Even though you're pretty well informed unlike your usa collegues who don't give a .... You probably missed a point or two about s478.
I must admit, i really didnt know that retention bracket on s478 did only need to withstand 70C. It's extremly low point. But also i think brittleness of overcooked bracket shouldn't be an issue here.
Things were different and much better in K7 time (
they have same retention on the socket like s370) did need to withstand 95C and have two auxilliar notches on every side (
if crappy socket manufacturers like FOXCONN didn't go with cost reductions and have only rightmost auxilliar notch installed on sockets they produce). So in that times we could more or less rely on socket retentions which were developed to really withstand heat and
idiot-user that could multiple times fiddle around heatsink for cleaning-removal-exchange purposes. Ntm, that
socket 370 was mostly 35W and less (
not pure heat obviously).
In the later socket 462 (socket A) they did start with additional proposal of for holes around socket obviously for some purpose (Thunderbird,Palomino exhaustive heating) and Zalman vf7000 use that feature for mounting their additional retention system cause of
heavyweight cooler (HWC) champions they started to introduce into market.
Socket 478 (like s775 original pushpin retention also, but lets focus on s478) is nowhere near as good to hold heavy coolers without above mentioned mods as for vf7000 and similar. If you recall original s478 intel boxed stock coolers have
PLASTIC RETENTION SYSTEM. And if you wonder why, probably cause plastic doesnt conduct heat as well as metal (otherwise we cold have full featuring plastic coolers
and that was dirty cheap and good solution for s423/478 to have plastic/plastic bond. It also shouldnt grind down weaker plastic on sockets s370/s462 (notches for coller) during multiple removals. And it could easily retain that 250g alu coolers that was usual for the boxed cpu of that time.
On the other side, many aftermarket cooler suppliers used metal retentions (and metal conduct heat) and if intel really had that kind of crappy specs for their retention brackets to only withstand 70C it's in fact good enough even for metal retentions. Because if some of older you remember P4 CPU class first introduce
throttling to reduce overheating and that jump into action at pretty low points (60C and there you have worse CPU than P3
and if core reaches 80C outer cooler was nowhere near 80C maybe 60C so in fact even metal retentions couldnt fry plastic brackets if CPU was not burning at 95-100C and that was a throttle all around with 25% efficiency so in fact you have 800MHz CPU that was cooking @3.2GHz. (Well it's not the point also
)
The point is that for outer plastic bracket, to be burnt as you claim, at 70C (according to specs as you say) you need to have cpu core burning at around 100C, metal retentions for cooler and cooler itself at 70C or up.
In fact, even that dear customer of yours bittering about "damage made by himself", if brackets were according to specs. And i myself saw quite enough FoXCONN brackets with broken one eye (retention holder) on systems that no one even tried to install anything better than boxed intel's stock cooler i feel free to say that you overestimate 80% of s478 retention brackets lifespan.
Even for s604 Xeon class i dont think they developed it with 10yrs life expectancy in mind.
In fact overproducing FOXCONN even go as far to deliberately weaken bracket holders to mobo (that with s478 pushpins to secure) so if one of four retention holder (eyes) wont break, at least 1-3 holders break after 3-5yrs (and the newest s478 is from 2003) with (again) only stock cooler ever installed (250g aluminium).
And after all, you're in selling aftermarket enthusiastic coolers businnes that goes up to 1kg and you (Scythe customer support) believe that some poor and deliberately Low Quality (LQ) s478 retention bracket should hold that
.
I believe, even avfm8 didnt know how to speak out that, that you should for s478 go beyond poor intel specs,
which is easiest way for you to claim compatibility. And go with retentions that are enough for themselves (and enthusiastic users --- the busseniss youre in) without installed brackets (Asus just like fc, while amco and lotes were 300% better). Yep it would mean that you need retention design for obsolete? spocket 478. But few M6 screws, standoffs (spacers) and screwing (alternative) mechanism to install cooler onto s478 could be just as cheap as they're now.
And as far s775 goes pushpins are just too down beneath humongous cooler that could be easily installed as some "service management" predicts. And while they're good enough (without
idiot-user) for small boxed coolers cause they could be easily accessed during install/uninstallment, for oversized AM coolers they certainly are not good on long run as you say. You also forgot about W=F*s when you have huge HP cooler which heavily leans on one side by its weight. In those cases "intel's specification" pushpin retention just dont hold against threatning spillout. HWC will grind that pushpin holder plastic against harder mobo pcb holes especially if you have some direct vibrations (from malfunctioning fan, or higere speed rotating fan even 2000rpm past 92mm makes good enough vibrations). So in fact you should work on that s775/1366/1156 retentions also.
And my bitter sweet is s754/939/AM2/AM3 retainer on Kama Angle. While that notches reduction system could be good for many coolers on Angle you're always had one side of retention directly beneath attached cooler fins. (While other is free to manipulate). And if you install it on third notch (and you really need to work around removal from notch which is time consuming even with flat pliers, not hands) you cant be safe that cooler is on that heatsink covered side properly fasten to bracket and in some cases even 3+3 doesnt help to fasten cooler without some 1-1.5mm spacer beneath cooler base
btw. hi, hello and cheers inside a thread (even when CS is in the matter) is too much patronising and not just as nice template you people are obliged, by bully management, to use in your responses.