Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
NMI (Non-Maskable Interrupts) crashes are usually northbridge related, even more so when the kext crashing is com.apple.driver.AppleIntelCPUPowerManagement. Inspect the CPU tray northbridge heatsink push-pins.

If the northbridge overheated because a broken push-pin and cooked itself, you will need a replacement CPU tray. There is no economical way to repair the CPU tray when the northbridge is damaged.

You could also have BootROM problems because of the crashes being saved inside the NVRAM for months, first PanicLog is from Sun Apr 16, but first inspect the push-pins.

Wouldn't the temperature readings I attached in the original post show if there was a NB heatsink issue?

I took the two processor heatsinks off and had a look at the NB heatsink push-pins. They look OK.
I have attached some photos.
Thanks.
 

Attachments

  • photo_5_2023-07-10_22-25-12.jpg
    photo_5_2023-07-10_22-25-12.jpg
    196.7 KB · Views: 172
  • photo_4_2023-07-10_22-25-12.jpg
    photo_4_2023-07-10_22-25-12.jpg
    348.3 KB · Views: 95
  • photo_1_2023-07-10_22-25-12.jpg
    photo_1_2023-07-10_22-25-12.jpg
    351 KB · Views: 96
Wouldn't the temperature readings I attached in the original post show if there was a NB heatsink issue?

Temperature is not a dead giveaway unless the delta is extreme - you have to know how much time the Mac Pro is on, if it's in load, ambient temperature and etc.

I took the two processor heatsinks off and had a look at the NB heatsink push-pins. They look OK.

This is definitively not ok, the arrow point is already too open and will break soon, but it's not the main problem.

I have attached some photos.
Thanks.

This CPU tray is too dirty, the oily residue from the thermal pads + accumulated dust becomes conductive overtime and you start to have crashes. Clean all the mess with isopropanol, re-assemble and test again.

Do you have access to another working CPU tray to check if the problem is only the CPU tray and the rest of the Mac Pro is healthy?
 
Temperature is not a dead giveaway unless the delta is extreme - you have to know how much time the Mac Pro is on, if it's in load, ambient temperature and etc.



This is definitively not ok, the arrow point is already too open and will break soon, but it's not the main problem.



This CPU tray is too dirty, the oily residue from the thermal pads + accumulated dust becomes conductive overtime and you start to have crashes. Clean all the mess with isopropanol, re-assemble and test again.

Do you have access to another working CPU tray to check if the problem is only the CPU tray and the rest of the Mac Pro is healthy?
I don't have another CPU tray.
I will clean the board.
I may as well replace the pins for the NB heat sink too. I bought some a few weeks ago just in case.
Thanks.
 
Temperature is not a dead giveaway unless the delta is extreme - you have to know how much time the Mac Pro is on, if it's in load, ambient temperature and etc.



This is definitively not ok, the arrow point is already too open and will break soon, but it's not the main problem.



This CPU tray is too dirty, the oily residue from the thermal pads + accumulated dust becomes conductive overtime and you start to have crashes. Clean all the mess with isopropanol, re-assemble and test again.

Do you have access to another working CPU tray to check if the problem is only the CPU tray and the rest of the Mac Pro is healthy?
Hi,
While cleaning the CPU tray, the cable that attaches to the NB heatsink disconnected.
It looks like it was glued. And I don't know what the grey stuff is on the end.
Can this be reattached?

Thanks
 

Attachments

  • photo_2_2023-07-13_21-25-11.jpg
    photo_2_2023-07-13_21-25-11.jpg
    383.7 KB · Views: 82
  • photo_3_2023-07-13_21-25-11.jpg
    photo_3_2023-07-13_21-25-11.jpg
    207.2 KB · Views: 95
  • photo_5_2023-07-13_21-25-11.jpg
    photo_5_2023-07-13_21-25-11.jpg
    226.6 KB · Views: 90
  • photo_1_2023-07-13_21-25-11.jpg
    photo_1_2023-07-13_21-25-11.jpg
    317.4 KB · Views: 87
  • photo_4_2023-07-13_21-25-11.jpg
    photo_4_2023-07-13_21-25-11.jpg
    157.5 KB · Views: 85
Can this be reattached?

This is the temperature sensor, you will remove the old thermal glue/epoxy the best you can and then re-apply it.

If you can't find thermal glue/epoxy easily, you can improvise it with 2/3 slow cure/24h bi-component epoxy and 1/3 white silicon thermal paste. Do not use fast cure epoxy, it will break easily when mixed with thermal paste. You will need at least a day to re-install it, but works fine and I've used this formula a lot over the years.
 
This is the temperature sensor, you will remove the old thermal glue/epoxy the best you can and then re-apply it.

If you can't find thermal glue/epoxy easily, you can improvise it with 2/3 slow cure/24h bi-component epoxy and 1/3 white silicon thermal paste. Do not use fast cure epoxy, it will break easily when mixed with thermal paste. You will need at least a day to re-install it, but works fine and I've used this formula a lot over the years.
Great.

