered by shell rim. Peak at front. No attached neck defense. One-piece pressed. Internal leather band system. Chin strap.

S2–A1–F0–B1–N0–M2–C1–L3–R2

Note: Use A2 if fit leaves ears partially exposed. Deep integral coverage captured by S2 rim drop.

51. U.S. M1917 (1917+)

Deep bowl (Brodie-derived). Steel. Open face. Shell rim sits high, leaving ears exposed. Full planar brim. No neck defense. One-piece pressed. Webbing suspension. Chin strap.

S2–A3–F0–B3–N0–M2–C1–L4–R2

World War II Era

52. British Mk II (1938+)

S2–A3–F0–B3–N0–M2–C1–L4–R2

Note: Morphologically a kettle hat with modern liner/retention. Brim does not count toward A-series.

35. French Adrian M15 (1915+)

S2–A3–F0–B1–N2–M2–C4–L3–R2

36. German Stahlhelm M16/M17/M18 (1916+)

Deep bowl with pronounced temporal/nape flare designed for ear and temple protection. Steel. Open face. Ears covered or partially covered by shell rim. Slight peaked brim. No attached neck defense. One-piece pressed. Internal leather band system. Chin strap.

S2–A1–F0–B1–N0–M2–C1–L3–R2

Note: Use A2 if fit leaves ears partially exposed. Deep integral coverage captured by S2 rim drop.

37. U.S. M1917 (1917+)

Deep bowl (Brodie-derived). Steel. Open face. Shell rim sits high, leaving ears exposed. Full planar brim. No neck defense. One-piece pressed. Webbing suspension. Chin strap.

S2–A3–F0–B3–N0–M2–C1–L4–R2

World War II Era

38. British Mk II (1938+)

Deep bowl. Steel. Open face. Shell rim sits high, leaving ears exposed. Full planar brim. No neck defense. One-piece pressed. Webbing suspension. Chin strap.

S2–A3–F0–B3–N0–M2–C1–L4–R2

Note: Same morphological signature as Brodie; changes are dimensional/liner refinements.

39. U.S. M1 Helmet (1941–1980s)

Deep bowl. Steel outer shell. Open face. Ears partially covered by shell rim. No brim. No neck defense. One-piece pressed outer + rigid liner shell carrying suspension. Two-shell system. Chin strap.

S2–A2–F0–B0–N0–M2–C7–L6–R2

40. Soviet SSh-40 (1940+)

Deep bowl with pronounced temporal coverage. Steel. Open face. Ears covered or partially covered by shell rim. No brim. No neck defense. One-piece pressed. Internal leather band system. Chin strap.

S2–A1–F0–B0–N0–M2–C1–L3–R2

Note: Use A2 if fit leaves ears partially exposed.

41. German Stahlhelm M35/M40/M42 (1935–1945)

Deep bowl with characteristic flared temporal coverage. Steel. Open face. Ears covered or partially covered. Slight peaked brim. No attached neck defense. One-piece pressed. Internal leather band system. Chin strap.

S2–A1–F0–B1–N0–M2–C1–L3–R2

Note: Refinement of WWI Stahlhelm; same typological signature. Use A2 if fit leaves ears partially exposed.

Modern Ballistic Helmets

42. PASGT Helmet (1983–2000s)

Deep bowl. Aramid composite. Open face. Ears covered by shell rim. Slight peaked brim. No neck defense. Molded composite shell. Webbing basket suspension (~12mm standoff). Four-point retention.

S2–A1–F0–B1–N0–M5–C6–L4–R4

43. ACH (Advanced Combat Helmet) (2002+)

Deep bowl. Aramid composite. Open face. Ears covered. No brim. No neck defense. Molded composite shell. Pad-based energy-absorbing liner. Four-point retention.

S2–A1–F0–B0–N0–M5–C6–L5–R4

44. ECH (Enhanced Combat Helmet) (2013+)

Deep bowl. UHMWPE composite. Open face. Ears covered. No brim. No neck defense. Molded composite shell. Pad-based liner. Four-point retention.

S2–A1–F0–B0–N0–M5–C6–L5–R4

45. Ops-Core FAST / High-Cut Helmets (modern)

Deep bowl with high-cut ear area for communications headsets. Ballistic composite. Open face. Ears exposed by shell cut. No brim. No neck defense. Molded composite shell. Pad-based liner. Four-point retention.

S2–A3–F0–B0–N0–M5–C6–L5–R4

Note: High-cut (A3) distinguishes from full-coverage shells (A1).

46. IHPS (Integrated Head Protection System) (2019+)

Deep bowl with modular appliqué options. UHMWPE composite. Open face in base configuration. Ear coverage depends on configuration. No brim. Modular mandible/visor options. Molded composite. Pad-based liner. Four-point retention.

S2–A1–F0–B0–N0–M5–C6–L5–R4

Note: Base configuration shown; add F7b if mandible fitted.

47. NovaSteel Helmet, Base Configuration (modern)

Deep bowl. Steel. Open face. Ears exposed by shell cut. No brim. No neck defense. One-piece pressed/formed shell. Pad-based energy-absorbing liner. Four-point retention.

S2–A3–F0–B0–N0–M2–C1–L5–R4

Note: Morphologically similar to 16th–17th c. secret helmets, with modern liner/retention.

48. NovaSteel Helmet with Fixed Mandible (modern)

Deep bowl. Steel. Rigid faceplate detachable by removing four bolts; horizontal slot aperture (frogmouth orientation). Ears covered. No brim. No neck defense. One-piece pressed shell. Pad-based liner. Four-point retention.

S2–A1–F3b(f)–B0–N0–M2–C1–L5–R4

The post Universal Combat Helmet Typology appeared first on Adept.

The LLNL SiC Mystery

Adept Armor — Mon, 08 Dec 2025 07:32:03 +0000

The LLNL SiC Mystery

I’ve posted previously about some oddities in the LLNL Light Armor Program’s ceramic test results. There’s another oddity that, in itself, was relatively minor, but proved unfortunately influential for decades: Their experiments with silicon carbide.

As you can see in the table below, silicon carbide (SiC) didn’t perform very well. Its figure of merit was 0.76 – well under boron carbide’s (B4C) baseline of 1.0, and only marginally better than high-purity alumina at 0.73. In practical terms: if it would take a 10-pound B4C tile to offer a given level of protection, you’d need a 13.16-pound SiC tile, or a 13.70-pound alumina tile.

That’s on a weight basis. On a thickness basis, it gets even worse: the LLNL’s SiC was the worst-performing material of the three. Their SiC required approximately 7% more thickness than B4C and 24% more thickness than alumina to provide equivalent ballistic protection.

Yet today SiC totally reverses this paradigm, outperforming both B4C and Al2O3 on a thickness basis. In modern ceramic armor plate production, given a certain margin of safety, a 7.8mm thickness of high-quality SiC will reliably stop a common reference AP threat, whereas high-grade boron carbide requires no less than 8.3mm; in this case, SiC needs approximately 6% less thickness than B4C. Compared against alumina, the relation varies (alumina’s performance depends heavily on its compressive strength, hardness, and density, all of which track with purity grade), but SiC generally outperforms Al2O3 on both thickness and weight.

In fact, the best grades of modern SiC approach the weight efficacy of today’s best grades of boron carbide, and militaries around the world have been moving to SiC as their workhorse ceramic armor material.

So why did the LLNL report such poor results?

The answer is that their tests were run in the late 1960s, and before 1974 high-quality SiC was effectively unsinterable. They were testing “SiC” that by today’s standards would be regarded as unfit for purpose.

The Sintering Problem

In the 1960s, the only known way to densify non-oxide ceramics such as carbides and borides to their theoretical full density was via hot-pressing. B4C and most other non-oxide ceramics were highly amenable to this process – but SiC was stubbornly resistant. It proved almost impossible to produce high-quality SiC via the straightforward hot-pressing of pure SiC powders.

The problem is thermodynamic. At elevated temperatures, silicon carbide doesn’t simply soften and flow like most ceramics. Instead, it decomposes: the silicon sublimates, leaving behind a porous carbon skeleton. This decomposition begins at temperatures well below those needed to achieve full densification through solid-state diffusion. Hot-pressing pure SiC powders typically yields a weak, porous body riddled with voids – or, if pushed too hard, a pile of graphite and silicon vapor.

Throughout the 1960s, researchers tried various workarounds. Some used extremely high pressures. Others experimented with additives. The results were inconsistent at best. The SiC that the LLNL tested was almost certainly produced by one of these early, imperfect methods – likely a hot-pressed material with significant residual porosity, decomposition damage, or both.

The Breakthrough

Everything changed in 1974, when Svante Prochazka at General Electric discovered that small additions of boron and carbon could enable pressureless sintering of SiC to near-theoretical density. The mechanism was elegant: Boron segregates to the grain boundaries and dramatically accelerates diffusion, while carbon removes the thin oxide layer (SiO2) that normally coats SiC particles and inhibits sintering.

This was a genuine revolution. Suddenly, high-density SiC could be produced in conventional furnaces without the massive hydraulic presses that hot-pressing required. Costs dropped. Quality improved. Complex shapes became feasible.

The sintered SiC that emerged from this process bore little resemblance to the porous, compromised material of the 1960s. It was fully dense, fine-grained, and mechanically excellent – with hardness and elastic modulus values far beyond anything the LLNL researchers encountered in their “SiC”.

The Legacy Problem

Unfortunately, the LLNL’s results had already been published and widely circulated. For years afterward, armor designers looked at that 0.76 figure of merit and concluded that SiC wasn’t worth serious consideration. Why bother with a material that barely outperformed cheap alumina and fell far short of B4C?

It took considerable time for the armor community to recognize that Prochazka’s sintered SiC was, for all practical purposes, a different material entirely. The chemistry was nominally the same – silicon carbide is silicon carbide – but the microstructure, density, and mechanical properties were transformed.

Today, SiC is one of the dominant materials in ceramic armor. It offers an exceptional combination of high hardness, good stability, moderate density, and acceptable multi-hit capability. The best commercial grades achieve densities above 3.15 g/cm³ (theoretical density is 3.21 g/cm³), with Vickers hardness values exceeding 2500 HV and fracture toughness around 4–5 MPa·m^½.

