Intel finally responds to bending 12th gen Alder Lake CPUs on LGA1700 motherboards

GiGaBiTe said:
At least it wasn't like Socket 462 where you had to heave your entire body weight on the retention spring and risk slipping and punching a hole straight through the motherboard, or at bare minimum, ripping chunks and components off of it.
Click to expand...
I don't recall having any issues with those, stock coolers, or water block clips.
 
The Mad Atheist said:
I don't recall having any issues with those, stock coolers, or water block clips.
Click to expand...

Even taking great care I dented a couple of mobos when the clip mechanism moved in an unexpected way, causing the screwdriver to slip off the retention clip.
It was fortunate no tracks on the mobo were cut during the impact.
Building/working on enough machines you encounter the significant issues or hear of them from colleagues.
 
kac77 said:
LMAO. OMG this was the scariest thing dealing with that socket. I've never applied so much pressure using a flat head screwdriver in my life.

Once it was on it was on. There was no IHS either so if you got that CPU to boot it was an act of God considering the amount of pressure applied to the CPU package.
Click to expand...

Companies started selling copper shims for Socket 462 CPUs because of people cracking dies or chipping the corners that resulted in a dead CPU. I still don't know why AMD opted to not have an IHS on that generation, when the K6 before it and the Athlon 64 after it did.

The Mad Atheist said:
I don't recall having any issues with those, stock coolers, or water block clips.
Click to expand...

I'd argue you've never installed a Socket 462 CPU, or forgot how hard it was. You can still punch "amd socket 462 cracked cpu" in Google today and see threads from 20 years ago about people cracking their Athlons, Durons and Semprons because of the garbage spring and latch mechanism. Even someone earlier in the thread said they had the same issue. I've built hundreds of Socket 462 machines, it was very much a problem, even if you were as careful as possible.

Companies didn't start selling metal shims for Socket 462 CPUs for no reason. I still have one in my parts bins somewhere.
 
Heh, I remember some motherboard mfg's started putting protective plastic near where the flathead would be, to help stop trace cutting from a slip.
 
GiGaBiTe said:
Companies started selling copper shims for Socket 462 CPUs because of people cracking dies or chipping the corners that resulted in a dead CPU. I still don't know why AMD opted to not have an IHS on that generation, when the K6 before it and the Athlon 64 after it did.



I'd argue you've never installed a Socket 462 CPU, or forgot how hard it was. You can still punch "amd socket 462 cracked cpu" in Google today and see threads from 20 years ago about people cracking their Athlons, Durons and Semprons because of the garbage spring and latch mechanism. Even someone earlier in the thread said they had the same issue. I've built hundreds of Socket 462 machines, it was very much a problem, even if you were as careful as possible.

Companies didn't start selling metal shims for Socket 462 CPUs for no reason. I still have one in my parts bins somewhere.
Click to expand...

I still have my shim from my 1.4 Tbird back in the day, although once I got my Alpha PAL 8045 I didnt have to worry about terrible heatsink latches. I did break the tabs off a socket 370 board once, along with wrecking a couple boards once the screwdriver slipped out of the heatsink latch. It took a lot of force to get them on, and if you didnt put it on even, or your corner pads were messed up, you could easily crack the corner off the core.
 
GiGaBiTe said:
Companies started selling copper shims for Socket 462 CPUs because of people cracking dies or chipping the corners that resulted in a dead CPU. I still don't know why AMD opted to not have an IHS on that generation, when the K6 before it and the Athlon 64 after it did.



I'd argue you've never installed a Socket 462 CPU, or forgot how hard it was. You can still punch "amd socket 462 cracked cpu" in Google today and see threads from 20 years ago about people cracking their Athlons, Durons and Semprons because of the garbage spring and latch mechanism. Even someone earlier in the thread said they had the same issue. I've built hundreds of Socket 462 machines, it was very much a problem, even if you were as careful as possible.

Companies didn't start selling metal shims for Socket 462 CPUs for no reason. I still have one in my parts bins somewhere.
Click to expand...
Or god forbid the “I slammed my screwdriver into the MB when it slipped off the latch and now my board is dead” crowd. That managed to somehow not damage the CPU to only destroy the board instead.

Boy did those make me appreciate the old Slot-1 and Slot-2 systems.
 
