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

Mac Pro 4.1 only boots after a rest

S

SteelBadger

macrumors newbie
Original poster
May 17, 2019
11
0
Hi all,

I've a Mac Pro flashed from 4.1 to 5.1 which requires a 'rest' of about two hours from shutting down to restarting properly. If I restart immediately then usually get a progress bar that stops at about 75% and never progresses beyond that. I have an internal spinning disk hard drive as a backup (boot disk is an ssd), and I can hear that tinkling away long after the progress bar has stopped moving. I've learnt that if I wait for a couple of hours before turning the machine back on then then it boots more or less reliably. This was an issue before the 5.1 flash, so I think that can probably be discounted. Booting form a USB stick has the same result. I assume it's some sort of thermal cutout, but I've really no idea.

The current situation is a bit more serious, I have a non flashed RX560, so I no longer get the progress bar, the screen is always blank until I get a successful boot. Now I've recently installed new X5680 cpus to fix the audio stuttering bug. It won't boot fully at all but I get the startup bong sound and the hard disk sounds the same as it did before, running for many minutes, so somethings going on but I don't know what. The symptoms seem the same as previous failed boots, but it might be the cpus or the installation.

Does anyone have any ideas on what I might do to fix it or investigate further?
 
Synchro3 said:
Check the northbridge heatsink rivets.
Click to expand...
Thanks, I'll have a look tonight. It may have got a slight knock during the CPU replacement that's made the problem worse.

I'm not sure how I would check the PSU.

A further bit of info, while it's in it's stalled during boot condition all the fans appear to work correctly, including the ones on the GPU. The B processor heatsink is notably warmer than the A.

When it did boot, it was totally stable and I used to run it for months without a restart.
 
Last edited:
Synchro3 said:
Check the northbridge heatsink rivets.
Click to expand...

A good point, but I would have thought if the NB overheating was the culprit it would show issues in use as well?

OP do you have any temp monitoring programs installed? if not go grab one and see what your NB is reporting. Typically if the heatsink clips have gone or it's making bad contact you'd see a very high NB temp, with a suspiciously low NB heatsink temp.
 
Sadly I'm in a position where the machine won't finish booting with very limited information about what's going on.

The symptoms I see now with the new CPUs appear the same as when I had the previous CPUs and restarted too quickly. It might be something unrelated though, like bent pins during the upgrade etc.
 
MIKX said:
Have you done the NVRAM reset ( 3 times in succession ) ?
Click to expand...

I've tried resetting the NVRAM as suggested. No dice.

I've investigated the northbridge. The push-pins are still intact and the heatsink wobble slightly but is firmly attached. I did intend to undo the pins and replace the heat transfer compound, but they look a bit fragile and I'm wary of making a bad situation worse by breaking one or both.

Anything else I could try?

I have the original (GT120?) graphics card, so I could boot with that and see if I get any additional diagnostics.

I could try reverting to the original CPUs, I'm getting quite handy at swapping out the heatsinks, but still approach it with a bit of trepidation. Does the fact that it gets as far as running the fans and the hard disk mean the CPUs are ok?

Further update, I noticed the heatsink for CPU A was much warmer than CPU B which was only just warm to the touch, so I swapped the CPUs, resulting in no effect. The new CPU A is still far warmer.

Thanks.
 
Last edited:
SteelBadger

You can use the "Nut & Bolt technique but the spring should have approx the same locking strength as the original . .. or use the original rivet springs.

This is how my spare 4,1 tray is set up.

Northbridge nut & Bolt setup 02.jpg
 
Well, after leaving it overnight it booted this morning and the new CPUs have fixed the audio glitch that started this all off. I'm going to leave it there, hopefully it's back in pre CPU upgrade state of requiring a couple of hours between boots, I've lived with that for years now even if it is a bit harrowing on every restart.

Thanks for all the help.

Definitely leaving it there. Hang on though, now it'll accept 1333Mhz memory won't it?
 
SteelBadger said:
I've tried resetting the NVRAM as suggested. No dice.

I've investigated the northbridge. The push-pins are still intact and the heatsink wobble slightly but is firmly attached. I did intend to undo the pins and replace the heat transfer compound, but they look a bit fragile and I'm wary of making a bad situation worse by breaking one or both.

Anything else I could try?

I have the original (GT120?) graphics card, so I could boot with that and see if I get any additional diagnostics.

I could try reverting to the original CPUs, I'm getting quite handy at swapping out the heatsinks, but still approach it with a bit of trepidation. Does the fact that it gets as far as running the fans and the hard disk mean the CPUs are ok?

Further update, I noticed the heatsink for CPU A was much warmer than CPU B which was only just warm to the touch, so I swapped the CPUs, resulting in no effect. The new CPU A is still far warmer.

Thanks.
Click to expand...
You should probably monitor the temps with MFC or similar and show some pictures of the temps. I would also recommend running AHT. That will certainly reveal any hardware issues.
 
SteelBadger said:
Well, after leaving it overnight it booted this morning and the new CPUs have fixed the audio glitch that started this all off. I'm going to leave it there, hopefully it's back in pre CPU upgrade state of requiring a couple of hours between boots, I've lived with that for years now even if it is a bit harrowing on every restart.

Thanks for all the help.