The LLNL’s poor results weren’t wrong, exactly – they accurately reported what they observed. But what they observed was a material that no longer exists in any meaningful sense. Their “SiC” and modern sintered SiC share a name and a chemical formula, but little else.

The post The LLNL SiC Mystery appeared first on Adept.

Adept Armor Threat Assessment Series I: Standard-Issue Military Rifle Ammunition

Adept Armor — Wed, 29 Oct 2025 08:19:19 +0000

Adept Armor Threat Assessment Series I: Standard-Issue Military Rifle Ammunition

When designing plates that are built to counter common military small-arms threats, it helps to have a reference guide which summarizes the various characteristics of those threats.

Below is the Adept survey of typical standard-issue small arms projectiles. Field order is:

US, Europe, Canada, UK

M193 Ball (5.56×45mm, US/NATO)
55 gr | lead-antimony core, FMJ | – | – | 12HV | 990 m/s (20″) | 3,248 fps | 1,289 ft⋅lb | 2

M855 / SS109 Ball (5.56×45, US/NATO)
62 gr | steel tip + lead rear | 4.6 mm | 10 gr | 40–45 HRC | 915–950 m/s (20″) | 3,003–3,117 fps | 1,240–1,330 ft⋅lb | 3

M855A1 EPR (5.56×45, US)
62 gr | exposed hardened‑steel tip + copper slug (lead‑free) | 4.3 mm | 19 gr (steel) | 58–60 HRC | 905–960 m/s (14.5–20″) | 2,970–3,150 fps | 1,214–1,366 ft⋅lb | 5

AP45 (5.56×45, Nammo) (Note: Very rare!)
69.4 gr | sub‑cal. WC‑Co penetrator + Al cup + Cu jacket | 3.9–4.0 mm | 41–43 gr | WC (70–75 HRC) | 900 m/s | 2,953 fps | 1,344 ft⋅lb | 7

Swiss P AP (5.56×45, Germany / RUAG)
63 gr | sub‑cal. WC‑Co core + Cu jacket | 4.0 mm | 34–36 gr | WC (70–75 HRC) | 930–960 m/s (20″) | 3,051–3,150 fps | 1,303–1,389 ft⋅lb | 7

L31A1 EP Ball (5.56×45, UK / BAE)
62 gr | full‑length hardened‑steel core + Cu jacket (lead‑free) | 5.0 mm | 52 gr | 58–62 HRC | 900–950 m/s | 2,953–3,117 fps | 1,201–1,338 ft⋅lb | 7

M995 AP (5.56×45, US/NATO / Nammo)
52 gr | sub‑cal. WC‑Co core + aluminum cup | 4.0 mm | 32 gr | WC (70–75 HRC) | 1,030 m/s (20″) | 3,379 fps | 1,319 ft⋅lb | 6.5

M80 Ball (7.62×51, US/NATO)
147 gr | lead‑alloy core, FMJ | – | – | – | 850 m/s (22″) | 2,789 fps | 2,539 ft⋅lb | 2

M80A1 EPR (7.62×51, US)
130 gr | hardened‑steel tip + copper slug (lead‑free) | 5.5 mm | 45 gr (steel) | 50–55 HRC | 847–915 m/s (16–22″) | 2,780–3,002 fps | 2,232–2,600 ft⋅lb | 5

Swiss P AP (7.62×51, RUAG)
196 gr | WC‑Co + lead hybrid construction | 5.7 mm | 86 gr | WC (75 HRC) | 790–810 m/s (22″) | 2,592–2,657 fps | 2,925–3,075 ft⋅lb | 9

M61 AP (7.62×51, US)
150.5 gr | hardened‑steel core + lead filler | – | 55 gr | 60–63 HRC | 855 m/s (22″) | 2,805 fps | 2,630 ft⋅lb | 5.5

M993 AP (7.62×51, US/NATO)
128 gr | WC‑Co core + aluminum cup | 5.5 mm | 91 gr | WC (70–75 HRC) | 930 m/s (22″) | 3,051 fps | 2,647 ft⋅lb | 8.75

XM1186 GP (6.8×51, US; EPR‑type)
135 gr | hardened‑steel penetrator + Cu slug; hybrid case | 5.5–6.0 mm | 30–40 gr (steel) | 58–62 HRC | 900–915 m/s (13–16″) | 2,953–3,002 fps | 2,613–2,702 ft⋅lb | 7

L2A2 / SS109 Ball (5.56×45, UK)
62 gr | steel tip + lead | 4.6 mm | 10 gr | 40–45 HRC | 920–940 m/s | 3,018–3,084 fps | 1,254–1,309 ft⋅lb | 3

L59A1 “High Performance” Ball (7.62×51, UK)
155 gr | hardened‑steel tip + lead core | 5–6 mm | 15–20 gr | 58–62 HRC | 838 m/s | 2,749 fps | 2,602 ft⋅lb | 5

DM151 AP / Hartkern (7.62×51, Germany)
151 gr | WC‑Co core, lead‑free | 5.0 mm | 70 gr | WC (75 HRC) | 825 m/s | 2,707 fps | 2,458 ft⋅lb | 9

SS109 Ball (MEN/BAE/etc., 5.56×45, Europe/NATO)
62 gr | steel tip + lead | 4.6 mm | 10 gr | 40–45 HRC | 920–940 m/s (20″) | 3,018–3,084 fps | 1,254–1,309 ft⋅lb | 3

C77 Ball (SS109‑equiv., 5.56×45, Canada)
62 gr | steel tip + lead | 4.6 mm | 10 gr | 40–45 HRC | 940 m/s | 3,084 fps | 1,309 ft⋅lb | 3

C21 / NATO Ball (M80‑equiv., 7.62×51, Canada/NATO)
146–147 gr | lead core, FMJ | – | – | – | 845 m/s (22″) | 2,772 fps | 2,492–2,508 ft⋅lb | 2

.338 AP (e.g., AP485/AP529, .338 Lapua Magnum, Finland/UK/NATO)
248–300 gr | WC‑Co core (tungsten AP) | 7.0 mm | 120–200 gr | WC (70–75 HRC) | 840–905 m/s (26–27″) | 2,756–2,970 fps | 4,186–5,878 ft⋅lb | 10

Russia, Ukraine

7N6 Ball (5.45×39, Russia)
53 gr | mild‑steel core + air cavity/lead | 4.0 mm | 20–22 gr | 40–45 HRC | 880 m/s | 2,887 fps | 981 ft⋅lb | 3

7N10/7N10M “EP” Ball (5.45×39, Russia)
56 gr | hardened‑steel core (U12A) | 4.1 mm | 25 gr | 60 HRC | 880 m/s | 2,887 fps | 1,037 ft⋅lb | 5

7N22 AP (5.45×39, Russia)
57 gr | hardened tool‑steel core | 4.0 mm | 27 gr | 60–65 HRC | 890 m/s | 2,920 fps | 1,080 ft⋅lb | 5.5

7N24 “Super‑AP” (5.45×39, Russia)
64 gr | WC‑Co core (1.8 g) | 4.0 mm | 28 gr | WC | 840 m/s | 2,755 fps | 1,080 ft⋅lb | 6

7N39 “Igolnik” AP (5.45×39, Russia)
63.3 gr | WC‑Co core (1.9 g) | 4.0 mm | 29 gr | WC | 850 m/s | 2,789 fps | 1,094 ft⋅lb | 7

57‑N‑231 “PS” Ball (7.62×39, Russia/North Korea/export)
122–123 gr | mild/heat‑treated steel slug + lead sheath | 5.6 mm | 55–60 gr | 35-45 HRC | 710–730 m/s (16″) | 2,329–2,395 fps | 1,471–1,567 ft⋅lb | 1

7N23 “BP” AP (7.62×39, Russia/North Korea)
122–123 gr | hardened‑steel core | 5.0 mm | 60 gr | 60 HRC | 710–730 m/s | 2,329–2,395 fps | 1,471–1,567 ft⋅lb | 5

LPS Light Ball / 57‑N‑323 (7.62×54R, Russia/export)
148–150 gr | steel core (mild) | – | – | – | 825–860 m/s | 2,707–2,822 fps | 2,408–2,653 ft⋅lb | 3

7N13 “BP” EP (7.62×54R, Russia)
145 gr | heat‑strengthened steel core | 6.5 mm | 70 gr | 55–60 HRC | 828 m/s | 2,717 fps | 2,378 ft⋅lb | 6

7N14 AP (sniper) (7.62×54R, Russia)
152–160 gr | hardened‑steel / special core | – | – | 60+ HRC | 840 m/s | 2,756 fps | 2,564–2,700 ft⋅lb | 6

B‑32 API (MG/sniper) (7.62×54R, Russia/Ukraine/export)
166 gr | steel core + incendiary | – | – | 60 HRC | 790–810 m/s | 2,592–2,657 fps | 2,477–2,602 ft⋅lb | 5

China

DBP87/95 Ball (legacy, 5.8×42, China)
64–67 gr | steel core + lead base | 3.8–4.0 mm | 20–24 gr | 50–55 HRC | 890–930 m/s (18″) | 2,920–3,051 fps | 1,211–1,385 ft⋅lb | 4

DBP10 Ball (current, 5.8×42, China)
71 gr | hardened‑steel core (lead‑free) | 3.8 mm | 25 gr | 55–60 HRC | 915 m/s (18″) | 3,002 fps | 1,421 ft⋅lb | 5

Tungsten‑core AP (“DVC‑12”, 5.8×42, China)
84–85 gr | WC‑Co core (3.5 g) | 4.5 mm | 54 gr | WC | 880–900 m/s | 2,887–2,953 fps | 1,555–1,646 ft⋅lb | 7.5

.338 AP (8.6×70 mm, China)
Estimates/class-typical:
248–300 gr | WC‑Co penetrator (tungsten‑carbide) + copper jacket (assessed, AP‑type) | ~7.0 mm (bullet) | ~120–200 gr core (class‑typical) | WC (70-75 HRC equiv.; ~1200–1500 HV) | 2,756–2,970 ft/s (840–905 m/s) (26–27″) | 4,200–5,900 ft⋅lb | FoM: 10

Rest of World

CBC M963 Ball (7.62×51, Brazil)
144 gr | lead core, FMJ | – | – | – | 833 m/s (24″) | 2,733 fps | 2,389 ft⋅lb | 2

