ValantarDude, get off your damn high horse. Please. If I am a 'wannabe righteous keyboard warrior' (righteous in regard to what? I'm genuinely curious.), I'm not even going to put into words what you are acting like. Stop making a fool of yourself. I am simply trying to have a cohesive on-topic argument here - which is what a forum is for, right? - you are the one stepping in with outlandish statements that you subsequently refuse to back up with any kind of proof. If you expect someone to trust what you say, you actually need to provide some sort of basis for trust. As of now, all you have provided is "look at me, I'm a big shot, stfu and listen to what I'm saying". Which is, as I said, empty posturing, entirely devoid of substance. As to what you are saying about NDAs, I guess that might apply in thoroughly corrupt states where the letter of the law has no value. In countries where laws work as they are supposed to (which for the record is the vast majority of applicable countries in a discussion like this) this won't be possible. And getting charged with industrial espionage for posting mostly general information on a forum? Don't be daft. That case wouldn't hold up in even the most corrupt court - espionage requires a competitor of the company in question to be involved and (trying to) benefit from it, after all. You might be liable for absurd sums of monetary damages, but jail time? Not a chance. Not to mention that what we are talking about here - the mechanical specifications of an LGA socket's mounting system - is an extremely simple and mundane thing. It might be included in a document requiring an NDA, but I will guarantee you that it isn't the key part of that document. I mean, anyone with a Solidworks licence and some mechanical engineering chops could simulate what we are talking about in half an hour. And said simulation would, of course, show that your statement that the vacuum force created by TIM adhesion to a heatsink is nowhere near strong enough to break a solder bond between an IHS and die or die and substrate. Now can we leave this dumb offshoot of the discussion, please? You're just painting yourself further into a corner with each post.

First and foremost, learn about criminal law jurisdiction before calling other people names.

Secondly, if my distinguished ME tells me the MCU BGA flip-flop and substrate resin joint will fail then I am confident he knows a bit more then a random stranger from the internet. Ever heard of fatigue cycling and ESC?
Thirdly, about that Soliworks comments - back in the day one was required to calculate vacuum pull force by hand, on a sheet od paper, if he wanted to call themselves engineers. It's not rocket science, just static mechanics.

Haile SelassieFirst and foremost, learn about criminal law jurisdiction before calling other people names.

Secondly, if my distinguished ME tells me the MCU BGA flip-flop and substrate resin joint will fail then I am confident he knows a bit more then a random stranger from the internet. Ever heard of fatigue cycling and ESC?
Thirdly, about that Soliworks comments - back in the day one was required to calculate vacuum pull force by hand, on a sheet od paper, if he wanted to call themselves engineers. It's not rocket science, just static mechanics.

Sorry, but who is calling who names exactly? I would recommend you reread each of our previous posts and get back to me with an answer to that question, as either your reading comprehension or your understanding of the term "calling someone names" is sorely lacking. Have I described your behaviour here in rather unflattering terms? Absolutely. I fully stand behind those descriptions, as they are entirely supported by your previous posts. You, on the other hand, have a) called me names, and b) refused to argue for your claims here, instead claiming that we should accept your word as truth based on... well, again, posturing. You have given us zero reason to believe what you are saying, beyond attempts to belittle me, deflect arguments, claim some secret and innate authority, and generally argue in bad faith. The onus is purely on you to change that, as you are the one bringing something to the debate that actually needs proving.

As to the rest of what you are saying: I would be asking that mechanical engineer what sort of thermal paste they are using in their models, as they might have mistakenly used numbers for a thermal adhesive instead. Even with fatigue due to thousands or hundreds of thousands of thermal cycles, there is no way the solder bonds will weaken beyond the strength of the thermal paste. Enough for one or a few bonds to break (even without mechanical stress) and render the CPU not working? Sure, that can likely theoretically happen (though it's highly unlikely to happen in practice - if this was common CPUs failing with age would itself be common, which it isn't), but that is in no way the same as the die or IHS being ripped off by the IHS sticking to the cooler. Which, to remind you, was what you said. Given that your statement is the one requiring the most new assumptions - namely, that dried up thermal paste can create a vacuum force strong enough to fully break even a severely weakened solder bond between either a <100mm2 CPU die and its IHS or the bottom of the die and its <1000 solder points connecting it to the substrate - the onus of proof falls on you. If not, your statement fails the simple test of Ockham's razor.

I also don't see what criminal law jurisdiction has to do with this. As I said, NDAs are purely a matter of civil law. Breaking an NDA is not a crime, but a breach of contract, for which you can be sued in civil court. Industrial espionage can fall under criminal law, but breaching an NDA is not equal to industrial espionage by a long shot. And if what you are saying was actually true, you know these forums have a DM function, right? You could make your argument there, we could delete the messages afterwards, and nobody would be any wiser except that you would actually be able to make your case, and I would gladly admit I was wrong here. But I strongly suspect you don't actually have a case to make at all.

And lastly, what does the Eurovision Song Contest have to do with thermal stress in solder joints? (I guess I have to point this out, but that is what you call a joke.)

