Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Snow Tiger

macrumors 6502a
Dec 18, 2019
854
634
Why not use bolts and nuts? I am using bolts and nuts with the stock springs and the space between the bolt's head and the (self saving) nut exactly equal to the stock pin's length. All works perfectly fine. Only difference to the pin solution is, that it is very unlikely to break.

It might work until it doesn't - i.e. crack the NB die ( because the NB HS is too close to the NB chip) or prematurely dry up the thermal paste ( because the NB HS is too far from the NB chip ) . You're dealing with a tiny range of movement ( and yes , that Northbridge heatsink has to move ever so slightly as the NB chip expands and contracts due to changes in its temperature ) . Either way , System goes offline .

You won't achieve factory tolerances using this method ( nuts and bolts ) , the pressure rating of the factory spring is an Apple secret anyways . I actually made industry inquiries into this years ago . Apple used push pins for a reason and not nuts and bolts with springs .

As far as i see, the pic shows a common pin like it has been used on many older GPUs. I wouldn't recommend these as their springs are much weaker than what is used on the cMP’s genuine northbridge heatsink pins.

Metal springs are not going to break . In fifteen years I haven't seen a broken spring from a GPU . What will break is the plastic pin itself , due to thermal fatigue . The best replacement is a push pin and installation method that matches the factory pin and pressure as closely as possible . This is something that can be done by hand ( by adjustments ) once the fasteners have been applied . You get a feel for how taut the NB heatsink is after installation by lateral and horizontal test movements . It should move , but extremely slightly - just like the Northbridge installed at the factory .

By the way , I have a container full of spare Apple factory NB fastener springs I can't use - you just cannot place them on after market push pins and expect them to work properly ( I tried that ) . The pins and their matching springs are designed as a single , properly working unit .
 

flyproductions

macrumors 65816
Jan 17, 2014
1,082
461
You're dealing with a tiny range of movement ( and yes , that Northbridge heatsink has to move ever so slightly as the NB chip expands and contracts due to changes in its temperature )
Exactly which part of „I'm using bolts and nuts with the stock springs“ You did not understand? So it should be clear from the reading, that the pressure to the die is given by the spring and not by screwing down the nut. And, as the length between head and nut is exactly matching the measure of the pins Apple uses - the pressure to the die should also be the exact same as with the pins before.

...which will not be the case with the pins your picture shows. The springs used in the stock configuration are much stronger than the the ones of those GPU clips. And even if those springs don't break. The pins themselfs might just as well as the ones Apple built into the cMP. Or even more likely.
 
Last edited:

Snow Tiger

macrumors 6502a
Dec 18, 2019
854
634
a
Exactly which part of „I'm using bolts and nuts with the stock springs“ You did not understand? So it should be clear from the reading, that the pressure to the die is given by the spring and not by screwing down the nut. And, as the length between head and nut is exactly matching the measure of the pins Apple uses - the pressure to the die should also be the exact same as with the pins before.

...which will not be the case with the pins your picture shows. The springs used in the stock configuration are much stronger than the the ones of those GPU clips. And even if those springs don't break. The pins themselfs might just as well as the ones Apple built into the cMP. Or even more likely.

If you insist on using plastic nuts to secure a vital component in a System , then at least secure it with some Loctite . Assuming you have the proper tension on that fastener , then the Loctite will seal it in place . If you screwed up and the System goes offline , then you can always break the seal on that nut with enough torque from a hand tool and install a proper fastener .
 
Last edited by a moderator:

designed

macrumors 6502
Original poster
Nov 8, 2005
286
1
Finland
Today I received the good thermal paste and TIM (thermal paste + chipset) cleaning solution and re-did all of the cooling, apart from replacing any fans. I cleaned up even more dust from the chipset, cleaned the heatsink surfaces, the CPUs tops and applied a decent amount of the new thermal paste.

I also swapped out the heatsink thermal pads cooling the misc circuitry. New pads were 2mm thick and I made sure they weren't electrically conductive. Fits perfect even if I was a little generous cutting them. Then I did all the screws until hand tight (but not arm tight).

Then I zapped SMC and PRAM to activate the higher memory clocks of the new DIMMs, that wasn't registered yesterday.

Just finished a Folding work unit with zero drama. CPU A diode stayed at 75°C the entire time, meaning I could swap some degress for a few hundred fan rpms. PMSET showed no issues for the entire time. Tomorrow it's marathon time.
 
  • Like
Reactions: h9826790

h9826790

macrumors P6
Apr 3, 2014
16,656
8,584
Hong Kong
Today I received the good thermal paste and TIM (thermal paste + chipset) cleaning solution and re-did all of the cooling, apart from replacing any fans. I cleaned up even more dust from the chipset, cleaned the heatsink surfaces, the CPUs tops and applied a decent amount of the new thermal paste.

