A friend of a friend of the blog FOIAd this information, which took him two requests. The first produced the Round One reliability testing, which downchecked 6 of the 10 submissions. During the first-round test and data analysis, the submissions were blinded by using a code letter ID from A through J. In the results below, the producers of the four proposals which advanced to Round Two are shown, while the downchecked candidates are still indicated only by code letter.
Ultimate winner — the HK M27 IAR, a version of the company’s HK 416 AR knock-off.
Results from Round One, in 2008:
e. All participants submitted three samples for testing.
f. Results for all Class I and II failures are listed below across all 3 UUTs from 9 of the 10 IAR Bid Samples.
- Colt proposal A: 60 Failures
- Colt proposal B: 28 Failures
- Competitor C: 23 Failures
- Competitor D: 78 Failures
- Competitor E: 39 Failures
- Competitor F: 12 Failures
- Heckler & Koch Defence Inc. proposal G: 27 Failures
- Competitor H: 124 Failures
- FN Herstal proposal J: 26 Failures
g. The 10th IAR Bid Sample, Competitor I, was determined unsafe for live fire due to a lack of proof marking. Live fire testing was not conducted.
Kind of a raw break for that unfortunate competitor, Code Letter I.
Note that some of the rejected proposals (C, E, F) had fewer failures than the ones that proceeded. This is presumably due to the distribution of the failures. (If you have only a few failures, but they’re take-the-gun-to-the-bench failures, that’s a whole other thing than a higher quantity of simple failures that are rectified in seconds by operator immediate action (a la SPORTS drill on the M16/M4 series).
This is the Round One definition of failures by class:
Reliability Testing
The Unit Under Test (UUT) shall have a Mean Rounds Between Failure (MRBF) of 900 for Class I and II failures combined (Threshold), 5,000 MRBF (Objective).
- Class I failure: A failure that may be immediately corrected by the operator within 10 seconds or less while following prescribed immediate action procedures.
- Class II failure: A failure that may be corrected by the operator, and that requires more than 10 seconds but not more than 10 minutes to correct (less the TM/OM defined cool down period if a hot barrel condition exists). Only the equipment and tools issued with the weapon may be used to correct the failure.
A very similar definition of failures, with a third, more serious, class, was used for Round Two in 2009.
Reliability/Endurance Testing Mean Rounds Between Failure (MRBF)
a. Three Units Under Test (UUTs) were provided for each model under evaluation.
b. The UUT shall have a Mean Rounds Between Failure (MRBF) of 900 for Class I and II failure combined (Threshold), 5,000 (Objective). The MRBF for Class III failures shall be 15,000 (Threshold), 20,000 (Objective).
- Class I failure: A failure that may be immediately corrected by the operator within 10 seconds or less while following prescribed immediate action procedures.
- Class II failure: A failure that may be corrected by the operator, and that requires more than 10 seconds but not more than 10 minutes to correct (less the TM/OM defined cool down period if a hot barrel condition exists). Only the equipment and tools issued with the weapon may be used to correct the failure.
- Class III failure: A failure of a severe nature. The failure (1) can be corrected by an operator but requires more than 10 minutes; (2) cannot be corrected by an operator and requires assistance (no time limit); or (3) requires higher level of maintenance or correction by an authorized operator cannot be accomplished because of unavailability of necessary tools, equipment, or parts.
This table is taken from the FOIA release, but we have added a column identifying the firearms, which in this test were coded 09 (presuably for the fiscal year) and a letter, thus 09A, 09B, etc.
|
UUT |
Manufacturer | MRBF Class I and II |
MRBF Class III |
Estimated Barrel Life |
|
09A |
Colt |
952 |
60,000 | 1800 |
| 09B | Colt | 1,277 | 15,000 |
1400 |
|
09C |
FNH USA | 5,000 | N/A* |
5,200 |
| 09D | HK USA | 1,622 | 20,000 |
16,200 |
Some interesting results here. The FN entrant had the highest rate of relatively minor Class I and II failures, but the lowest rate — zero — of Class III failures. (That’s why it’s “N/A”. You can’t calculate an MRBF with zero failures). And the HK example was distinctly mediocre compared to these competitors, on this one measurement. Conversely, it had far and away the highest barrel life — an important statistic for the always-broke Marines.
Update
Apologies to all for leaving off the document. This was actually two separate FOIA releases, of three and two pages, but I’ve combined them into one document and OCR’d them for your convenience (well, I also OCR’d them so I could pull those quotes above).
The initial page with the ID of the requestor has been deleted as he has requested privacy.
Kevin was a former Special Forces weapons man (MOS 18B, before the 18 series, 11B with Skill Qualification Indicator of S). His focus was on weapons: their history, effects and employment. He started WeaponsMan.com in 2011 and operated it until he passed away in 2017. His work is being preserved here at the request of his family.