I am wondering if that cable was a little loose, it may have been giving a false readout showing the NB HS as lower than it actually was?

I found this. Would it work?



Product Overview​

TBS20S


Adhesive, Thermal Bonding System, Epoxy - 2 Part, Syringe, Blue, Cream, 20 ml


Thermal Bonding Compound is a two part epoxy bonding system which utilises metal oxides to provide excellent thermal conductivity whilst being electrically insulating. It is especially useful in the manufacture of heatsink assemblies where 'piggy back' arrangements are applied and where the manufacture design of heat sinks does not allow for welding or brazing techniques to be employed due to complexity or geometry of the fins. TBS is also ideal for use as a bonding medium in surface mounting assemblies. Key Properties: Two part epoxy bonding system Very high bond strength High thermal conductivity: 1.10 W/m.K Eliminates need for mechanical fixing by providing a permanent bond Wide operating temperature range: -40°C to +120°C Includes glass beads for a set thickness to be applied


  • Excellent tensile strength.
  • Very good thermal conductivity
  • Excellent electrical insulation characteristics
  • Part A colour: Blue
  • Part B colour: Cream
  • Package: 20 ml twin syringe system


Thanks
 
Great.

I am wondering if that cable was a little loose, it may have been giving a false readout showing the NB HS as lower than it actually was?

Very possible.

I found this. Would it work?






Thanks

It's not the same type used by Apple, but should work. Any thermal glue with good conductivity will work, some will take more time to transfer the heat to the sensor, but in the end you just need a reliable reading.

Try to remove all the old thermal glue before applying the new one, also clean the heatsink pocket for the sensor.
 
  • Like
Reactions: cakey
Very possible.



It's not the same type used by Apple, but should work. Any thermal glue with good conductivity will work, some will take more time to transfer the heat to the sensor, but in the end you just need a reliable reading.

Try to remove all the old thermal glue before applying the new one, also clean the heatsink pocket for the sensor.
The grey stuff on the sensor, is that the thermal glue too? It looks like something else compared to the orange coloured stuff left on the cable.
 
I glued the sensor to the NB HS.
One question. The CPU heatsinks have this pink strip as shown in the photo. On touch, it feels a little tacky, like there was something on it.
Do I need to put anything on it, or just clean it?
Thanks.

1689950079620.png




1689950270302.png
 
If the thermal pads are degraded, you need to replace it. There is a thread about the thermal pads.

I'd clean the current one first, reassembly everything and test the CPU tray to be sure that it is working correctly.
 
If the thermal pads are degraded, you need to replace it. There is a thread about the thermal pads.

I'd clean the current one first, reassembly everything and test the CPU tray to be sure that it is working correctly.
I re-assembled everything.
It booted normally, and I had it on for a couple of hours last night. The NB diode temperature peak was 77c peak and NB HS peak was 65c.
This morning, I turned it on for 20 mins.
Then restarted and it ran into issues.
First, I got the Apple logo and a prohibitory symbol - flashing between the two.
I turned it off and took out all the other drives, bar the OS drive.
Turned it on, then got the load up screen which hung a few times, then rebooted a few times, then the black screen.
Turned it off and restarted, and it loaded up, I typed my password and rebooted. That happened twice, followed by black screen.
Then I turned it off. Turned on again and it booted up normally.
The NB D was 77c and NB HS is 65c after almost 30 mins.
I saved the crash report.
I restarted, and it booted up normally and has been on for 40 mins. Same temperature rates as above.
Thanks.
 

Attachments

  • cp7.pdf
    24.6 KB · Views: 82
Last edited:
Same problem. I'd test a known working CPU tray, to be sure that this CPU tray is toast.
 
  • Like
Reactions: cakey
Same problem. I'd test a known working CPU tray, to be sure that this CPU tray is toast.
I had a feeling you were going to say that.
If I was to buy a CPU tray, I would prefer getting the same as mine, in case it needed to be replaced. But, looking at the prices of the same CPU tray I have, it may be better replacing the whole computer and selling my computer.
 
Sure, but if you have a little patience and search everyday the auction sites, call your nearby e-recyclers/etc, you'll probably gonna find a whole dual CPU mid-2010 Mac Pro on the cheap, then you will need to buy just a single CPU tray to make a spare.

Get the best parts of each Mac Pro to make a very good one for you and then you have another one to sell and recoup a big chunk of the expenses.
 
  • Like
Reactions: cakey
It took a few months, but I managed to pick up a 2nd hand 5,1 for a very reasonable price. The whole computer was cheaper than the price of the trays for sale.
The specs aren't as good, but everything seems to run fine.
I swapped the trays around and have run both computers (playing videos online) for 15 hours. Neither of them have crashed so far.
As I am trying to figure out what is wrong here, is there anything else I could do to see if there are any issues, bar wait for a crash?
Thanks.
 
Hello,
Had this issue for a while.
While running, it freezes. I have to power off as a result. It could be after 10 minutes use or it could happen after being on for 10 hours (- not using any apps/programs in that time), then using an app.
Running latest Ventura with latest OCLP.