CBC “Perfurante” AP (7.62×51, Brazil)
148 gr | hardened‑steel core | 6.5 mm | 65–70 gr | 58–62 HRC | 838 m/s (24″) | 2,749 fps | 2,483 ft⋅lb | 6

SS109‑equiv. Ball (5.56×45, CBC/SEDENA/INDUMIL; BR/MX/CO)
62 gr | steel tip + lead | 4.6 mm | 10 gr | 40–45 HRC | 900–940 m/s | 2,953–3,084 fps | 1,201–1,309 ft⋅lb | 3

NATO Ball (M80‑equiv., 7.62×51, Mexico, Columbia)
147 gr | lead core, FMJ | – | – | – | 830–850 m/s | 2,723–2,789 fps | 2,422–2,539 ft⋅lb | 2

PS Ball (Type 56/AK‑103, 7.62×39, Venezuela)
122–123 gr | mild/heat‑treated steel core + lead | 5.6 mm | 55–60 gr | 35-45 HRC | 710–730 m/s | 2,329–2,395 fps | 1,471–1,567 ft⋅lb | 1

7N23 “BP” AP (if issued, 7.62×39, Venezuela)
122–123 gr | hardened‑steel core | 5.0 mm | 60 gr | 60 HRC | 710–730 m/s | 2,329–2,395 fps | 1,471–1,567 ft⋅lb | 5

Type 89/20 (SS109‑equiv. Ball, 5.56×45, Japan)
62 gr | steel tip + lead | 4.6 mm | 10 gr | 40–45 HRC | 900–940 m/s | 2,953–3,084 fps | 1,201–1,309 ft⋅lb | 3

M80-equivalent NATO Ball (MG use, 7.62×51, Japan)
147 gr | lead core, FMJ | – | – | – | 830–850 m/s | 2,723–2,789 fps | 2,422–2,539 ft⋅lb | 2

SS109‑equiv. Ball (Poongsan, 5.56×45, South Korea)
62 gr | steel tip + lead | 4.6 mm | 10 gr | 40–45 HRC | 900–940 m/s | 2,953–3,084 fps | 1,201–1,309 ft⋅lb | 3

M80-equivalent NATO Ball (K12 MG, 7.62×51, South Korea)
147 gr | lead core, FMJ | – | – | – | 830–850 m/s | 2,723–2,789 fps | 2,422–2,539 ft⋅lb | 2

The Figure of Merit –

The numeral at the end of each projectile’s line is a general estimate of their ability to penetrate hard armor. This was derived via calibrating to a well-documented 5.56mm dataset in 500HB steel at 900 m/s and extended by conservative physics-based scaling to other rounds. It’s a rough guideline and is not intended to be authoritative, but simply to indicate how these bullets compare in general terms. Without such a heuristic, many people default to kinetic energy, but KE alone is a very poor measure.

Notes on Armor Plate Design –

By a wide margin the most common military standard-issue small-arms threats outside Russia and China are M855/SS109-type 5.56mm, M80-type 7.62mm, and 7.62x39mm MSC. The M855/SS109 type is particularly ubiquitous. In itself, this makes a pretty good case for the military relevance of the NIJ 0101.07 RF2 specification.

But if China and Russia are considered, RF2 plates seem woefully inadequate. The standard “ball” rounds 7N10/7N10M and DBP10 – to say nothing of 7N39 and DVC-12 – are considerably tougher threats than M855, and would likely penetrate light RF2 plates even at a significant standoff. Our modeling suggests that building a plate to stop DBP10 requires a ceramic (SiC-TiB2 or B4C) thickness of roughly 6mm – not terribly far from what an RF3 plate would minimally require. And DVC-12, in particular, appears to require something beyond the average Level IV/RF3 plate.

From a Western perspective, the optimal plate for war is a lightweight RF3 that can also reliably handle threats like 7N39 and DVC-12. Something is in the works.

The post Adept Armor Threat Assessment Series I: Standard-Issue Military Rifle Ammunition appeared first on Adept.

Understanding the Ballistic Event in Ceramic Armor

Adept Armor — Wed, 01 Oct 2025 07:07:54 +0000

Understanding the Ballistic Event in Ceramic Armor

When a hard projectile impacts a ceramic-faced armor plate at velocities ranging from 400 to 2000 meters per second, two primary factors must be considered: (1) the performance characteristics of the projectile (penetrator) and (2) the response of the ceramic armor plate. The interaction between them occurs on an extraordinarily short timeframe – typically just microseconds – and is governed by dynamic effects such as high strain rates, localized deformation, and complex stress wave phenomena. It can generally be split into three phases.

Phase 1: Dwell

Projectile Behavior:

When an armor-piercing (AP) bullet core comes into contact with the ceramic strike-face of an armor plate, it is momentarily arrested before resuming forward motion – a phenomenon known as “dwell.” This brief event lasts an average of 2-12 microseconds. During dwell, the penetrator experiences rapid deceleration, and local temperatures rise significantly due to kinetic energy conversion. The penetrator’s tip, typically composed of hardened steel or a tungsten carbide-cobalt cermet (WC-Co), begins to erode or shatter as it’s shocked upon contact.

The aforementioned 2-12µs range is broad for an average, but it’s dependent on a number of variables, most critically the stiffness of the backing layer behind the ceramic layer. There is a direct correlation between backing layer stiffness and the duration of this dwell phase. In extreme cases, with a very thick backer such as a “semi-infinite” steel block, it could far exceed 12µs. Other factors include the acoustic impedance of the ceramic layer, the backing layer behind the ceramic strike-face, and the adhesive holding them together.

Armor Response:

Upon impact – assuming compressive failure dominates and tensile fracture is suppressed – fractures rapidly propagate in the ceramic in a characteristic inverted cone pattern known as the “fracture conoid.” The conoid typically measures (approximately!) three projectile diameters in base diameter, with a semi-angle ranging from 60° to 70°, commonly approximated as 68° in ballistic modeling. This conical fracture pattern is critical, as its formation and size determine the amount of ceramic material actively resisting projectile penetration.

In a very real sense, the conoid is a momentum trap. The momentum of the projectile is converted to the momentum of the (usually) much larger and more massive conoid, which leads to deceleration and a steep reduction in applied force per unit area.

Current (2025) research in ceramic armor seeks to determine how various mechanical properties – such as hardness, fracture toughness, grain size, and Young’s modulus – impact conoid volume and mass. Though it’s intuitive that stronger ceramics with higher flexural strengths, and thus a lower defect density, would appear beneficial, no definitive, universal correlation has yet been established between any property and conoid side angle.

Phase 2: Stress Waves and Deformation

Projectile Behavior:

Immediately following the initial dwell phase, a tensile stress wave is reflected back into the projectile from the ceramic surface, rapidly following the initial compressive wave. These alternating stress conditions place extreme mechanical demands on the projectile, often causing internal cracks and subsequent fracture. The penetrator’s materials must exhibit resilience against these rapidly shifting loads or fragmentation occurs. Steel penetrators often outperform brittle WC-Co cores under these conditions due to their higher ductility and better shock resistance, although exceptionally hard steels may still fail catastrophically.

Armor Response:

Simultaneously, stress waves propagate through the ceramic armor plate itself. A shear wave travels axially and radially through the ceramic, eventually reaching the backing layer – usually made from high-performance fiber composite materials such as Dyneema, Spectra, or Kevlar. Upon interaction with the backing layer, the shear wave triggers deformation and compression, leading to energy absorption through plastic deformation, fiber breakage, and interlaminar delamination. The backing layer’s deformation also serves to mitigate stresses at the interface between the ceramic and the backer.

There’s an excellent video of stress wave propagation in a transparent cube here:

Phase 3: Final Defeat or Penetration

Armor Response:

By the end of the second phase, the ceramic within the fracture conoid region is extensively comminuted – in other words, pulverized – broken into fine, closely-packed granular fragments. Larger radial and axial fractures significantly compromise the structural integrity of the ceramic plate beyond the immediate conoid. The effectiveness of the ceramic’s fragmented particles plays a crucial role; tightly packed granular debris provides substantial resistance to projectile penetration and can serve to erode the AP bullet’s core, dramatically reducing its mass.

Projectile Behavior:

If the penetrator survives initial fragmentation relatively intact, it attempts to continue through the now granular ceramic debris. During this stage, the projectile experiences significant erosion, reducing its effective mass, diameter, and kinetic energy. This remnant projectile, greatly diminished in its threat potential, then interacts with the backing layer.

The backing layer’s primary role at this juncture is to dissipate the residual kinetic energy through deformation, stretching, delamination, and fiber breakage. Effective backing materials catch projectile fragments long before they perforate the plate. Foams are often added behind the backing layer – on the body side of the armor plate – for deformation mitigation and impact energy management. And, very frequently, for comfort.

Optimizing the synergy between ceramic and backing layers is central to modern armor design.

Impact Velocity Considerations Low-Velocity Impacts (<400 m/s):

At velocities below 400 m/s, particularly below 250 m/s, structural factors dominate armor performance. Damage to the ceramic strike-face can be relatively extensive, and a fracture conoid will still usually form, but the low energy of the impact event will be spread out over a wide area and will be easily captured by the ceramic backing layer. Failures in ceramic armor systems at these lower velocities are rare. However, other considerations such as multi-hit capability and deformation of the backing layer become increasingly relevant.

High-Velocity Impacts (>2000 m/s):

At velocities above 2000 m/s, conventional solid mechanics approaches become insufficient. The extreme pressures and strain rates render projectile and armor behaviors fluid-like, which necessitates hydrodynamic modeling. This regime is particularly relevant for military armored vehicles facing shaped-charge threats, where projectile-armor interactions resemble fluid dynamics more than conventional solid impacts.

Ultravelocity Impacts (7000–12,000 m/s):

At these extraordinarily high velocities, such as those encountered in space environments by micrometeoroids, impact energies are so immense that both projectile and armor materials partially or completely vaporize upon contact. For this reason, satellites and spacecraft typically employ multi-layer spaced armor systems known as Whipple shields, incorporating an initial thin bumper layer to disrupt and vaporize micrometeoroids, followed by a substantial air gap, a secondary layer of ceramic and aramid fabric materials to capture remaining fragments, and a second aluminum bumper or structural skin. This arrangement effectively mitigates catastrophic damage from hypervelocity impacts.