Definitely leaving it there. Hang on though, now it'll accept 1333Mhz memory won't it?
Click to expand...

Well now you've managed to boot it get some temperature monitoring on there and see what's going on.
 
amedias said:
Well now you've managed to boot it get some temperature monitoring on there and see what's going on.
Click to expand...

Northbridge Diode 72.C
MCH Heatsink 53.C

Nothing else above 40.C

If I compare the logs of a unsuccessful boot and a successful one then they diverge with the unsuccessful one having the entries...

May 28 23:29:40 Mac-Pro com.apple.xpc.launchd[1] (com.apple.netbiosd[231]): Service exited due to SIGHUP | sent by killall[232]
May 28 23:29:40 Mac-Pro com.apple.xpc.launchd[1] (com.apple.netbiosd): Service only ran for 0 seconds. Pushing respawn out by 10 seconds.

After that it looks like it never really recovers. Netbiosd never gets a mention is a successful boot.

Any clues there?
 
SteelBadger said:
Northbridge Diode 72.C
MCH Heatsink 53.C

Nothing else above 40.C

If I compare the logs of a unsuccessful boot and a successful one then they diverge with the unsuccessful one having the entries...

May 28 23:29:40 Mac-Pro com.apple.xpc.launchd[1] (com.apple.netbiosd[231]): Service exited due to SIGHUP | sent by killall[232]
May 28 23:29:40 Mac-Pro com.apple.xpc.launchd[1] (com.apple.netbiosd): Service only ran for 0 seconds. Pushing respawn out by 10 seconds.

After that it looks like it never really recovers. Netbiosd never gets a mention is a successful boot.

Any clues there?
Click to expand...
I wonder if it did not detect your HDD /SSD and tried booting from the network (and of course did not find anything) and shut down
 
How long had it been running at that point and what’s the ambient temp?

If you do something load intensive for a while what kind of temps then?

72c is high but not outrageously high for a dual CPU 4,1/5,1, especially if it’s been on for hours in a warmish ambient, or under load, but if that is minutes after booting then possibly concerning.

For ref, I have 3 single CPU 4,1/5,1s and although their NB will never run as hot as a dual I have never seen a NB diode temp over 70 even when it’s been running at near 100% load for >24hrs. Normally my NBs would be in region of 50-55c at idle and 60-65c under full load.

I’m still suspicious of PSU as I would expect an overheating NB to cause issues in use, not the weird rest-recover cycle you’re experiencing.

Does it sleep and wake from sleep OK?

If it were one of my machines and I didn’t have another to start playing swapies with trays and PSUs I think I’d start by removing all non essential components and testing with just a single SATA HDD/SSD and a low power OEM graphics card like a gt120 to see if it behaved itself under lower power draw conditions.

I might even start testing with single CPU on the tray...

I think there is an Apple service doc somewhere detailing PSU test procedure to make sure you’re getting the right voltages on the right lines, might be worth having a dig about.

Know anyone local with a similar model that you can play tray/backplane/psu swapsies with?
 
The 72.C is after running for 12 hours, at very light loads.

I think the stability of the northbridge is ok, it's pretty solid. I wanted it to be that as it's something I could fix. I agree that an overheating northbridge should lead to intermittent crashes, instead once booted it's rock solid.

Sleeping is no issue, neither is the 'soft' part of an OS update where it does a kind of restart without the chime followed by a progress bar. However for the second part of the OS install where it does a proper reboot with the chime I have to turn it off and leave it.

When the problem first started 3 or 4 years ago i stripped everything right back, removing everything I could and booting from a USB stick. For 3 solid days over one Easter I did little more than change things and reboot or leave it on stuck progress bars, all with no result. It was only when I finally gave up and switched it off and left it that it finally came back to life.

I don't know anyone else with a machine for swapping parts. I did think that for a ten year old machine I could probably pick up spares for pennies, but a quick trip to Ebay disabused me of that opinion.

I'm inclined to agree that it could be something to do with the power supply, I'll have an inspection next time it's turned off for an OS update and blow some compressed air at it.

Thanks everyone for all the help.
 
Last edited:
amedias said:
72c is high but not outrageously high for a dual CPU 4,1/5,1, especially if it’s been on for hours in a warmish ambient, or under load, but if that is minutes after booting then possibly concerning.
Click to expand...

4,1>5,1 dual sits consistently on 77c at idle, with a system ambient of 28c, and only takes a couple of minutes to reach that after wake.
 
  • Like
Reactions: owbp
mattspace said:
4,1>5,1 dual sits consistently on 77c at idle, with a system ambient of 28c, and only takes a couple of minutes to reach that after wake.
Click to expand...

I don't own a dual any more but mine was always ~62-65c at idle and ~75-78c under heavy load, but ambient in my office is normally 19-21c which would account for the difference. I always found the duals to run ~10deg hotter NB than the singles.
 
MIKX said:
My 4,1 is " stock "

Adding a USB fan or a PCIe card makes it non-stock ?
Click to expand...

Adding additional cooling fans to augment the factory fans counts as not stock in my eyes.

I’m not saying it’s a bad idea, just that for the purposes of evaluating the OPs running temp, comparing it to an non-standard setup doesn’t really help him, same problem as comparing temps with fans idling at 600rpm vs 2000rpm.
 
  • Like
Reactions: crjackson2134
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back