Here are two reports from last month and today.

Is that enough to get the gist of what the issue is?

Thanks
 

Attachments

  • 20251.pdf
    34.1 KB · Views: 26
  • 2025 6.pdf
    35.1 KB · Views: 29
Your Kernel Panics are not OCLP related:

Code:
mp_kdp_enter() timed-out on cpu 4, NMI-ing
mp_kdp_enter() NMI pending on cpus: 0 1 2 3 5 6 7 8 9 10 11 12 13 14 15 mp_kdp_enter() timed-out during locked wait after NMI;expected 16 acks but received 1 after 16938939 loops in 1196999846 ticks

Code:
mp_kdp_enter() timed-out on cpu 11, NMI-ing
mp_kdp_enter() NMI pending on cpus: 0 1 2 3 4 5 6 7 8 9 10 12 13 14 15 mp_kdp_enter() timed-out during locked wait after NMI;expected 16 acks but received 1 after 16924772 loops in 1196999846 ticks

NMI related Kernel Panics is the classic crash caused by overheating, another major hint is com.apple.driver.AppleIntelCPUPowerManagement right in the start of the backtrace - check if the push-pins of CPU tray northbridge heatsink are not broken, replace the broken pins, clean throughly everything and re-paste it.



Could also be something else defective on the CPU tray, if the push-pins of the heatsink are still attached and the CPU tray is clean. This takes another known working Mac Pro to diagnose. If you have another Mac Pro, install the known working CPU tray to the crashing Mac Pro and see if the crashes stop, install the crashing one to the known working Mac Pro. The KPs now should happen with the known working Mac Pro.

This is very important, constant Kernel Panics will brick the Mac Pro backplane. The crashes are saved inside the NVRAM volume and since the management/garbage collection of the VSS store is prone to fail when over pressure, the constant crash logs being saved inside the NVRAM volume will overwrite the VSS store and corrupt the BootROM. This is very common with Mac Pros constantly crashing and a major cause of bricks.

Never let a Mac Pro crash it frequently - after you get it working you most probably will need the BootROM reconstruction service to repair the BootROM.
 
Your Kernel Panics are not OCLP related:

Code:
mp_kdp_enter() timed-out on cpu 4, NMI-ing
mp_kdp_enter() NMI pending on cpus: 0 1 2 3 5 6 7 8 9 10 11 12 13 14 15 mp_kdp_enter() timed-out during locked wait after NMI;expected 16 acks but received 1 after 16938939 loops in 1196999846 ticks

Code:
mp_kdp_enter() timed-out on cpu 11, NMI-ing
mp_kdp_enter() NMI pending on cpus: 0 1 2 3 4 5 6 7 8 9 10 12 13 14 15 mp_kdp_enter() timed-out during locked wait after NMI;expected 16 acks but received 1 after 16924772 loops in 1196999846 ticks

NMI related Kernel Panics is the classic crash caused by overheating, another major hint is com.apple.driver.AppleIntelCPUPowerManagement right in the start of the backtrace - check if the push-pins of CPU tray northbridge heatsink are not broken, replace the broken pins, clean throughly everything and re-paste it.



Could also be something else defective on the CPU tray, if the push-pins of the heatsink are still attached and the CPU tray is clean. This takes another known working Mac Pro to diagnose. If you have another Mac Pro, install the known working CPU tray to the crashing Mac Pro and see if the crashes stop, install the crashing one to the known working Mac Pro. The KPs now should happen with the known working Mac Pro.

This is very important, constant Kernel Panics will brick the Mac Pro backplane. The crashes are saved inside the NVRAM volume and since the management/garbage collection of the VSS store is prone to fail when over pressure, the constant crash logs being saved inside the NVRAM volume will overwrite the VSS store and corrupt the BootROM. This is very common with Mac Pros constantly crashing and a major cause of bricks.

Never let a Mac Pro crash it frequently - after you get it working you most probably will need the BootROM reconstruction service to repair the BootROM.
Thanks for the response.
I actually have two MPs as I had the same issue before. You recommended I get hold of a spare, which I did for a good price.
I replaced the pins, cleaned and repasted on the original tray. That tray runs with no issues on the spare MP - or it has never had a KP running videos all day a few times (uses the ram that came with the spare).
The tray inside the current original MP is from the spare, although the ram is not from the spare. Everything else is from the original MP. So, the original MP is now also having KPs with the spare tray, too.
What would you recommended here apart from replacing the pins and cleaning and replacing the paste on the spare tray and trying again?
Could it be something not related to the tray, or have I been unlucky with two dodgy trays?
Thanks
 
NMI crashes are always a hardware defect and 90% of the time is something defective with the CPU tray, but defective RAM and a marginal PSU could be the reason that the CPU tray that does not crash with one Mac Pro crashes when installed to the other.

Run AHT first, then ASD, maybe you'll get a clue on whatever else is wrong.
 
I tried doing an AHT and cannot. Holding D after power on, I get the grey startup screen, then the EFI Boot screen. Holding option D I get the black startup screen and the EFI Boot screen.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.