I also swapped out the heatsink thermal pads cooling the misc circuitry. New pads were 2mm thick and I made sure they weren't electrically conductive. Fits perfect even if I was a little generous cutting them. Then I did all the screws until hand tight (but not arm tight).

Then I zapped SMC and PRAM to activate the higher memory clocks of the new DIMMs, that wasn't registered yesterday.

Just finished a Folding work unit with zero drama. CPU A diode stayed at 75°C the entire time, meaning I could swap some degress for a few hundred fan rpms. PMSET showed no issues for the entire time. Tomorrow it's marathon time.
75°C is a nice number, not too high, Turbo Boost always avail.

Also not too low, not too much unnecessary fan noise.
 
  • Like
Reactions: designed

flyproductions

macrumors 65816
Jan 17, 2014
1,082
461
If you insist on using plastic nuts to secure a vital component in a System , then at least secure it with some Loctite . Assuming you have the proper tension on that fastener , then the Loctite will seal it in place .
So, as you seem to have reported my last post, which has been deleted after for absolutely NO reason after you blamed me for „foolishly“ doing things, i try „to make my point without taking it to a personal level“ now:

a) I don't know where you have that „plastic nuts“-thing from. I didn't write this anywhere nor am i using them.

b) I am not using loctite because it would a) be aggressive to plastic if i would use plastic bolts and nuts nor would it work with this kind of material and i b), as i have written, am using self locking nuts on metal bolts, which lock in any position they are set to in a much better, cleaner and equal secure way than loctite would.

c) I have a lot of older GPUs with their heatsinks attached with pins like the one shown in your picture. The wire their springs are made of does not even have half the diameter of the one used in the Apple parts. For that reason they can give only a small portion of the pressure Apple intended to hold down the northbridge heatsink in the Mac Pro. In other way, using the springs of the (broken) Apple parts with these fragile pins would most likely instantly break them. So i can in no way suggest to use those cheap GPU-heatsink-pins to keep the northbridge heatsink in place in a cMP!
 
Last edited:
  • Like
Reactions: h9826790

Jamie Kehoe

macrumors newbie
Dec 17, 2019
14
1
Sydney, Australia
I've had my cMP 4,1 dual CPU since new in 2009 but only decided to upgrade it now as a part of my "killing time during corona" project.

I purchased a duo of X5690s already de-lidded and installed them few days ago. It wasn't immediately a huge success, but "re-refitting" the heatsinks solved it mostly. My problem is that the CPU A still runs a bit hot even after re-pasting couple of times. This means I need to have more rpms in MFC than I'm comfortable with, even if my temp goals are little bit ambitious compared to maximum allowed. I know there's a considerable TDP difference between the old 2,26 GHz Nehalems and the full-fat Westmeres.

I also realise the CPU A heatsink does the Northbridge cooling. One of the NB pins was broken and while examining the other it broke off too, but I replaced them with equivalents from an old Asus motherboard. I re-pasted the NB while at it and the NB temps have been sweet ever since.

I read in an old post here that you can alternatively skip the plastic spacers used with de-lidded CPUs to better equalise the temp between CPU diode and the heatsink. Has any current user done this to any risk or effect? I'll also clean any excess silicone gunk left over from the de-lidding.

I have the exact same setup as you do and replaced these processors early on this year. Mine is sitting at about 53 degrees celsius for CPU1 and around 40 degrees C for CPU2 - I guess I never bothered to fiddle to get them close as they are both within spec. I used a good quality thermal paste and reseated mine easily without any hassles. I knocked off the lids with a small vice. I think CPU1 is used more as not all apps are multicore apps. So it may be that CPU2 is sitting idle some of the time? Might need to do a taxing test with a CPU benchtesting app.
 

designed

macrumors 6502
Original poster
Nov 8, 2005
286
1
Finland
I think CPU1 is used more as not all apps are multicore apps. So it may be that CPU2 is sitting idle some of the time? Might need to do a taxing test with a CPU benchtesting app.

Yea, CPU A does most of the work by default. My temps are not very far off yours during regular use and the machine is quiet even compared to 2009 stock configuration. Fan profiles and the better cooled display adapter help.

I'm more interested in 100% load temps 'tho, I currently have my cMP Folding for ~12 hours a day. Later it will re-become my desktop system, which should be a bit easier on the old girl.
 

Jamie Kehoe

macrumors newbie
Dec 17, 2019
14
1
Sydney, Australia
Yea, CPU A does most of the work by default. My temps are not very far off yours during regular use and the machine is quiet even compared to 2009 stock configuration. Fan profiles and the better cooled display adapter help.

I'm more interested in 100% load temps 'tho, I currently have my cMP Folding for ~12 hours a day. Later it will re-become my desktop system, which should be a bit easier on the old girl.

They are champion machines, she might be an oldie, but she's a goodie. They are such battle hardened machines, it will keep on ticking.
 
  • Like
Reactions: designed
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.