GiGaBiTe said:
Companies started selling copper shims for Socket 462 CPUs because of people cracking dies or chipping the corners that resulted in a dead CPU. I still don't know why AMD opted to not have an IHS on that generation, when the K6 before it and the Athlon 64 after it did.



I'd argue you've never installed a Socket 462 CPU, or forgot how hard it was. You can still punch "amd socket 462 cracked cpu" in Google today and see threads from 20 years ago about people cracking their Athlons, Durons and Semprons because of the garbage spring and latch mechanism. Even someone earlier in the thread said they had the same issue. I've built hundreds of Socket 462 machines, it was very much a problem, even if you were as careful as possible.

Companies didn't start selling metal shims for Socket 462 CPUs for no reason. I still have one in my parts bins somewhere.
Click to expand...
Owned about 5 Athlons back in the day in my Soyo Dragon MOBOs and remounted them countless of times, no issues what so ever. Didn't have problems with 386-P5 chips while getting my AS in computers and electronics back in 96 either. I was more worried about connecting AT power, or jumpers wrong. I remember shims being sold and ,maybe, a thread about them on Directron, still don't recall them being widely used.
 
I milled a TRUE flat (or TRUE haha) verified on a CMM to +/- 0.0001" precision. In other words as flat as can be! Looked beautiful.
However the nightmare ensued, temperatures were WORSE and a lot of the boards were missing memory channels. They are convex for a reason...

Still better than bare die socket 462 Athlons. Even though I never killed or chipped one, installing some of the coolers with strong springs over those socket tabs (and even worse removing them!) made me nervous. Loved the through bolt options of Swiftech and Alpha though. Swiftech even mounted one up, installed the motherboard in a chassis and threw it out a window destroying the box and the chip survived. This demonstrated it was safe to ship a system with one of their heavy heatsinks mounted.
 
Had a lot of 462 Athlons, Was fortunate to never damage one or burn it up because the heat sink wasn't attached / well. Although I do remember that happening to a few folks because they didn't have thermal protection back then.

My 12th gen might be getting bent but the temps are just fine under full load. But I'm also on an MSI board with the Lotes.

intel_getbent.jpg
 
The_Heretic said:
Although I do remember that happening to a few folks because they didn't have thermal protection back then.
Click to expand...

While the CPU did not have thermal protection, motherboards usually had thermal protection. Unfortunately, 99% of the time it was disabled by default, you had to manually go in and specify a thermal warning temperature and thermal shutdown temperature. I always made sure to turn both of these on because customers would never clean or perform regular maintenance on their machines. The CPU heatsink would invariably get 100% plugged, or the fan would fail and at the machine shutting off spontaneously would force them to do something to fix it.

Motherboard thermal protection existed at least as far back as Socket 7.
 
GiGaBiTe said:
While the CPU did not have thermal protection, motherboards usually had thermal protection. Unfortunately, 99% of the time it was disabled by default, you had to manually go in and specify a thermal warning temperature and thermal shutdown temperature. I always made sure to turn both of these on because customers would never clean or perform regular maintenance on their machines. The CPU heatsink would invariably get 100% plugged, or the fan would fail and at the machine shutting off spontaneously would force them to do something to fix it.

Motherboard thermal protection existed at least as far back as Socket 7.
Click to expand...
Unfortunately in the event of a heatsink falling off, the chip will die in a near instant!

As we've all seen many times LOL!

 
GiGaBiTe said:
While the CPU did not have thermal protection, motherboards usually had thermal protection. Unfortunately, 99% of the time it was disabled by default, you had to manually go in and specify a thermal warning temperature and thermal shutdown temperature. I always made sure to turn both of these on because customers would never clean or perform regular maintenance on their machines. The CPU heatsink would invariably get 100% plugged, or the fan would fail and at the machine shutting off spontaneously would force them to do something to fix it.

Motherboard thermal protection existed at least as far back as Socket 7.
Click to expand...

I have an aging memory now, but I do not remember any thermal protection for those Athlons. If that HS came off while it was running, it fried. The first boards didn't even have a temp sensor, but that was added later. However don't remember that being included in any protection/prevention. It was just a way to read a CPU temp.
 
