M50 is only half a crash rating

There is no residential crash-rating standard. Nothing in a building code obliges a private driveway gate to stop a vehicle, and no agency certifies estate barriers as such. So when a rating appears in an estate security proposal, it has been borrowed from federal perimeter security, where the ratings mean something specific and were written for embassies, military access control points and critical infrastructure.

Borrowed is fine. The standard is good and the testing is real. The problem is that the code gets borrowed in half.

The rating has always been two parts

The Army's transportation engineering agency states the structure in one sentence, in a public bulletin on passive barrier systems:

The condition designation is the maximum kinetic energy the system can absorb for a given vehicle type, mass and velocity. The penetration ratings categorize the range of allowable penetration for the barrier system.

Two properties, measured in one test, reported as two values. The condition designation is the thing you have seen: a letter for the vehicle class and a number for the impact speed in miles per hour. M50 means the standard test truck at 50 mph. The penetration rating is the thing you have not seen, and it is the one that answers the question an owner actually has, which is not "did the barrier survive" but "where did the truck end up."

What P1, P2 and P3 mean in feet

The Federal Highway Administration published a primer on vehicle security barriers in December 2018 that reproduces the ASTM tables with permission. Its penetration table is short enough to quote in full. P1 is one metre or less, which it gives as 3.3 feet. P2 is 1.01 to 7.0 metres, given as 3.31 to 23.0 feet. P3 is 7.01 to 30 metres, given as 23.1 to 98.4 feet.

Read that again with a driveway in mind. A barrier certified at P3 has been tested, has passed, and is eligible for the Department of Defense's approved list, having let the truck travel up to ninety-eight feet past the barrier line.

The measurement is defined precisely, and the definition matters:

The penetration measurement rates the maximum dynamic distance, or the total distance traveled by the front edge of the cargo bed beyond the pre-impact, inside edge of the barrier.

Note three things in that sentence. It tracks the cargo bed, not the bumper, because in a vehicle-borne explosive the cargo is the threat and the cab is just the delivery mechanism. It is a dynamic maximum, the furthest point reached during the event, not where the wreck came to rest, because the FHWA explains that a vehicle may retract back toward the impact point afterward. And it is measured from the inside edge of the barrier, so the distance is measured into your property.

The FHWA also states the impact geometry that earns the best result:

The purpose of a VSB is to prevent vehicles from gaining unencumbered access beyond the barrier, with the highest performance measure earned when the dynamic penetration distance of less than 1 meter with the impact angle of 90 degrees is achieved.

Ninety degrees. Square on. That is the tested case.

The government's own list proves the point

If the pairing were academic, every listed product would be P1 and nobody would need to ask. It is not academic. The Army bulletin publishes a summary of the barrier types on the DoD Anti-Ram Vehicle Barrier List with the ratings each type carries, and the spread is the story:

Bollards appear at M30 through M50 with penetration ratings across P1 to P3. Cable barriers, M30 to M50, also span P1 to P3. Portable modular barriers are listed at M30 with P2. Portable bollards are listed at M30 with P3. Post and beam systems run M30 to M50 at P1 to P2. Inertial barriers are listed at M50 with P1.

So on one government list, "M30" covers a portable bollard whose certified penetration band tops out at 98.4 feet and a post-and-beam system rated at 3.3 feet. Same condition designation. A thirty-fold difference in the number that describes what happens to your driveway.

This is not a secret grievance of mine. It is the industry's own complaint. A barrier manufacturer, writing this February for its own customers, put it bluntly:

It is common to see a spec that says "M50 barrier" without stating the penetration level. That is incomplete.

When the people selling the product are telling specifiers that the shorthand is incomplete, and the shorthand is still what reaches the homeowner, the gap is not a knowledge problem. It is a formatting problem with consequences.