42 thoughts on “USMC IAR Reliability Testing Results”
Wouldn’t the FN entrant have the LOWEST rate of Class I and II failures since the Mean Rounds Between Failure is 5,000 rounds?
I came to say what Ken said. 5000 MRBF is too round a number to be real, but if it is, FNH is the rifle I want.
No, Colt’s 09A barely passed the Threshold MRBF for Class I and II failures. It was the least reliable in that regard..
The FN entry was the only one that met the Objective MRBF for Class I and II failures.
It isn’t hard to imagine how a nice round number could result. Say for example, they fired 120,000 test rounds and experienced 24 Class I or II stoppages. That works out to 5,000 MRBF.
Upon rereading, I see it was 60,000 rds fired, so that works out 12 Class I and II stoppages.
Ah, you beat me to it.
also the barrel life numbers sound really odd. only 1800 rounds? My understanding was your average colt barrel life was 10k-12k rounds, obviously depending on frequency and rate of fire. i’m curious what the accuracy requirement was to determine barrel life.
Hours, maybe?
My first thought also. Less than 2K rounds barrel life for an infantry rifle? Seems bizarre…I have more ADs than that each year…no wait…just kidding!
Possible stupid question, but isn’t barrel life mostly a matter of the various barrel dimensions, metallurgy, and any treating/finishing?
That is, shouldn’t it be relatively independent of MRBF aside from issues like chamber dimensions and shape?
Thus, shouldn’t it be theoretically possible for FN to fix their barrel longevity (relatively) simply without affecting the reliability of the rest of the platform, putting them clearly in the lead?
Ah, my fine inherited-trait expert, test is probably run more by riflemen than by engineers. Did I include the .pdf of the results with the testing methodology, and the failure Class definitions, but not the evaluation rubric? I’ll check and if not, I’ll put it up here…
… and you will see that it does not answer your question.
Heh, I’ve always enjoyed this pseudonym, if only as a reminder that like the real man, I am -not- as smart as I sometimes think.
That said, thanks for the PDF. It may not answer the question, but it’s still interesting.
Shawn says there is one more document coming. When it comes it’ll show up here.
Barrel life is a function of several factors:
1. How “overbore” the cartridge is. Some “hot” rounds (like the 6.5-284, used by F-class and long range shooters), might have significant throat erosion in 1200 rounds or so. Rifles with very modest pressures and powder charges (eg, a .22LR) might have a barrel life over 50K rounds.
2. Wear is worst in the throat area, modulo Bubba see-sawing a brush back and forth through the bore. Properly treated, the area of concern will be the throat, just forward of the chamber neck.
3. Some coatings help increase barrel longevity – nitriding is the current hot lick. Chrome wash or plating also can add significant life, at the expense of ultimate accuracy.
I’ve talked to competitors shooting AR’s in tight competitions who tell me that they notice the typical stainless match AR barrel might last 5 to 7K rounds, at their level of precision/accuracy, which will be much better than a issue rifle/carbine. For an issue weapon, I’m sure that the accuracy/precision is acceptable to specifications longer than 5K rounds.
How do you properly clean a barrel then?
I’ll write up something for Hog to post, as it will be longer than what I want to write up in a comment. I’ll try to cover more than just black rifles, since few folks here have just an AR.
If you’d like, I can give you the “keys” to the blog and your own login, and you can draft whatever you like and I will edit as much or as little as you like and put it on the schedule. The content manager makes working with others very easy. The only one who’s taken me up on this before is Andrew Branca, the Law of Self Defense wizard.
Ok, so I was aware of those three, and figured that 1 is a constant, not a variable in terms of military procurement. If you need to sell them on a new 5.56 cartridge along with your new rifle, your bid is probably in big trouble.
2 is a training issue. I know that I now know a lot more about how to properly maintain an AR-pattern weapon than I did when I was in the Army, and was fed instructions and tips re: cleaning back then that I have since learned are absolutely horrible ideas, most of which have been commented on by our host in previous posts.
That leaves 3, which is why I’m sort of puzzled by those numbers. Did FN submit a SCAR variant with a non chrome-lined (or nitrided, or what-have-you) barrel to an automatic rifle competition just to drive down unit cost? I have to cop to not having read the full need statement that presumably covered things like -minimum- acceptable barrel life, but surely they would be aware that barrel life would be a big factor in which platform was selected, and would have planned accordingly?
Based on the round two testing it appears that estimated barrel life was the only metric that mattered.
Colt’s unit 09A outperformed the H&K 09D in every other way shown on the table.
No, Colt’s 09A barely passed the Threshold MRBF for Class I and II failures. It was the least reliable in that regard.. The only way it was superior was in terms of Class III stoppages.
Think about that. That’s 20,000 rounds per UUT, 3 UUTs.
That’s a difference of 27 Stoppages.
Personally 63 Stoppages per 60,000 rounds is incredibly minor and can be chocked up to bad magazines(since type of stoppage isn’t listed).