ValantarSorry, but who is calling who names exactly? I would recommend you reread each of our previous posts and get back to me with an answer to that question, as either your reading comprehension or your understanding of the term "calling someone names" is sorely lacking. Have I described your behaviour here in rather unflattering terms? Absolutely. I fully stand behind those descriptions, as they are entirely supported by your previous posts. You, on the other hand, have a) called me names, and b) refused to argue for your claims here, instead claiming that we should accept your word as truth based on... well, again, posturing. You have given us zero reason to believe what you are saying, beyond attempts to belittle me, deflect arguments, claim some secret and innate authority, and generally argue in bad faith. The onus is purely on you to change that, as you are the one bringing something to the debate that actually needs proving.

As to the rest of what you are saying: I would be asking that mechanical engineer what sort of thermal paste they are using in their models, as they might have mistakenly used numbers for a thermal adhesive instead. Even with fatigue due to thousands or hundreds of thousands of thermal cycles, there is no way the solder bonds will weaken beyond the strength of the thermal paste. Enough for one or a few bonds to break (even without mechanical stress) and render the CPU not working? Sure, that can likely theoretically happen (though it's highly unlikely to happen in practice - if this was common CPUs failing with age would itself be common, which it isn't), but that is in no way the same as the die or IHS being ripped off by the IHS sticking to the cooler. Which, to remind you, was what you said. Given that your statement is the one requiring the most new assumptions - namely, that dried up thermal paste can create a vacuum force strong enough to fully break even a severely weakened solder bond between either a <100mm2 CPU die and its IHS or the bottom of the die and its <1000 solder points connecting it to the substrate - the onus of proof falls on you. If not, your statement fails the simple test of Ockham's razor.

I also don't see what criminal law jurisdiction has to do with this. As I said, NDAs are purely a matter of civil law. Breaking an NDA is not a crime, but a breach of contract, for which you can be sued in civil court. Industrial espionage can fall under criminal law, but breaching an NDA is not equal to industrial espionage by a long shot. And if what you are saying was actually true, you know these forums have a DM function, right? You could make your argument there, we could delete the messages afterwards, and nobody would be any wiser except that you would actually be able to make your case, and I would gladly admit I was wrong here. But I strongly suspect you don't actually have a case to make at all.

And lastly, what does the Eurovision Song Contest have to do with thermal stress in solder joints? (I guess I have to point this out, but that is what you call a joke.)

I don't need to give you any proof other than to look at any interconnect vendor's ZIF FCPGA socket offering to realize why this is a bad idea. Maybe take a look at some LGA construction and try to make some parallels... Some vendors are in this exact business for over 40 years. Kids think technology didn't exist pre social media era? And that Gamers Nexus guy know it all is the second coming of J.H.C.?

Lol at Eurovision (??), talking mechanical part reliability with anyone not familiar with basic ESC, let alone FCPGA packaging on organic substrate or god forbid the basic layer construction of a simple flip chip is an utter waste of time, sorry.

Why pull Tech Jesus into this...why? All I can say at this point is it might be a fun video for GN to test this bracket concept. Like if you agree!

I'm just trolling you, @Valantar You were a far too easy target to upset and get into posting excessively long emails. You must excuse me, shelter in place is taking a toll in the boredom department.

So, I'm not really an eng PM at a certain electrical interconnect company. I am an EE at the said company and my ME colleagues tell me we are indeed working on a solution to prevent pullout in a standard ZIF socket via a mechanical leaver. There are some details I can share freely, the rest is strictly new IP and as such classified.

Haile SelassieI'm just trolling you, @Valantar You were a far too easy target to upset and get into posting excessively long emails. You must excuse me, shelter in place is taking a toll in the boredom department.

So, I'm not really an eng PM at a certain electrical interconnect company. I am an EE at the said company and my ME colleagues tell me we are indeed working on a solution to prevent pullout in a standard ZIF socket via a mechanical leaver. There are some details I can share freely, the rest is strictly new IP and as such classified.

Well, sorry to disappoint I suppose - the goal of trolling seems to mostly be to get people to flip out, while I just got a bit annoyed. Think I've seen better executions, frankly. Too transparent. A lot of people think not feeding the trolls is the only way to get them to shut up, but in my experience that just makes them move on and keep going, so actually arguing against them while identifying their tactics and calling them out works much better - as me tiring you out here goes to show.

Still, I guess it's good to know you're not bonkers enough to believe thermal paste can create a stronger bond than solder.

ValantarWell, sorry to disappoint I suppose - the goal of trolling seems to mostly be to get people to flip out, while I just got a bit annoyed. Think I've seen better executions, frankly. Too transparent. A lot of people think not feeding the trolls is the only way to get them to shut up, but in my experience that just makes them move on and keep going, so actually arguing against them while identifying their tactics and calling them out works much better - as me tiring you out here goes to show.

Still, I guess it's good to know you're not bonkers enough to believe thermal paste can create a stronger bond than solder.

Ah, it's all fun and games :D