The post Understanding the Ballistic Event in Ceramic Armor appeared first on Adept.

Ballistic Shields: What They Are and How to Use Them

Adept Armor — Thu, 28 Aug 2025 07:57:11 +0000

Ballistic Shields: What They Are and How to Use Them

Note: This is a pared-down and updated version of two articles we wrote for Police and Security News in early 2023, titled “Ballistic Shield Technologies and Tactics.” The technologies haven’t changed much, if at all, since early 2023; the tactics are a summary of best practices developed by American and European teams over the past twenty years.

Modern ballistic shields are man-portable hard-armor barriers. They often share little with wood and iron ancestors besides the name and the shape; they’re held differently on account of their weight, and they’re utilized very differently. They’re closest to the historical pavise — a large shield that doubles as mobile cover — for indeed most contemporary shields allow a trained user to carry cover with them, shape angles in rooms and hallways, and buy the seconds needed to communicate, move, or solve a problem with less risk.

The construction of ballistic shields mirrors the construction of hard body armor. The most common recipe consists of nothing more than a pressed UHMWPE fiber composite. When steel-core threat defeat is required, a technical ceramic strike face can be bonded to a UHMWPE backer; a layer of aluminum oxide, silicon carbide, or boron carbide up front breaks and blunts the M855 penetrator or AP bullet core. Para-aramid fabric in resin or UHMWPE tape or fabric in polyurethane behind it traps fragments and absorbs residual energy. Ceramic layers on the order of 0.13 to 0.40 inches over composite backers around 0.25 to 0.60 inches are typical. Handgun-rated shields are virtually always all-composite. Some specialized or budget designs use ballistic steel. The right choice follows two questions that should be answered first: The threat level the shield must defeat and the weight the user can truly carry and control.

Shields are often described with NIJ 0108.01. Levels II and IIIA cover common handgun threats. Level III addresses most rifle ball rounds. Level IV is intended for steel-cored AP threats and all rifle ball, including magnum and high velocity. You will also encounter “Level III+.” That label has no formal definition, though it most frequently implies performance beyond Level III against specific steel-core ball threats. The standard dates to 1985 and remains serviceable for shield discussion.

Sometimes the new 0101.07 and 0123.00 standards might be used, in which case HG2 basically corresponds to IIIA, RF1 to Level III, RF2 to III+M855, and RF3 to Level IV.

Weight governs everything from carry time to whether the tool gets used at all. Here a thought experiment helps: If a shield covered only one square foot, a typical IIIA would weigh about two pounds, a Level III roughly 4.2 pounds, and a high-quality ceramic-composite Level IV just over 8.5 pounds, inclusive of paint, trim, and a light handle. Real shields are much larger. An 18 by 24 inch shield triples that one-square-foot area, which is why IIIA dominates day-to-day use, III is less common outside specialized teams, and most Level IV shields end up on wheels and behave more like compact mobile barricades than handheld tools.

Transparent armor is not weight-efficient. A rifle-rated 24 by 36 inch shield with a viewport can weigh 46 pounds while the same shield without glass weighs 30. The 16-pound penalty buys true see-through and safer aiming from behind armor, but fatigue is a real cost. An emerging alternative is to mount small, efficient cameras or sensors to the strike face and feed a tablet-class display on the user side, which approximates a multi-spectrum window at an added mass measured in ounces. The trade-off is that aiming through a simulated viewport is less intuitive and demands careful setup of camera placement and viewing angle, plus dedicated practice to prevent disorientation.

Support hardware changes what is possible, especially with heavier shields. The Army’s steadicam-style Third Arm concept was built to offload weapon weight, reduce fatigue, and improve stability. Shield work is rarely done in the prone, and shields are heavier than rifles, so the benefits transfer cleanly. Developers have already produced shield-specific brackets. Exoskeletal carriers that couple the load to the hips or torso, such as the Reaper, let an operator move and shoot accurately while bearing very heavy rifle-rated shields. Expect future shields to include dedicated hard points so the operator can free a hand for reloading, opening doors, or running a long gun without losing coverage.

Handling fundamentals are simple and, with training, can turn into reflex. Hold the shield away from the body, square to the threat, with little or no cant. If there is a viewport, bring the helmet into light contact with the top rim to stabilize eye relief while keeping the head and neck behind armor. Without a viewport, either keep the face fully behind the shield and take brief, deliberate peeks, or run the top rim at roughly nose level to protect torso, neck, and much of the face while retaining visibility. Balance well with a tall forward stance, the lead foot a half-step behind the strong-side foot, and the lead toes pointed forward. The shield operator sets the pace for the element, so lateral movement, quick pivots, and short controlled steps must be drilled until smooth.

Shields change handgun ergonomics. Contact between the pistol slide and the shield invites malfunctions, and bracing the forearm against the shield induces horizontal drift. Train malfunction clears and reloads explicitly. Reloading behind a shield is slower in the best case, which is why many users stage a second loaded sidearm. If your shield includes long-gun cutouts, use them only if you can keep a proper hold, full mobility, and the shield square. If supporting a carbine forces you to shunt the shield aside or angle it down, switch to a handgun.

The shield opens tactics that are unique to it. Lipping anchors a shield edge on a wall or door frame so you can pivot around that feature while keeping coverage for both the operator and people behind. Splitting biases coverage toward an expected threat vector, for example rotating the shield outward and left for a right-handed operator when danger is likely from that side. In close control or during an arrest, lower your center of gravity, widen the stance, and cant the shield outward slightly so it cannot be easily grabbed. If a compliance strike is unavoidable and your handle allows it, strike with the edge rather than the face. Edge strikes concentrate force, resist grabs, and reduce the risk of damaging the armor package. Remember that non-steel ballistic shields are not riot shields; even moderate blunt impacts on the order of a few dozen joules can delaminate UHMWPE backers, and ceramic faces can crack. Treat the shield as precision equipment and inspect it after abuse or a ballistic event.

Door control is a core skill. Practice one-handed pull-open with the shield offset just enough to keep the muzzle line clean, then re-square immediately so the door edge does not trap the shield. Learn to use the door as a movable piece of cover: open six inches, lip, read the slice, then open farther. On hinged-in doors, plan the arc so the door does not knock the muzzle or hand. On push-in entries, drive the door with the shield edge, accept the brief exposure as the slab moves, then reseat on the frame and clear. When the door opens into a hard corner, bias the split toward that corner and clear pie slices before committing through the threshold.

Hallways reward speed and discipline. On center-feed rooms, the shield takes the midline and holds center chest height, allowing the second operator to roll past the strong side. On corner-feed rooms, rotate the shield toward the long wall (split to the threat) and lead with short, quiet steps. When a hallway is too narrow for two-abreast, adopt a “bump and drag” rhythm where the second operator briefly steps up to shoot, then settles back behind cover as the shield advances a shoulder’s width.

Stairs require deliberate geometry. On the ascent, keep the shield slightly elevated, square to the risers, and minimize the gap under the rim where an elevated threat could see shoes first. On the descent, keep the rim lower than usual and bias the split toward the most open flank of the stairwell. If a second shield is available, stacking them vertically in a bookend arrangement protects both head and lower legs during long climbs.

In low light, a shield-mounted white light aimed slightly down prevents washout in viewports and reduces back-scatter into the eyes. If you run optics or a laser on the handgun, co-witness the light to avoid chasing two points in peripheral vision. With simulated viewports that use cameras and a screen, rehearse the cadence for switching from IR to visible and back so you do not lose the reticle in a mode change at the threshold.

Vehicle work benefits from a clear sequence. On approach, angle the shield to mask the operator’s hips and thighs from the suspect’s probable firing arcs. At the door, lip the B-pillar for a safe peek into the passenger compartment, then either wedge the shield vertically at the hinge line to hold the door open or retract and move to the next window. When extracting a noncompliant subject, use the shield to isolate a single arm for the contact officer rather than trying to pin the entire torso.

Multi-shield tactics scale coverage and let teams cross open danger areas. Two shields can form a wedge where the lead keeps the long axis forward and the wing shield covers the flank with a slight overlap. In a long hallway, a T-shape works well, the second shield forming the cross of the T to protect against lateral doors while the lead addresses the forward threat. During casualty movement, place the shield over the casualty’s chest and face, then two rescuers lift and drag from the shoulders with short steps while a third walks backward, keeping the shield square.

Who benefits most depends on mission and training time. Patrol officers get a flexible piece of cover that turns many unknown-risk calls into manageable problems. Tactical teams use the shield to set the pace, control angles, and cross open danger areas with a safety buffer. Protective details and armed security can exploit buckler-style shields held at arm’s length. A well-designed buckler — such as our NovaSteel Buckler — protects far more than its diameter suggests, including the vulnerable areas of the throat and face that soft vests leave exposed, thus a 12.2 inch ballistic buckler held forward can cover disproportionately large areas compared with chest plates.

Inside the home, the geometry is unforgiving and distances are short. A shield makes sense when you must move through the structure, for example from a bedroom to collect a child, along a hallway to a safe room, or to manage a forced entry at a door. For most private citizens, the realistic sweet spot is a compact Level II or IIIA shield. These stop common handgun threats at weights that can be managed quickly through doorways and around furniture. True rifle-rated shields exist but the mass and bulk slow you at the worst time. Compact rectangles around 18 by 24 inches work well, and the buckler concept is compelling in tight quarters because it protects head and neck when held forward, can be used as a compliance tool or bludgeon, and is extremely easy to maneuver and reposition. Skip anything with a viewport unless you have a specific reason. Sighting is straightforward with a red-dot or laser on a handgun, and brief peeks from behind a solid edge often beat pushing a heavy window around corners. Stage the shield upright near your primary egress with an attached light, electronic hearing protection, and a phone. Run one-minute home drills that include opening and closing doors while maintaining coverage, and then confirm mechanics at a range that allows shield work. Unless you’re running a buckler, treat the shield as a sensitive article of armor, not a battering ram, and inspect it after any abuse.