The old K ratings were a pair too

Plenty of estate security literature still quotes K4, K8 and K12. Those come from the State Department standard that preceded ASTM. The FHWA explains why they persist:

The DOS performance rating designation (K4, K8, and K12) is still used by manufacturers who had their systems certified before 2009.

The State Department stopped certifying systems in January 2009. The K numbers map cleanly onto the modern ones: the FHWA's comparison table gives K4 as the equivalent of M30P1, K8 as M40P1 and K12 as M50P1, all of them describing a 15,000 pound diesel truck whose cargo bed travelled no more than 3.3 feet past the barrier. Under the 2003 revision of that standard, the FHWA notes, a bed travelling farther than 3.3 feet simply got no designation at all.

But the K system was also a pair, and the Army bulletin preserves the other half, which has largely fallen out of the conversation. Under the State Department standard, kinetic energy carried a K label and penetration carried an L label, with L1 covering 20 to 50 feet, L2 covering 3 to 20 feet, and L3 covering less than 3 feet. The DoD list still shows both: reinforced fence, for instance, appears with K8 to K12 alongside L1 to L3.

I could not reconcile those two facts against the primary list itself. The USACE Protective Design Center server did not respond when I tried to open the barrier list directly, so I am relying on two federal documents that describe it rather than on the list. The most likely explanation is that the L1 and L2 entries come from testing under the earlier version of the State Department standard rather than the 2003 revision the FHWA describes, but that is my inference and I could not confirm it. Take the safe reading: a K number on its own does not tell you the penetration band either.

The vehicle class is the other half of the other half

The letter in front of the speed is a vehicle, and the vehicles are not close to each other. The FHWA table gives median test weights of 2,430 pounds for a small passenger car, 4,630 for a full-size sedan, about 5,000 for a pickup, 15,000 for the standard test truck that carries the M designation, and 60,000 for the heavy goods vehicle class. The Army bulletin lists the heavy class at 65,000 pounds.

The kinetic energies in the FHWA table make the gap concrete. M50 is listed at 1,250 foot-kips. H50, the heavy goods vehicle at the same 50 mph, is listed at 5,430 foot-kips. That is roughly four and a third times the energy, at the same speed, from a different truck.

So a gate advertised as M50 has been tested against a medium-duty truck. It has not been tested against a loaded tractor-trailer, and there is no dishonesty in that: the M class is the historical DoS test vehicle and it is the right benchmark for most threats. It is simply not the same claim as "stops a truck," which is what the marketing sentence lands as.

There is one more designation worth knowing about because it looks like a rating and behaves differently. The Army table includes U30, U40 and U50, where the vehicle weight is given as user-defined, described as any specific vehicle type per the end user's requirements. A U-rating tells you a test happened. It does not, by itself, tell you against what.

What the certification actually guarantees about the test

Two facts from the FHWA primer are worth keeping, because they cut in opposite directions and both are reassuring in their way.

The first is that this is a physical test, not a simulation. The FHWA is explicit: certification "is achieved through actual performance impact testing and is not provided based on modeling and simulation," at an accredited independent laboratory operating to ISO/IEC 17025. Somebody drove a real truck into a real barrier.

The second is about who owns the result. The FHWA notes that when industry funds the tests, the results belong to industry and are obtained from them. So the test report exists, and you are entitled to ask for it, but it is not a public record by default. The DoD list is the free public index, and its own disclaimer, as quoted in the Army bulletin, is narrower than people assume:

The list does not represent an overall endorsement of any product or design or address its operational suitability or maintainability.

Verification that a test was passed. Not a recommendation, and explicitly not a statement about whether the thing will keep working after five years in your climate.

The four questions

If a barrier, gate or bollard line is being proposed for a property, the complete rating has four parts, and any one of them missing makes the other three unreadable.

