Mac Pro 3,1 black screen then fans on full

Have bought (from reputable business supplier servicing UK AV industry) another PSU and once fitted experienced exact same issues so that’s highly doubtful. They think it’s a hardware issue (due to years of heat shock) with logic board. Another co that has a technician who deals with backplanes for defence industry and has experience with these exact models cites frequent issues With the 3,1 motherboards. What do you think? A mobo strip out and replace is extreme but with tested items from reputable dealers worth a try? IF you agree how would you source a decent Mobo? Anyone?

Hi Grant Thanks again for swift reply

Yes I think you are correct; it's a distinctly different sound to that the dvd drives make when they are searching for an AHT disc ;-) ... it's not like arcing electricity or anything frightening but I'd bet $$ it's in that PSU / top of mobo area. I had the PSU fully out of the chassis earlier today on the strip down and there was way more dust than I would have expected given it's only about a year since I last stripped it and ran a can of air through it (and we have had a robovac thing for most of the year that keeps the dust down daily).... and its fan is pulling air in 20cm above the floor.

PSU location 1 is running at 40degC and PSU location 2 at 42degC. Hottest thing on the rig is one of the Northbridge chips at 62degC - the 2nd is 51degC ....everything else is running cooler sub 40degC.

But looking at the fuzz and dust that was around the PSU I would not be surprised if it is end of life :(

How reliable are the diagnostic LEDs ? - the 'power rail fail' definitely showed up within a second of that click (I was waiting ready with my plastic stick to press the diag button) and there was no bad news from the two CPU ones at any point whilst testing.

Thanks

Andrew

further - I have tried it with 'everything out' and it still fails similarly after some time :-

I have run it headless, remoted in with VNC viewer, i.e. no GPU (I also tried this with no GPU nor optical drives, & there's nothing else in PCIe slots)

on boot I could use VNC to get into it and log in but the same crashes happened after similar random periods as short as a couple of minutes, to as long as an hour or more (even when the machine was in sleep mode - my son being in the room and knowing the drill ... "to power it down when the fans come on loud" and let me know it had crashed and when

then with just one boot HDD, i.e. only one sledded spinning HDD drive, and then logically going through the memory risers each in turn, with just one stick of RAM (yes I tried each in sequence - even though the memory board red lights flash correctly once upon boot then go off) .... still the same crashes

Hi Grant - as I have made a couple of replies inadvertently to myself you might have missed the developments here but other than that please note the original boot drive now mounted in my new 5,1 doesn't show partitions within volumes even after repairs, .... and opencore cannot see it as a bootable volume (it's listed in bootpicker, but selecting it for startup disk gives a no entry icon)

ideas ?

many thanks

Andrew

ps having spoken to an Apple hardware specialist technician (who I got in touch with to get the original PSU revamped) I have discovered that the logic boards on these frequently get heat shocked and once they reach operating temps the failures I have been experiencing occur without any logs (as it is an immediate power fail shutdown - something goes over the rated temp and system protection kicks in, power off, fans on full) - having tried two PSUs in this with the exact same result I am inclined to agree - what do you think ? are there further OS based avenues to diagnose ? - I have the ASD tool for this exact machine (ASD-3S123.zip)

And more importantly is there any way to make those Yosemite HDDs bootable and usable in the new 5,1 ?

at the very least I would wish to recover the Photos.app library which having been rebuilt from the main drive (Yosemite 2) is missing thousands of pics, but hopefully the timemachine backup on Yosemite HDD from the original 3,1 which is recent enough will be able to rectify that (not sure how to use it in the new machine though).

Hi Grant

Thanks for the reply. - Yes that makes sense to me and I think I've discovered the missing element; El Cap does not have the necessary Kexts / drivers etc for the GPU card in that machine (an OTT 8GB radeon monster). Minimum I think is Mojave.

Does that sound correct?

PS how do I boot into the 3,1 with the ASD boot USB I have prepared (the 3,1 has unflashed non Apple GPU - an old 512 nvidia so I am getting NO boot picker screen) ?

I have tried guessing and using the L/R arrows and return key after 5 mins to no avail ....

thanks

Andrew

additionally - I have fully recovered the photos library from another backup early December when the old Mac 3,1 started acting up (and added recents from icloud) - I only mention this as now there is no urgent need to focus on that from my perspective.

Every time I put any bootable OS drive in the 3,1 and leave it with no load the problem will occur - it happens after a few minutes or a couple of hours but the machine has never stayed on for more than 2hrs 45mins.

FYI In case I hadn't mentioned it above, I have eliminated the PSU being the source of the problem by repeating the same issue with a known good PSU (6 hours bench tested by Apple authorised service technicians who were good enough to refund me the PSU cost once I returned it).

What I would like to do, and you may be able to advise on this, is use the ASD-3S123 USB (with EFI and OS boot partitions) to more thoroughly diagnose (what I suspect to be) a logic board / logic board component issue (perhaps even the thermal paste of the CPUs - although I have never seen the service_diag LEDs light up for the CPUs and monitoring has given the temps as 49degC to 63degC)

My issue currently is that I cannot get the 3,1 to boot from that USB (this is made trickier by the fact that the GT 9500 GPU is not flashed so I get no bootpicker screen)

Any ideas ?

thanks again

Did you ever get any crash report logs that pointed towards hardware at all? And I assume you did the work swapping out the motherboard and reseating the thermally pasted components ie the processor heatsinks yourself?

I know what you mean - in spite of the excellent modular design of these cMPs that's quite a bit of precision labour to get to the backplane

So - I'd love to logically go through this and compare notes as what I have found (when known good Mobo swapouts don't instantly fix the shutdowns) leads to suspicions around the controller chips on 3,1s

& this is going to sound weird but ....I'm not beyond ruling out UEFI bootkit malware (intended for Intel EPROM & Xeons in Wintel architecture) somehow causing damage to firmware/controllers at this point

Yours is a 3,1?

Unaltered/bios unflashed/no opencore?

& It initially still kept crashing (but less so) after the mobo swap ?

then you worked on the GPU Nvidia & CUDA versions to get the right ones

(for which MacOS version?)

Sorry for the extra questions Sailordoom but you are probably the ONLY person who has sussed this and whilst I appreciate you live in fear (all old mac users do ;-) - it's like owning an old Healey 3000 vintage sportscar !) AND I know you don't want to jinx it, you just might have found the answer for everyone on the hardware/firmware side ....

PS Grant (kindly responding to most threads here) is OS uber-expert level at those logs and a lot more - did you ever get to ask him to cast an eye?

PS - As I no longer have the Apple Diagnostics Cd for this box I have found a copy of ASD-3S123 online, so can run that too - if you think it can help inform?