Technology trends are favorable. UHMWPE developments already yield Level III plates around 1.8 pounds in a 10×12″ format, which implies rifle-ball shields trending toward today’s handgun-shield weights once designs are optimized. New ceramics and amorphous materials are pushing steel-core rifle performance down the weight curve. Optics are shifting from heavy viewports to camera-assisted views. Geometry is converging as shaping and molding improve, with curved profiles, purposeful cutouts, interlocking edges for ad hoc barriers, and the return of buckler-style shields sized for modern threats.

The practical bottom line is unromantic. Choose the lightest shield that matches the likely threat, because the tool you can carry fast and control under stress is the tool you will actually use. For most patrol work that means a handgun-rated composite. For specialized teams, technique and mobility matter as much as square inches, and a slightly smaller rifle-rated shield may be warranted. For private citizens with a realistic plan for moving inside a house, a buckler used with skill and purpose turns risky movement into mostly-covered movement without downside.

The post Ballistic Shields: What They Are and How to Use Them appeared first on Adept.

Towards an accurate, back-of-napkin mathematical model of the whole ceramic armor plate II: Knockdown and uplift adjustment to reflect dynamic properties

admin — Mon, 14 Jul 2025 21:00:05 +0000

Towards an accurate, back-of-napkin mathematical model of the whole ceramic armor plate II: Knockdown and Uplift Adjustment to Reflect Dynamic Properties

As discussed in the previous post in this series, the baseline BE equation offers a good approximation of ceramic performance under most conditions, yet it assumes that the ceramic’s quasi-static mechanical properties carry over into the ballistic regime without drastic changes. This assumption generally holds true for common scenarios, including steel-cored AP threats impacting boron carbide, silicon carbide, or aluminum oxide at typical small-arms velocities. However, certain specific conditions can undermine that assumption.

Two key scenarios stand out for their common nature. The first involves boron carbide (B4C) and its stress-induced amorphization under very high-impact pressures, commonly encountered when facing tungsten-cored or tungsten carbide-cored penetrators at sufficiently high velocities. The second scenario involves any ceramic (not just B4C) whose hardness (H) is substantially lower than that of the incoming penetrator core. In both cases, the ceramic cannot maintain its effective hardness and compressive strength during the crucial dwell period, and may not be able to sufficiently wear the projectile’s penetrator via abrasion, resulting in deeper penetration than the simple BE equation would predict.

To handle these cases, we have to introduce a “knockdown” adjustment factor. This factor does not alter the fundamental form of the BE equation; what it does, instead, is create a template for the modification of mechanical property inputs. By adjusting hardness and compressive strength downward under certain conditions, we can then capture the real-world degradation in ballistic performance due to non-quasi-static effects.

B4C Amorphization

Boron carbide’s crystal structure can undergo a phase transformation (amorphization) under high shear stresses and pressures, especially when impacted by tungsten or tungsten carbide penetrators at very high velocities. This transformation dramatically reduces B4C’s local effective hardness and, consequently, its ability to erode the incoming projectile. Empirical studies demonstrate that, under these conditions, B4C’s effective compressive strength and hardness can fall significantly below their nominal quasi-static values.

We represent this empirically with a knockdown factor λ, selected based on ballistic test data. When conditions indicate that amorphization will occur (e.g., tungsten-cored threat at >900 m/s impact velocity), we reduce the affected terms accordingly:

H_amorph = H × (1 − λ)

CS_amorph = CS × (1 − λ)

For example, if testing shows a 30% drop in effective mechanical properties upon amorphization, we set λ = 0.3, resulting in H_amorph = 0.7H and CS_amorph = 0.7CS.

Hardness Mismatch with High-Hardness Penetrators

Even without undergoing a phase change like B4C’s amorphization, a ceramic that is substantially softer than the penetrator becomes less effective at blunting and eroding the projectile. For steel-cored AP projectiles (hardness ~750–800 HV), ceramics like SiC or B4C (≥2600 HV) are usually harder, so no knockdown is required. However, if the penetrator core hardness equals or exceeds that of the ceramic, the erosive mechanism diminishes significantly. This often occurs when Al2O3, AlN, MgAl2O4, and other oxide ceramics and glasses face tungsten carbide-cored threats.

In these scenarios, we apply a partial knockdown factor scaled to the hardness ratio Hp/H, where Hp is projectile hardness. Specifically, if Hp > H, define a hardness mismatch factor δ that increases as Hp/H grows. For instance, if Hp is 10% harder than the ceramic, we might reduce the ceramic’s effective hardness by approximately 20%. This relationship can be linear or defined using an experimentally tuned parameter. For example:

δ = α × (Hp/H − 1), where α is chosen based on empirical data.

Then we adjust:

H_adj = H / (1 + δ)

CS_adj = CS / (1 + δ)

Thus, as the penetrator overmatches the ceramic in hardness, we proportionally reduce the ceramic’s effective mechanical properties in the BE equation.

Strain-Rate-Enhanced Ductility in Nominally Brittle Ceramics

Not every deviation from quasi-static behavior negatively impacts performance. A subset of ceramics – including aluminum nitride (AlN), beryllium oxide (BeO), magnesia (MgO), and certain perovskites and MAX phases – exhibit measurable plasticity at ballistic strain rates once pressures exceed approximately 5–10 GPa. Under these conditions, their compressive strength notably increases due to pressure-hardening, and limited dislocation glide allows these ceramics to absorb additional work before fracture. Historical penetration tests have documented cases such as AlN surpassing both B4C and SiC performance when impact velocities exceed approximately 1.7 km/s. High-pressure mechanical data similarly indicate substantial increases in strength for BeO and AlN under dynamic loading conditions.

To capture this effect, we introduce an uplift factor ϕ that scales the quasi-static mechanical inputs upward for ceramics known to strengthen under relevant impact conditions:

CS_dyn = CS × (1 + ϕ)

H_dyn = H × (1 + ϕ)

When to apply ϕ: (1) When impact velocity exceeds ~1.2 km/s and estimated contact pressures surpass approximately 8 GPa. (2) The ceramic is among the known ductility-enhanced ceramics (high-purity AlN, BeO, MgO, SrTiO3, Ti3SiC2, etc.). (3) The penetrator does not significantly exceed ceramic hardness; otherwise, the hardness-mismatch knockdown factor δ remains dominant.

Typical magnitudes: Published dynamic compression and lateral-confinement tests report dynamic-to-static strength ratios between approximately 1.3 and 2.0 for ceramics such as AlN and BeO within 10–20 GPa pressure ranges. Conservatively setting ϕ = 0.4 (40% uplift) is reasonable for AlN impacted by steel-cored AP projectiles at approximately 1.6 km/s, while values up to ϕ = 0.8 may be appropriate for BeO at higher velocities. Where direct experimental data are unavailable, extrapolate uplift estimates from logarithmic fits of dynamic strength versus strain-rate. (For e.g., split-Hopkinson data.)

Interaction with λ and δ: The uplift factor ϕ should be applied after any hardness-mismatch reduction (δ) adjustments, but in place of amorphization adjustments (λ), as ductility uplift does not apply to ceramics experiencing detrimental amorphization effects.

Example: High-Velocity AlN vs. Steel-Cored AP

Consider a sintered AlN tile (density 3.26 g/cm³, CS = 3 GPa, hardness = 1400 HV) impacted by tungsten alloy penetrators at two velocities:

At 900 m/s (moderate velocity), we set ϕ = 0:

CS_dyn = 3.0 GPa, H_dyn = 1400 HV.

At 1650 m/s (high velocity), we set ϕ = 0.4:

CS_dyn = 4.2 GPa, H_dyn = 1960 HV.

Inserting these adjusted properties into the BE equation predicts roughly a 12–15% improvement in ballistic effectiveness, aligning closely with published historical depth-of-penetration data, where AlN performance surpasses B4C at these elevated velocities.

Velocity and Threat Type Dependencies

Knockdown and uplift adjustments ideally scale with specific impact conditions. At lower velocities, B4C amorphization may not meaningfully occur, making λ negligible. As velocity rises and tungsten-based penetrators become relevant, both λ and δ should increase to reflect growing mismatches and the onset of amorphization. Conversely, uplift factor ϕ becomes significant only when impact pressures and velocities are high enough to activate ceramic ductility. Experimental ballistic test data must guide the selection of thresholds and scaling parameters.

Adjusted BE Calculation

After applying knockdown (λ, δ) or uplift (ϕ) adjustments to hardness and compressive strength, the BE formula remains unchanged structurally. For example, for B4C impacted by a tungsten-cored penetrator:

BE_knockdown = 1.677 × D + 5 × ((T × D)/D) + 0.003 × CS_amorph + 0.005 × H_amorph

For hardness mismatch in another ceramic, such as Al2O3:

BE_knockdown = 1.677 × D + 5 × ((T × D)/D) + 0.003 × CS_adj + 0.005 × H_adj

For strain-rate-enhanced ceramics (e.g., AlN at very high velocities):

BE_uplift = 1.677 × D + 5 × ((T × D)/D) + 0.003 × CS_dyn + 0.005 × H_dyn

These adjusted property inputs more accurately reflect real-world high-rate behaviors within the established BE framework. Just please note that all of this is intended as a simple toy model of ceramic performance in armor systems – a model that a human without a computer can handle with ease – and that, while it’s generally very accurate, it simplifies matters quite dramatically.

The post Towards an accurate, back-of-napkin mathematical model of the whole ceramic armor plate II: Knockdown and uplift adjustment to reflect dynamic properties appeared first on Adept.

Towards an accurate, back-of-napkin mathematical model of the whole ceramic armor plate I: The BE equation

admin — Wed, 02 Jul 2025 19:04:55 +0000

Towards an accurate, back-of-napkin mathematical model of the whole ceramic armor plate I: The BE equation

I recently published a paper on how it’s possible to accurately estimate a ceramic armor strike-face’s performance from a small set of mechanical properties that are easy to define. That paper, “a facile method for the estimation of ceramic performance in light armor systems,” is available here as an open-access paper: https://ceramics.onlinelibrary.wiley.com/doi/10.1002/ces2.10227

Now I realize that it could use some background, a short summary, and I can also devote a few words to how we’re extending the method.

Ceramic armor mechanical properties and performance

To simplify things a bit, until very recently nobody knew which mechanical properties were responsible for good performance in ceramic armor systems. There were many false starts, with various methods, like the D-Value equation, proposed and then eventually discarded.