The_Heretic said:
I have an aging memory now, but I do not remember any thermal protection for those Athlons. If that HS came off while it was running, it fried. The first boards didn't even have a temp sensor, but that was added later. However don't remember that being included in any protection/prevention. It was just a way to read a CPU temp.
Click to expand...
My Abit socket A boards had a cpu temp thermistor in the middle of the socket. That was it.
 
GiGaBiTe said:
Not true at all. Motherboard temp sensors go back almost a decade before even Socket 462.



Most definitely existed, you just never looked for it.
Click to expand...
I wasn't speaking about anything other than Socket-A, specifically some of the very first boards. However as I said that's been quite a while ago.

Go look up an Asus A7V and you'll see it had the ability to read the CPU temp, but no setting to perform a shutdown.

1650231455970.png
 
cpufrost said:
That looks like Al and considering the cost, is quite remarkable. I certainly could never machine something for that price!
Click to expand...
Yea didn't catch the price at first. $6 bucks is pretty cheap for a solid AL part like that. It is probably stamped unless this is going to be a stupid low volume item.
 
cpufrost said:
That looks like Al and considering the cost, is quite remarkable. I certainly could never machine something for that price!
Click to expand...
It's probably molded, polished, then anodized. way cheaper that way. You get a lot less waste too.
 
cpufrost said:
Unfortunately in the event of a heatsink falling off, the chip will die in a near instant!

As we've all seen many times LOL!

Click to expand...


Missed this. Yes, this will happen on boards with no thermal protection enabled/available. I have saved many Socket 462 CPUs with motherboard based thermal protection though. I've tested removing the heatsink a few times with crappy CPUs, motherboard thermal protection will shut the board off near instantly when the heatsink is removed or compromised and lets the temperature pass the set threshold. My own Athlon XP 3200+ rig back in the day was also saved many times when the fan failed while I was away, and those crappy high RPM fans failed quite a lot.

vegeta535 said:
Could just 3D print something like that.
Click to expand...

There's no 3D print filament I know of that has the rigidity of solid metal at that size. PLA is rigid but brittle and would crack. PETG is stronger but too flexible, same with ABS. There are "metal" printer filaments, but that's just finely ground metal particles that use PLA as a binder.
 
GiGaBiTe said:
Missed this. Yes, this will happen on boards with no thermal protection enabled/available. I have saved many Socket 462 CPUs with motherboard based thermal protection though. I've tested removing the heatsink a few times with crappy CPUs, motherboard thermal protection will shut the board off near instantly when the heatsink is removed or compromised and lets the temperature pass the set threshold. My own Athlon XP 3200+ rig back in the day was also saved many times when the fan failed while I was away, and those crappy high RPM fans failed quite a lot.
Click to expand...
That's good to know. I was worried that without a heatspreader or heatsink the semiconductor die would rapidly develop hotspots that might burn before a temperature sensor might register.
 
Last edited:
serpretetsky said:
That's good to know. I was worried that without a heatspreader or heatsink the die would very rapidly develop hotspots that might burn before a temperature sensor might even register.
Click to expand...

Depends on the motherboard and how the thermistor was situated. The best ones I've seen had them on a bit of spring loaded flexible ribbon cable directly under the CPU die in the middle of the socket so they pressed up against the bottom of the CPU. Most of them had the thermistor on the motherboard in the middle of the socket though. But in either case, it gets hotter than the blazes of hell in a hurry under there.
 
GiGaBiTe said:
Depends on the motherboard and how the thermistor was situated. The best ones I've seen had them on a bit of spring loaded flexible ribbon cable directly under the CPU die in the middle of the socket so they pressed up against the bottom of the CPU. Most of them had the thermistor on the motherboard in the middle of the socket though. But in either case, it gets hotter than the blazes of hell in a hurry under there.
Click to expand...
Oh.. i was assuming it was using an on-die sensor. What you described makes me even more worried. :p
 
I watercool and lap both cpu and block for every new install. Been doing this as a matter of course for years. IHSs have always been a problem in some form, but it's better than a cracked cpu.
 
spine said:
I watercool and lap both cpu and block for every new install.
Click to expand...
Wouldn't you have to lap the CPU while it's in some kind of fake socket that replicates the tension it's under while installed?

Just like woodworking planes, you can't flatten them without a blade clamped in there or else you still won't end up with something flat when you're ready to work.
 
You must log in or register to reply here.
Back
Top