Personally I would rather have a gun with slightly high Class I/II Stoppages, and far fewer Class III Stoppages.
Because Class III Stoppages means my guns can’t be fixed by anyone below an armorer.
For the barrel life…Yeah I’m not touching that. Those numbers are all sorts of screwed up and make no sense.
That looks like some fancy pencil whipping if you ask me.
I wonder if the barrel life stat comes from a high temperature, high sustained rate of fire test.
D’Oh! I should have re-read the .pdf. There was no sustained rate testing in 2009.
While you can’t calculate a point estimate (mean) of MRBF with zero failures, you CAN statistically calculate confidence bounds around where it is, AND you can use the 50% confidence limit to estimate it as well. We do this frequently in the RAM Engineering world, and it’s always a pet peeve when I see “N/A”, “-“, a blank, or worst of all, a big fat goose egg in a table like this. What’s even worse is the tendency to just report the statistical mean with no context as to how confident you are in that mean.
>it’s always a pet peeve when I see “N/A”, “–“, a blank, or worst of all, a big fat goose egg
Not just your pet peeve. Implying that (worse, presenting as if) you have data when you don’t is WAY worse than simply admitting you have no data. Makes me crazy. Crazier. Making shit up is one of the few things over which I will absolutely lose it on someone.
>even worse is the tendency to just report the statistical mean with no context
True, yet this is probably a different category than making shit up; I tend to put lack of confidence intervals-type stuff down to genuine ignorance rather than laziness and ass-covering. It never ceases to amaze me how even basic statistical concepts elude otherwise very bright, capable and technically-trained people. Not that I am any statistical expert; I barely know enough to know how little I know.
Okay, color me confused. If you’re talking mean rounds between failure, wouldn’t you want to have the highest number possible? Or what am I missing here?
MRBF is not a rate. It is the inverse of a rate (failiures per round). High failure rates are bad and equal low MRBFs. For things like guns, think of “rounds” as a unit of time, and failures as a category of incidents that can be counted. The equivalent in a system like an aircraft would be Mean TIME Between Failure, and the failure rate would be Failures per time unit (typically hours).
Yes, confused me, too. Busy feeling with leakage testing of welds today, and that is pleasantly binary.
> lack of proof marking
Eh, wot? Since when do US military rifles need proof marks?
Though I imagine some people were pretty upset when that linered aluminum barrel AR-10 failed during testing…
There is a reason why Colt and FN barrels are marked “MP.”
I thought no US made weapons carried proof marks.
Without reading the specs the entrants were given I’m betting they had to have fired an HPT round and then have had an MP inspection.
it’s not some law they don’t HAVE TO. but quality makers will mark if they take the time and money to properly test
Where in the heck are they getting the “5,200” estimated barrel life for the FN entrant? I was under the understanding that FN barrels, especially MG barrels, are universally cold-hammer forged and regarded as some (if not the most…) of the best barrels under mass production.
Anybody else read that Battlefield Vegas synopsis regarding their MRBF/reliability metrics on FA guns rented on their range (http://www.calguns.net/calgunforum/archive/index.php/t-1098622.html)?
HK barrels are good. I get it. But something doesn’t make sense here…
Agreed. The HAMR basically used a 249 barrel.
You mean to tell me 249 barrels are toast in 5,000 rounds? Yeah not buying it.
Just at a glance from the outside, the IAR looks pretty good as a possible contender for the next generation of M4A-whatever-we’re-up-to; I’ll reserve mt judgement for until I get one in MY hands and put about 5000 rounds through it, using MY ammo- Probably Mk262.
Except for that pistol grip. It looks to be even worse than that of the GI M16A2 and M4, and I had not thought such a thing was possible. I’m working on a fix for my own Stoner/AR15 family rifles, still in the tinkering , and for good or evil, you’ll see it here first. It is not an original design, but an adaptation; I’m not that smart.
I confess, I must be a very special case of stupid here, so please bear with me.
What problem are we solving again? ie, what’s wrong with the M16/M4 that this weapon solves, or purports to solve?
In other words, as a taxpayer, what am I getting for my money?
HK gets to make gun in the US and get a big fat contract after they lost out over the XM8 and some senator gets a factory in his home state while a few retired Generals get a pie job shilling for HK. thats the problem being solved here
It’s the government bureaucracy so in terms of us the tax payers…nothing.
In terms of the senior officers, civilian government officials and company officials…lots and lots.
Keep your powder dry and your faith in God.
It’s also worth noting this is similar to what the Army saw in the ICC trial.
The M4A1 God slightly more Class I/II Stoppages than Gun C(the best entrant), however it had significantly less Class III Stoppages than Gun C.
You could have said you got the IAR stuff from me, after I got it from my friend. It would not have caused him grief or me any problems.
I think there is s still ome more info on the IAR tests coming from him soon.
Ummmm… think the recommendation might have been “FNH…with a better barrel”?