This started changing about ten years ago, when Erik Carton’s group at TNO developed their Inertia Method for ceramic armor analysis. What they realized is that a ceramic strike-face’s performance is highly correlated with the volume of the conoid that forms upon impact.

In practice, assuming the cone side angle doesn’t vary widely between ceramic types, we’re already very familiar with the implications: At an equal weight, denser ceramics tend to perform

much worse than lighter ones, because the volumes of the fracture conoids they form upon impact are typically that much lower.

(Image courtesy: Carton E, Roebroeks GHJJ, Weerheijm J, Diederen A. Inertia as main working mechanism for ceramic based armour, Personal Armor Systems Symposium, International Personal Armor Committee. Washington DC. 2019. https://www.researchgate.net/publication/374925700_Role_of_inertia_in_armour_ceramics Downloaded 14 May 2024)

In other words, lower-density ceramics, being thicker at an equal weight, perform better because the fracture conoids they form are more massive.

At a glance, this result explained just about everything we knew of the performance gap between common armor ceramics such as boron carbide, silicon carbide, and aluminum oxide. Or so we thought.

Years later, in reviewing large libraries of ballistic data alongside our in-house R&D reports, it became apparent that the Inertia Method tended to overestimate the performance of boron carbide and, simultaneously, underestimate the performance of silicon carbide. The actual performance discrepancy between them, on a weight basis, was slightly but durably less than a pure conoid-based analysis predicts. For e.g., in Level IV plates, a B4C tile will perform roughly 17% better than a SiC tile on a weight basis, though the difference in conoid mass is on the order of 60%.

Needless to say, this called for further investigation. Fracture conoid side angles proved difficult to measure empirically, and predictive equations and models were equivocal at best. Besides, it’s not entirely clear that the projectile interacts very much with ceramic material on the margins of the conoid — a major failure mode for AP projectiles is erosion, and only ceramic material in direct contact with the projectile can contribute to erosion — so it didn’t seem likely to me that increased conoid volume necessarily translates to improved performance in a linear fashion.

So then the question became: “Is there a difference in mechanical properties that can explain why the relative performance gap is narrower than expected?” The two ceramic types were of similar, if not practically identical, hardness; the samples we tested ourselves measured at ~2600HV1 (SiC) and ~2650HV1 (B4C). Hardness was therefore dismissed as a significant factor. But, as it turned out, SiC had a tremendous advantage in compressive strength — a measure of resistance to bulk deformation that is in many respects similar to a hardness test, yet one that takes place on a larger scale and is more sensitive to grain size, defect density, and other large-scale and microstructural features. In fact, SiC’s compressive strength was consistently nearly 35% greater than boron carbide’s.

As a general rule, ceramics fail in compression because, under a uniaxial compressive load, they experience tensile stresses perpendicular to the applied force due to Poisson’s effect. These tensile stresses act on pre-existing microcracks within the ceramic, causing them to propagate and eventually link up to form a crush zone. Unlike tensile failure — where the largest, most favorably oriented flaw leads to sudden fracture — compressive failure in ceramics involves the stable growth and coalescence of many cracks.

It is therefore natural that compressive strength would be important to ballistic performance, for when a projectile impacts ceramic armor, the armor must withstand intense localized compressive stresses. A ceramic with higher compressive strength can resist the initiation and propagation of cracks more effectively, thereby absorbing more energy from the projectile and enhancing the armor’s ability to stop or erode the incoming threat, e.g. in a prolonged dwell period.

Preliminary models were built, and regression analysis was applied to see whether variance in compressive strength has explanatory power in the case of SiC vs. B4C. That indeed proved to be the case. But also, when looking across different ceramic types, hardness also stood out as a property strongly correlated with performance. Both compressive strength and hardness — two different but related properties — proved to be individually important factors. Other properties, such as fracture toughness, tensile strength, and Poisson’s ratio, had no correlation
with performance at all.

Further analysis led to the equation below.

BE = 1.677⋅D+5⋅((T⋅D)/D)+0.003⋅CS+0.005⋅H
Where:
D = Density in gm/cc
CS = Compressive strength in MPa
T = Thickness
H = Vickers hardness (HV1)
BE = Ballistic efficacy figure of merit, where a value of approximately 70 corresponds to the ability to defeat the .30-06 M2 AP in a typical light body armor or vehicular armor system and a value of approximately 100 corresponds to the ability to defeat .50 BMG AP in a light armor system.

The BE equation proved to have tremendous explanatory power that held across all ceramic types at thickness ranges typical of body armor.

Extending the method: Other materials

Interestingly, the BE equation also appears to work when applied to certain other brittle materials. Consider ultra-high hardness steels.

Maraging C350, given: D=8.1, H=700, CS=2675, T=9
BE = 1.677⋅8.1+5⋅((9⋅8.1)/8.1)+0.003⋅2675+0.005⋅700
BE ≈ 70.1

The BE equation predicts that a 9mm thickness of an ultra-hard high-compressive-strength maraging steel, in fully-hardened condition and over a backing layer, will suffice to defeat 7.62mm AP projectiles — if only barely. This is highly credible in light of the fact that a 12mm thickness of that steel will almost certainly defeat that same AP threat without a backer, and in light of a large body of existing data that compares vehicular ceramic armor with steel armor in terms of ballistic efficacy. Though it requires further experimental validation, the BE equation provides credible values for high-hardness steel in steel-composite armor systems, and may be used to optimize steel alloys for target mechanical properties — namely low density, high compressive strength, and high hardness.

For example: Adding 3% Si to a steel alloy significantly reduces its density, and can simultaneously improve its compressive strength and hardness.

All of the above also applies to glass and glass-ceramics. Though generally far softer than ceramics, they fail in compression in much the same way, and the same properties that influence ceramic armor performance also influence glass and glass-ceramic armor performance. So, as above, it makes sense to optimize them for low density, high compressive strength, and high hardness.

Note only that it is considered axiomatic that the armor strike-face should be as hard or harder than the penetrator core, which, in steel-cored AP rounds, is typically 59-63 HRC. Therefore the formula might not be reliable for steels, glasses, or other materials at hardnesses under 60 HRC or 700 HV1. Low-hardness RHA and similar grades of steel would fail via a ductile plugging mode that is not contemplated by the BE formula, and low-hardness glasses or tool steels might not be capable of eroding the threat projectile’s core.

The post Towards an accurate, back-of-napkin mathematical model of the whole ceramic armor plate I: The BE equation appeared first on Adept.

The Breastplate Through History – Characteristics From the 14th Century Through the Present Day

Adept Armor — Wed, 07 May 2025 12:53:43 +0000

The Breastplate Through History – Characteristics From the 14th Century Through the Present Day

1. Breastplate thickness from the 14th through 17th centuries
In the 14th century, just as plate armor was becoming commonplace, all of it was rather thin, with most pieces, including the breastplates at Churburg, under 2mm in average thickness.

As time went on, plate armor gradually got thicker to contend with heavier crossbows and early firearms. “The Knight and the Blast Furnace,” which surveyed the thicknesses of breastplates at the Royal Armouries, the Wallace Collection, and the Landeszeughaus in Graz, paints an interesting picture.

With “+” denoting armor for horsemen and “△” denoting armor for infantry, a few trends are readily apparent:

(1) Cavalry breastplates grew much thicker with time, starting from a roughly 1.9mm average – with substantial variance – around the year 1500 and increasing to about 4.5-5 mm on average, with exceptional pieces over 7 mm.

(2) Breastplates for infantry also grew thicker, to a roughly 3mm average by 1550 – but then actually thinned out a little bit, and by the 17th century they stabilized in a range from 1.8mm to 3mm in thickness: Effectively 2.4±0.6mm.

Now we can build our own, second scatterplot with additional data from the Allen study collection, a Slovenian survey of armor plates, and other sources,:

This fully corroborates the earlier picture – that, though there’s an overall trend towards thickening, there are many examples of plates from 1580 and 1590 that are even thinner than breastplates from 1480 – which implies that there are essentially two trendlines: Breastplates for cavalry got much thicker, whereas breastplates for infantry hardly changed or even grew thinner.

Now there are a few caveats to keep in mind. First is that there might be more than a few tourney or parade breastplates hiding in both datasets, though I don’t believe this to be the case. Second is that corrosion and cleaning/polishing both affect a steel or iron armor plate’s thickness – with corrosion thickening plates as the steel expands, and polishing thinning plates as material is removed – so we must assume that even the most meticulously drawn average has an error range of roughly ±0.25mm. Third is that all breastplates were of uneven thickness, and they all varied to some extent.

On that note, it’s virtually always the case that the breastplate and helmet were the thickest parts of plate armor. The edges and faulds of a breastplate were often much thinner, and arm and leg harnesses were generally thinnest of all, often under 1mm thick. The thickness range for breastplates is 1.25mm to 7.8mm in all of the samples measured above – but the average gauntlet or fauld thickness might be well under 1mm.

Breastplate areal density

There’s more information on the weights of breastplates than there is on their thickness, thickness being much more difficult to accurately measure. Yet weight itself doesn’t tell us very much, as breastplates were made in a very wide variety of sizes and cuts, and some were of elaborate peascod designs. We might benefit here from a degree of abstraction.

Low-carbon steel has a density of 7.85 gm/cc and iron has a density of 7.87 gm/cc – so close that they need not be considered separately. Let’s take 7.86 gm/cc as an average value.

With that, we can construct a useful table of areal densities:

Thickness (mm)	Areal Density (kg/sqm)	Areal Density (lb/sqft)
1.0	7.86	1.611
2.0	15.72	3.223
2.5	19.65	4.028
3.0	23.58	4.834
3.5	27.51	5.639
4.0	31.44	6.446
5.0	39.30	8.057
6.0	47.16	9.669
7.0	55.02	11.280
8.0	62.88	12.892

The average 14th-17th century breastplate has always covered roughly two square feet, or slightly under two square feet, in area. (Though, with few exceptions, they were made for shorter men than this era’s. A 16th century design to fit a 6’ man would be somewhere over two square feet.) For infantry, 10 pounds was considered the maximum usable weight, and this corresponds well to a thickness of just over 3mm. A typical infantry breastplate would be around 8 pounds at a central average thickness of 2.5mm. The heaviest musket-proof cavalry breastplate would have weighed over 22 pounds, and at that weight would have been a real burden.

