Nothing wrong with good old face to face communication from trusted sources. However, this does not conflict with the fact that most NON ENGINEERS/NON SCIENTISTS, especially in the firearm domain most often have very heavy biases. Heck engineers can too, but I was referrring more to the actual engineers working on firearms systems. Anecdotal evidence is not really evidence from my perspective. A detailed look at statistical failure rates and/or a detailed engineering analysis approach is required to truly understand the failure of such engineered systems.
The Importance and Relevance of Engineering
I agree 100%. I grew up the son of an uber-engineer who worked on some of the most advanced aerospace systems in DOD. This guy would come home from running detailed and extensive test procedures at one of the CTF's at the Air Force Flight Test Center, Germany, (and other bases) and get into what he really loves with his own research in physics. One of my uncles was a nuclear engineer, and my grandfather was a machinist for Douglas Aerospace, so my childhood, while I thought it was perfectly normal, was anything but.
I self-rate somewhere between a nuckle-dragger and an engineer, and am mostly comfortable in engineering discussions, whether we're talking statistical and elemental analyses, or fundamental aspects of applied physics. I'm not formally-trained, but was raised on an engineering/physics vocabulary during my formative years in a home education environment, in addition to a combination of military, private, and public elementary education. Shooting sessions with my dad involved discussions in physics, while visiting grandpa meant measuring cases with calipers, micrometers, and getting more of his copies of American Rifleman.
Failure Rate
So yeah, I agree that anecdotal evidence is just that. However, when an entire Platoon or Company of Rangers breaks 100% of their optics within a 1 week range package with the SCAR, that tells anyone from the janitor to Einstein that we have a problem.
Taking a Platoon of 42 guys, and all 42 sets of optics broke, when this didn't ever happen anywhere near that failure rate before with the M4A1, I think we can skip the large population, long-term T&E schedule and go back the pre-production RDT&E guns and apply some instrumentation to diagnose what exactly is causing this.
I suspect the knuckle-draggers didn't conjure up the harmonic bolt impact phenomenon theory terminology by looking it up in the rifle manual, and the claims that Eotech and Elcan had to re-engineer their systems to compensate for the unusually high-rate of failure with one particular weapon system do coincide with new models from both companies, namely the XPS and SpecterDR.
End-User Perspective
You have to keep in mind that the sources of these claims are units where untold millions of rounds are burned through M4A1's and other self-loaders on a shoot schedule that was never meant to be digested by the AR15 family of weapons, yet it has performed beyond exceptionally well since the early 1960's in these small communities. During a CQM, SFAUCC, SOT, or whatever the unit calls its high volume individual marksmanship regimens, it's common to burn at least 500 rounds through an M4A1 per day per shooter, often more. I've participated in sessions that have reached 1100rds of 5.56 through my M4 in a 4hr period, and didn't think much of it at the time because everything worked as I expected, and the gun had been beat-up for years.
Now imagine an entire operational element of no less than 42 shooters, running their CQM package like that with a new toy, and 100% of the optics fail. The only factor that changed in our equation is that you removed the M4A1, and substituted it with the SCAR. From a scientific standpoint, there is clearly merit to looking into the source of the problem more. And that's just one Platoon. If I understand correctly, the entire Company had these issues, which would take the number up to well over 100.
Time/g Graphs
As to the graphs that you posted from Hootiewho's analysis on M4carbine.net, those are a time & g-force graph of the cycle of operation comparison of a Colt Model 6933, which is an 11.5" CLGS carbine, and a 16" SCAR. Before you indicated, I
did guess that the top graph was for the AR15, and the bottom for the SCAR. Can you guess why I guessed that?
I looked at the "g" values of both, and saw that one of them exceeded +40g and -60g during the bolt closure time window, while the other did not. You'll notice that even the typically-violent 11.5" Carbine, with its much faster cyclic rate, still did not exceed the +40g/-60g threshold. And which firearm DID exceed those limits when the bolt closed? That would be the SCAR-16.
I think in attempting to prove your point, from an engineering standpoint, you added credence to what I have heard, and this is literally the first time I have seen those graphs. While 20g's over the threshold of the more violent 11.5" AR might not seem significantly more to the layman, I can assure you that -20g on top op of an existing -40g's is no joke when it comes to electro-optical aiming systems. To further quantify that we are spiking that g load within .004 seconds, you can see the problem, if you have even a Reader's Digest understanding of materials strengths.
Excellent graphs, as I really haven't seen them before. I also like how he pointed out what I feel is one of the main strong points of the SCAR (cam track geometry), but also could be the contributor to the increased g's on bolt closure. You will notice one flaw in the comparison to the operating parts weights, in that the analysis starts out with weighing the bolt carrier group of the AR15 without the buffer, and the complete reciprocating group of the SCAR, since the SCAR has no buffer. Notice what happens to the weight differences when you add any of the AR15 carbine buffers to the AR15 BCG when compared to the SCAR BCG/PISTON...the H buffer takes you .7 oz heavier than the SCAR's total reciprocating mass, the H2 buffer takes you 1.5oz heavier, and the H3 takes you 2.3oz heavier. He used the H3 in his 11.5" 6933 Colt, as indicated in the graphs.
Most people aren't aware, but the first prototype AR10A had a one-piece BCG/buffer, that slid forward out of the gun when disassembled, as did the upper receiver.
http://i13.photobucket.com/albums/a261/jimbo1227/AR10SMRJan1998Page1.jpg
Some Pertinent History
For some reason, Stoner went to a sectional design that separates the BCG from the buffer. When the AR10 was scaled down to the AR15 around the .222 Remington cartridge, a telescoping buffer assembly was eventually used, called the Edgewater, which was replaced several years later with the current buffer design after the Colt Model 602, when the XM16E1/early Model 603 was introduced. This current buffer design has reciprocating masses within an aluminum buffer body, which segments the rearward
and forward recoiling impulses to deal with dwell time and carrier bounce. What that means is that instead of having one large operating part slam back into the breach or extension, the impact is divided between the reduced weight of the carrier and the follow-on impact of the buffer weights.
Contrasting Designs w/ Recoil Mitigation
Another thing that I also would suspect as a culprit to this increased g spike on the SCAR's return stroke impact is that the operating parts travel in the SCAR is significantly longer than the AR15's. This increases the travel time of course, but also allows the mass to gain more momentum during the return stroke in the cycle of operation.
Great Discussion
I'm actually glad we had this discussion, because a lot of clarity has potentially emerged for me. I'm sure the engineers at FN are well-aware of this, and I suspect that the desire to have a common upper receiver for both the SCAR-L and SCAR-H created the excessive travel time of the SCAR-L's operating parts.