Which vehicle class, which impact speed, which penetration band, and which edition of the standard. On that last one: the tables quoted in this report come from ASTM F2656/F2656M-15 as reproduced by the FHWA, and the Army bulletin was written against F2656/F2656M-18a. The current active edition in ASTM's catalogue is F2656/F2656M-23, approved 14 November 2023. I did not open the current standard, which is sold rather than published, so I am not asserting that its penetration bands are unchanged. Ask which edition a certificate is written against and check the date.

Then ask the question the rating cannot answer, which is what sits inside the penetration band. A P2 barrier at the property line puts up to twenty-three feet of travel between the gate and whatever is behind it. If what is behind it at twenty-three feet is a lawn, the rating is generous. If it is a parked car, a guard position, a gas meter or a bedroom wall, the rating is a warning. The Army's own design guidance treats standoff as the variable that selection serves, which is the right way round: you decide what distance you need first, and the penetration band tells you whether a product delivers it.

One more thing from the Army bulletin that has nothing to do with ratings and everything to do with installations. For a contiguous passive barrier line, breaks are not permitted to exceed three feet in width for traffic approaching at ninety degrees, or four feet for traffic running parallel to the barrier. A perfect barrier with a four-foot gap beside it is a decoration. The gap, not the product, is usually where residential installations fail, and no rating on any datasheet will tell you about it.

Why this beat cares

I help design the AI security systems for a veteran-owned (SDVOSB) home-security company run by fellow veterans. I do not own that company and earn nothing from this link. Full policy here.

My first career was counter-IED and electronic warfare, and the vehicle-borne threat is where I learned to distrust single numbers. In that world nobody says a barrier stops a truck. They say what mass, at what speed, arriving at what angle, held at what distance from what you are protecting. Standoff is the whole discipline. The charge does not care how impressive your gate looked, only how many feet of air sat between the detonation and the thing that mattered.

The ASTM system encodes exactly that discipline, which is why it is worth reading carefully rather than quoting loosely. The penetration band is a standoff figure wearing a product code. Every time it gets dropped from a spec, an owner is buying a distance without being told what the distance is.

What I could not confirm

I did not open ASTM F2656/F2656M in any edition. The standard is sold, not published. Everything quoted here about vehicle classes, speeds, energies and penetration bands comes from two federal documents that reproduce those tables, one with ASTM's written permission, and from the ASTM catalogue entry for the designation and approval date. Two independent federal reproductions agreeing is a defensible substitute for the original, and the P-band numbers match exactly across both. It is not the same as reading the standard.

I could not reach the DoD Anti-Ram Vehicle Barrier List itself; the Protective Design Center server refused the connection. The barrier-type ratings and the list's disclaimer are as reported and quoted in the September 2018 Army bulletin. That bulletin is eight years old, the list is updated quarterly, and the products on it today are certainly not identical. I am using it for the structure of the rating system, not for a current product inventory.

The two documents disagree slightly on two figures. The FHWA table gives the pickup class a median weight of 5,070 pounds and the Army bulletin gives 5,000. The FHWA gives the heavy goods class a median of 60,000 pounds with a 65,000 pound test weight, while the Army table lists 65,000. I have reported both rather than picking one. Neither discrepancy affects any argument here. I also did not quote the FHWA's worked K4 example verbatim, because the metric conversion in that table cell contains an obvious typographical error.

I did not evaluate, test or inspect any product, manufacturer or installation, and nothing here recommends one. I have no engineering role in barrier design, and this report is not a substitute for a security engineer's assessment of a specific site. My published research is in microwave spectroscopy; the barrier test data is ASTM's, and the guidance is the FHWA's and the Army's.

The signal

A crash rating is a sentence with a subject and a verb. The vehicle class and speed tell you what hit the barrier. The penetration band tells you what happened next. Marketing prints the first half because it sounds like an achievement, and drops the second half because it sounds like an admission.

Both halves are published, both are free to look up, and the difference between the best and worst penetration band on certified products is roughly thirty to one in feet of your own property. Ask for the whole code. If the answer comes back as one number, the specification is not conservative or simplified. It is unfinished.