III. Breastplates, not backplates

In all of the above, we’ve only considered the weight and thickness of individual breastplates. Backplates, throughout most of history, were very different: Thinner, lighter, and in some cases – especially in light cavalry and in the Austrian cuirassiers – not worn at all. In the Allen collection, backplates appear to be generally around 50-65% the thickness and weight of breastplates, which seems like a pretty fair generalization. It was not uncommon for an 8-pound breastplate to be paired with a 5-pound backplate.

In modern conflict soldiers are much more likely to be shot from the front – i.e. the bullet approaches from the front and impacts over the front of the body – and this was likely even more the case in the open-field conflicts of the 16th and 17th centuries. As such, it would have been perfectly rational to wear a thicker breastplate and a thinner backplate – or no backplate at all. (This applies particularly to mounted soldiers. In infantry, it can be exceedingly uncomfortable to wear a breastplate, or a modern 10×12” armor plate, without a backplate. Modern armor carriers and plate carriers generally match plates for this reason.)

The Cuirassier Breastplate

Though plate armor declined in popularity through the 17th century and was generally viewed as archaic and obsolete by the mid-18th, it did survive in the breastplates of heavy cavalrymen in Europe – the cuirassiers of France and their counterparts in other armies. These breastplates were still in service at the outbreak of WWI. Most 19th and early 20th century examples followed the 17th century tradition, with breastplates ~5-6mm thick, and much thinner backplates or no backplates at all. What’s important to highlight is that these were never, as a general rule, 1mm-thick costume pieces – they were always functional, and arguably far thicker than they needed to be.

Today’s ceremonial cuirassier breastplates are generally antiques, or very closely modeled after antiques. As far as can be ascertained, the final French cuirass is the so-called “3e republique” model which was produced from roughly 1870 until just after the outbreak of WWI – and they have not been modified since.

The Italian Corazierri breastplate is much the same – it hasn’t been substantially modified since 1900, though there is a dedicated facility responsible for the cleaning and maintenance of those fine nickel-plated antique breastplates. Though they’re quite functional and not exactly paper-thin, both the French and Italian models are used only for parade, ceremonial, and honor guard functions.

So the old European tradition of armor survives to this day, however in vestigial and much-reduced form. Still, it’s nice, isn’t it?

Steel Body Armor in WWI

The steel body armor of WWI owes nothing to heavy cavalry cuirasses. It followed the introduction of steel helmets over the course of that war, and was exhaustively chronicled by the Metropolitan Museum of Art’s arms and armor curator, Bashford Dean, in his 1922 book “Helmets and Body Armor in Modern Warfare.”

To briefly summarize:

France was the first nation to issue steel helmets, and General Adrian – the man most responsible for those helmets, and after whom they are named – also recognized the high mortality abdominal wounds among soldiers, which prompted him to develop an abdominal armor plate. In those days before antibiotics, abdominal wounds were apparently more fatal than chest wounds!

The abdomen plate, weighing two pounds, was held by a belt, though it was less favored by soldiers compared to the helmet. An attempt to add hip and groin guards was abandoned due to impracticality, but about 100,000 abdominal plates were produced and fielded. Adrian also tested a five-and-a-half-pound breastplate and gorget that mounted to the abdominal plate, though the weight of this assembly prevented its widespread use.

Later in the war, with both the abdominal plate and the breastplate abandoned, he produced steel epaulets – like scales – which were inserted into pockets in the uniforms of soldiers. These were generally made from helmet trimmings, and provided some small measure of protection from shrapnel.

As with their helmets, the French body armor systems were generally very thin – on the order of 1mm or less – and were made from their low-carbon helmet steel.

The British experimented along different lines. They produced a number of soft or “yielding” body armor systems from silk, linen, and other fabrics. They also experimented with steel plates in fabric carriers – remarkably similar to modern steel plates in plate carriers.

An example below is the “Portobank armoured waistcoat.”

An even more strikingly modern-seeming design is the 1917-1918 British Expeditionary Forces armor system, one piece of which is on display at the Royal Armouries in Leeds.

Neither of the above designs was popular or issued in large numbers. Both were replaced by a 1917 design developed by the The British Munitions Inventions Board. This, called the “E.O.B. corselet,” weighing 9.5 pounds, with a breastplate, backplate, and abdomen/groin plate. (Note: Those were Bashford Dean’s own measurements of one article of armor and there might have been some variance.) Like the Brodie helmet, it was made from 1.2mm-thick Hadfield manganese steel. Unlike the helmet, the plates came in a khaki carrier that was heavily padded for comfort. It provided reasonable protection from shrapnel and grenade fragments, and the E.O.B. corselet was issued “in pretty large quantities” to select units until the closing days of the war.

The most iconic body armor system of the First World War, the Germans issued a heavy, multi-plate body protector that could be worn on either the front or back, consisting of four plates with three abdominal plates that hung freely below a larger chest plate. The chest plate was 18.5 inches high and shaped to protect the chest and throat, with two shoulder plates (9 by 4.5 inches) to help support the armor. The abdominal section included three descending plates, each progressively smaller and attached by webbing with felt pads to reduce noise. Two sizes were available: a smaller version weighing 19-22 pounds with plates of 2.7mm to 3.4mm in thickness and a larger version, 31.5 inches high and weighing about 24 pounds with very slightly thinner plates (roughly 0.1mm thinner on average).

The German “sentinel” plate system was made of two different grades of steel. Type one was silicon-nickel steel 0.2C-1Mn-4.12Ni-2Si. The second, doubtless on account of nickel’s high cost, was a leaner and higher-carbon alloy 0.38C-0.6Mn-1.55Ni-1.75Si-0.2Cr. The first grade of steel has no modern equivalent; the second resembles modern AISI 3140, though it contains much more silicon.

These “sentinel” breastplates were deemed unsuitable for maneuver and regular infantry operations, so they were not kept in service after the war’s end. Further, not that many were issued: A report in Dean’s book indicates that they were issued in 1917 at a rate of two per company, including machine gun companies. A WWI German company consisted of anywhere from 160-260 men, for an issue rate of approximately one per hundred line troops. Thus they were one-size-fits-all and passed around between men on static guard duty. They appear to have been issued in larger numbers in 1918, but were still a relative rarity.For all this, the German breastplate almost certainly inspired the most common and widely-issued body armor system of both world wars: The Soviet Steel Bib.

The Soviet Steel Bib

There’s a tremendous amount of information in Warspot.ru’s series which begins with this article:

Стальная броня для красноармейца: рождение | Warspot.ru

For pictures and much more detail, go to them. For an extremely short summary, read on.

From the early 1930s, the Red Army (RKKA) worked on armor development programs. Their first project was steel helmet design, which ran concurrently with research into steel alloys that combined good ballistic resistance with ductility. By late 1935, they had created the SSH-36 steel helmet. In 1937, engineer I.M. Veynblat proposed a follow-up project: An “Armor Breastplate” designed to protect soldiers’ chests and abdomens from 7.62mm bullets (the type wasn’t specified, but likely 7.62x25mm), intended for assault units, motorized infantry, and cavalry. Early prototypes were made and successfully tested.

Despite initial support, production of Veynblat’s breastplate was delayed due to leadership changes and his eventual arrest by the NKVD. In October 1938, People’s Commissar Lazar Kaganovich ordered the Lysva Metallurgical Plant (LMZ) to develop and produce an experimental batch of steel breastplates based on Veynblat’s design, leading to the creation of the CH-38—the first serially produced steel breastplate in the USSR. By January 1939, LMZ had manufactured 491 CH-38 breastplates in various configurations, differing in weight, thickness, and design. There were basically two types: A “heavy” variant that was ~3.5mm thick and weighed just over thirteen pounds, and a “light” variant from 1.25 to 1.6mm in thickness, weighing around 7.25 pounds.

These breastplates underwent extensive ballistic testing and field trials, which confirmed their bullet resistance and practicality, although some issues were noted, such as discomfort, interference with aiming, and the need for design adjustments. Soldiers found, overall, that the breastplates offered significant protection and did not greatly impede mobility, and suggested only minor modifications for comfort and enhanced functionality.

Despite positive results, and internal government reports noting that “a steel breastplate in the next war will save the lives of many soldiers, officers, and political workers”, mass production of the CH-38 was not initiated. Work on steel breastplates stalled until August 1939.

Subsequently, and until the end of the war, a variety of breastplates were developed, though all were quite similar in appearance to the CH-38. The CH-40 was produced in two variants: A “light” variety at a roughly 3.5mm thickness and a “heavy” variety at 4.4±0.4mm in thickness. The “light” weighed roughly 8.7 pounds (3.95kg); the “heavy” 12.8 pounds (5.8kg).

Both variants of the CH-40 were criticized for excessive weight and poor ergonomics – and, at the same time, military planners criticized their poor performance against rifle threats, as they were readily penetrated by rifle and machine gun bullets within 350 yards. This performance issue was addressed in the CH-40A, which came in two new varieties: A “light” at 4.2mm thickness and a weight of 13.6 pounds (Size M), and a “heavy” at 5.2mm and 15.8 pounds. The latter could resist the 7.62x54mmR Model 1908 (ball) bullet at 100 yards; against the 7.62x54mmR B30 (AP) it required a standoff of 320 yards and an impact angle of 20°, which is still reasonably good performance under the circumstances.

These breastplates were stuck in the bureaucratic valley-of-death for a little while, and ultimately the decision was made not to issue them. They were deemed too heavy, and, according to the GAU, “one of the main types of small arms of all branches of the armed forces is the submachine gun” – which implies that the breastplates could be made much thinner and lighter, and still offer reasonable protection from the small arms threats of the time. So the CH-40A cleared the path for the CH-42, which was essentially the same design, but at an average thickness of 2mm (±0.2mm) and an average weight of 7.5 pounds.

The CH-42 breastplates were initially made of 36СГН helmet steel – nominally 0.36% C with roughly 1-1.5% silicon (“C”), 1-1.2% manganese (“Г”), and 1% nickel (“H”) as alloying constituents. They were hardened to roughly 45 HRC and were of a mostly-martensitic microstructure. They were resistant to MP-40 rounds at all distances, rifle ball rounds from ~300 yards, and grenade/shrapnel fragments

After initial testing but prior to fielding, the alloy type changed to 36СГ – removing nickel to keep production costs down. This does not appear to have affected the performance of the breastplates, for they remained in the same 2±0.2mm thickness range, and their ballistic rating does not appear to have changed.

Approximately 80,000 of these CH-42s were issued to certain front-line engineering and infantry troops, and they quickly became the war’s iconic piece of body armor – made famous by its use in close-quarters urban fighting.

Immediately after the war, feedback from soldiers informed the development of the “modernized” CH-42, which was thickened to 2.5-2.7mm, incorporated various design improvements to minimize noise and discomfort, and reverted to the 36СГН steel alloy. This design was made in small prototype quantities, but was never issued.

Through 1946 and 1947, the modernized CH-42 heavily informed the development of the CH-46, which incorporated a few small changes to the shape of the lower plates, but was otherwise identical to the modernized CH-42. Like its immediate predecessor, it was made in small prototype quantities, and was never issued. It’s said that military planners felt that the development of intermediate cartridges for CQB obviated the need for a steel breastplate, and with the shooting war over, there was no urgency and less available funding for the development of personal protective gear.

VII. The Novasteel Breastplate

It took almost a century, but by late 2023 it was time to revive the concept. Two things came together to make it viable: (1) The NovaSteel alloy produces superior results at a lower thickness than any 20th century breastplate or steel helmet, which makes for a lighter breastplate. It’s 2.3mm thick – in keeping with historical breastplates for infantry – but beyond Level IIIA in terms of ballistic resistance, and has class-leading knife and spike resistance. (2) Lightweight UHMWPE panels which leverage the performance characteristics of the breastplate enable it to be up-armored to defeat rifle threats, at a negligible weight penalty; the breastplate plus up-armor is in many cases lighter than a carrier plus armor plates, at the same level of protection. Future advances in up-armoring the breastplate will make it a truly universal armor system – a real alternative to textile carriers – and a spiritual successor to the steel bib and the breastplates of still older times.

The post The Breastplate Through History – Characteristics From the 14th Century Through the Present Day appeared first on Adept.

What Are Impact Foams?

Adept Armor — Thu, 27 Mar 2025 07:48:43 +0000

Impact Foams: Why they’re so important in body armor and how they work

Summary of Key Points:

Foams play a critical, underappreciated role in modern body armor, serving three vital functions:

Mitigating Deformation: They absorb and dissipate energy from ballistic impacts, reducing trauma to the wearer and protecting brittle materials like ceramics.
Structural Protection: Foam layers shield armor components (e.g., ceramic strike faces) from shocks, drops, and environmental damage.
Comfort and Wearability: By distributing forces over time and area, foams reduce pressure points and improve mobility.

Mechanisms of Action:

Absorption: Foams convert kinetic energy into heat/strain via cell collapse, minimizing energy transfer to the body.

Dissipation: They spread impacts over larger areas and prolong collision duration, lowering peak force.

Elastic Rebound: Resilient foams avoid “bottoming out” by partially rebounding, ensuring gradual deceleration.

Material Limitations and Innovations:

Traditional foams (EPS, EPP, shear-stiffening variants) often lack balance—sacrificing rebound, temperature resilience, or energy absorption.

IvoryGuard exemplifies next-gen foam design: closed-cell structure, ultra-lightweight (~0.068 gm/cc), and temperature-stable performance. It optimizes all three mechanisms while maintaining flexibility across environmental conditions.

Why This Matters:

Foam is often the largest component by volume in armor systems. Advancements in foam technology directly enhance protection, comfort, and weight savings.

Foams in Body Armor

When people talk about contemporary armor materials, the focus tends to be on composites or ceramics. Those are, after all, the strong materials that actually stop bullets in today’s most advanced armor systems. But body armor – armor that people wear against their bodies – is about more than just stopping bullets. It’s also about mitigating armor deformation upon ballistic impact, about protecting ceramic tiles that have poor shock and drop resistance, and about comfort. For these reasons, the average armor plate or soft armor panel is often as much as half foam by volume.

Lightweight level III/RF1 UHMWPE plates can be more than half foam. UHMWPE tends to deform a lot, and that deformation usually needs to be mitigated for compliance with ballistic standards or specifications. So if, for instance, an armor plate is 1.1” thick, roughly 0.6” of that thickness could very well consist of foam padding.

Heavier ceramic armor plates, built to counter tougher threats, are usually made up of roughly a half-dozen different materials. From front to back:

Front cover
Foam (ceramic strike-face protection)
The ceramic itself
A backing layer (usually UHMWPE in a resin matrix)
Foam again (backface deformation mitigation)
Rear cover.

And there are adhesive layers between them all, some of which can be nearly 1mm thick.

Of these, on a volume basis, foam is usually the largest single material component. There’s almost always more foam than ceramic.

Performance improvements in foam materials therefore translate directly to performance improvements in armor plates and panels. How foams work, and how to best utilize them, is of paramount importance in body armor design. (Though, needless to say, foams are far less important in other forms of armor design; they don’t matter much if you’re building an armored vehicle.) It was also the case that foams were the “low hanging fruit” of body armor system design – there was a lot of room for optimization. So we’ve worked to optimize them. To understand how, let’s review how they work.

Protective Foam Mechanisms of Action

Protective foams in armor work by a combination of key physical mechanisms that cushion and manage impact forces. The three primary mechanisms are absorption, dissipation, and elastic rebound.

Absorption: When an impact occurs, foams absorb energy by deforming internally. The foam’s cell structures compress and collapse, converting kinetic energy into other forms (primarily internal strain and heat) instead of letting it transmit to the body behind the armor.

In effect, the foam soaks up impact energy by crushing cell walls, stretching polymer chains, and generating a bit of heat, rather than transferring that energy to your body. A foam’s ability to deform plastically under a heavy impact ensures a large portion of the impact energy is absorbed and not returned. Essentially, the more energy the foam absorbs internally, the less reaches you.

In closed-cell foams specifically – such as IvoryGuard – the sealed air bubbles compress, increasing internal pressure, converting impact energy into stored gas pressure and additional heat. This further enhances the foam’s energy-absorbing capability.

Dissipation: Foams also dissipate and distribute the force over a longer time and a larger area. When a foam is compressed, it slows down the impact, stretching out the collision duration. By prolonging the impact event, the foam lowers the peak force. (Since the same change in momentum over a longer time means less force at any instant.) At the same time, foams spread the load over a larger area by deforming and contacting more surface. A soft layer will distribute a concentrated blow across a broader region, reducing pressure on any single point. The combined effect is that the force is diluted – lower pressure and lower peak acceleration – which greatly reduces injury risk.

A well-designed foam pad behind a hard armor plate will deform and stretch laterally, spreading the force of a bullet impact over the whole pad area rather than a small spot on the body. IvoryGuard foam is engineered to excel at this kind of dissipation. Its closed-cell structure and high ductility allow it to compress predictably without cracking, so it can prolong and spread out an impact even at high strain rates. Notably, because IvoryGuard is extremely light (density ~0.068 gm/cc, even lighter than most commercial forms of silica aerogel), one can use a thicker layer of it without adding much weight. Thicker foam means more distance and time to cushion an impact, which further improves dissipation.

Elastic Rebound: Unlike sand or clay which just deform and stay deformed, polymer foams often have elasticity – meaning they spring back to shape. This elastic rebound is the foam returning some energy after absorbing the rest. When an impact compresses the foam, the material stores a portion of the energy as elastic stress, like a compressed spring. After the impact, the foam expands back, releasing that stored energy by pushing the impacting object away in a controlled bounce. A small amount of rebound is useful: It means the foam didn’t completely collapse – it had some “spring” left to give – and it helps separate the impacting object from the protected surface, finishing the impact with a gentle push-off rather than a hard stop. Elastic rebound also allows the foam to recover its shape for the next impact. However, too much rebound, as you’d get with solid rubber, would return a lot of energy to the impactor, potentially causing a ricochet or a second hit. The key is that some energy return is desirable – it indicates the foam is resilient and avoided bottoming out – but it must be combined with absorption. Many impact foams are thus viscoelastic, meaning part springy (elastic) and part lossy (viscous damping). On impact, they absorb most energy as heat but return a fraction as a bounce.

Balancing these mechanisms is crucial for optimal protection. If a foam only absorbed energy but had no rebound at all, it might “bottom out” on a severe impact – compressing completely such that the foam pad eventually becomes a hard unpadded surface. Bottoming out causes a sudden rise in force, as any remaining impact energy transmits directly, negating the benefits of the foam. Conversely, if a foam was very elastic and barely absorbed energy, it would simply spring the object back, leading to high peak forces and not enough cushioning. The best protective foams carefully tune absorption vs. rebound. They absorb as much energy as possible, to minimize what’s transmitted, while being just stiff and resilient enough to slow down the impact gradually and avoid bottoming out.

Most foams don’t make full use of all three mechanisms. EPS, though very light, doesn’t rebound at all – it crushes and stays crushed. EPP is slightly better in this regard, but it’s still permanently spent on impact, if it’s hit hard enough. Rubber foams, though very elastic, don’t absorb much energy internally. Shear-stiffening foams like D3O and Poron XRD excel at rapidly stiffening under high strain rates, absorbing significant energy internally through molecular rearrangements and increased internal friction. However, despite their elastic nature, these foams do not fully leverage elastic rebound, resulting in less-than-optimal performance.

There’s another problem with shear-stiffening foams, and it’s a big one: Their efficacy is closely tied to their glass-transition temperature (Tg). These foams work best when their polymer chains are near – but above – their Tg; at ambient conditions they’re soft and flexible, but under rapid impact, the high strain rate temporarily “freezes” molecular motion, yielding a stiff, glassy-like state that dissipates energy. This response is reversible and not due to a permanent shift in Tg. The problem is that at temperatures below their Tg – which can be as high as 0°C – these foams are stuck in the stiff phase and can be brittle and rigid, whereas at high temperatures they’re stuck in the soft phase and lose the ability to stiffen upon impact. In both cases performance is severely impaired.

Such challenges have inspired innovations in foam technology, with advanced formulations – such as those used in IvoryGuard – designed to optimize the balance of absorption, dissipation, and elastic rebound, over a far broader range of temperatures.

Future posts will go into more detail and will examine how this looks in practice.

The post What Are Impact Foams? appeared first on Adept.