Resolved Mac Pro Server (Mid 2012) strange antics powering up

[JedNZ]

Just bought another cMP 5,1 Server (Mid 2012) Quad-Core (single CPU). When I first plugged the power cable in it started automatically before I even hit the power button. All good, so I upgraded the firmware from the original 7F to 89, then popped a RX 580 in and upgraded to 144.0.0.0.0, but didn't install Mojave.I then put the 5770 back in because I'm running High Sierra (note: this is not the same cMP as in my sig). All went well, until I installed High Sierra (10.13.6 - latest version) on one of the internal spare hard drives. When setting up the initial User account, I set my home country and the machine immediately powered off.

And then there was no sign of life. The power button would not activate. Dead. No fans gently whirring. No hard drive clatter. Nothing! When I plugged the power cable in I could see the top drag LED flash once. If I pushed the diag button in the 5V STBY LED shines orange while I hold it down.

I removed the CPU tray and checked for dust bunnies, checked that the Northbridge rivets look intact, and checked for no bent pins on the backplane. I removed and reseated all 4x 2GB ram modules. I removed GPU (there are no other PCIe cards), and all the HDDs. I removed and inspected the PSU - nothing out of place - fairly clean. I removed the PRAM battery - no change - still dead. I put the battery back in - still no change. No matter what I did the cMP is simply dead, except for some signs of life with the diagnosis LEDs when I plug the power cord in (flashes red) or hit the diag button (5V STBY shines amber).

This morning I decided to replace the PRAM battery after reading that as a suggestion on a thread in here. Already depressed by the disappointment of a dead cMP, I pressed the power button not expecting anything, and woe and behold it POSTs and it's running again. I've since done three more shutdowns/restarts and no issues at all that I can tell. See iStat Menus Sensors snapshot below.

Couple of questions:
1. I see many references about leaving the cMP to sit for an hour or more to rest the PSU if it fails to start up. Do the observations above suggest a falling PSU?
2. Could a replacement, brand new CR2032 battery make all the difference? I mean, could a failed/dead battery even stop the cMP from POSTing and powering up? Do the settings for country get saved to PRAM, and could a dead/faulty battery therefore cause issues?
3. What other thoughts/troubleshooting do you suggest I do to test my cMP?
4. What else do you think might be going on?

 

[tsialex]

PSU dying causes this type of weird behaviour, happened to me and after some investigation I found two bulged capacitors on the PSU PCB.

Open the PSU, take a look for any leaking or bulging caps, clean it and then move it to your main Mac. Use it for some days and see if it's working correctly. Test your new Mac Pro with a know working PSU. This is what I do to track weird power problems.

Mac Pro 4,1 or 5,1 don't have PRAM. PRAM is a relic from 68K and PowerMacs. MP4,1/MP5,1 have NVRAM and it's a volume inside the SPI flash that stores the BootROM.

The battery only power the Real Time Clock chip (RTC), it has a 256 byte memory that stores only date related parameters, like time and time since boot - most of it is unused. This is a example of a RTC memory dump, made by DarwinDumper:

CMOS Dumper V0.1.:)
DSDT has RTC Length of 4. Dumped 256 Bytes.
--------------------------------------------------
  0  1  2  3  4  5  6  7  8  9  A  B  C  D  E  F  |
--------------------------------------------------
0|37 14 45 42 21 16 04 11 04 13 26 02 50 80 c9 00 |
1|eb f7 59 ee ff 7f ac 79 bf f8 ef 3b bf fa ef fe |
2|9f fe ff fb bf bf 85 ff 7f ff d4 bc ff fe bf 5b |
3|3f 73 20 6f bf ff 7e 6f bf e9 fe f7 ff 00 21 16 |
4|00 00 00 00 fe 00 00 00 00 00 00 00 00 00 00 00 |
5|01 3f 80 00 01 00 01 be 2e f3 80 00 00 00 00 00 |
6|54 4f 4f 42 00 00 00 02 00 00 1f 02 00 00 00 00 |
7|00 00 00 00 00 00 00 00 00 00 00 00 e6 fe 6e fd |
8|00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |
9|00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |
a|00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |
b|00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |
c|00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |
d|00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |
e|00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |
f|00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |
--------------------------------------------------
By observtion of the CHANGED bytes from Sleep/Wake Cycles
You MAY get some idea as to why CMOS RESET bug exists!!!!

A BR2032 battery below 1,9V can cause weirdness like wrong time, but AFAIK can't make a MP4,1 or MP5,1 un-bootable, you can boot even without a battery. Btw, you have to replace the RTC battery after the voltage drops below 2,6V.

Don't use a CR2032 with a Mac Pro, the heat from the GPU will make a CR2032 fail within 6 months and the chemistry formulation is not really adequate for continuous low discharge done by the RTC chip.

 

[JedNZ]

Thanks for that - makes sense to me now because my new cMP shutdown after a couple of hours while I was away from it. I pushed the power button to turn it back on and heard a clunk sound from the area of the PSU and there’s no sign of life again. So I agree with your diagnoses that it must be the PSU and therefore I can disregard my concerns about the backplane and CPU tray.

I’ll do as you suggest and swap my PSUs over and see how that goes. I haven’t been able to find the BR2032 battery in NZ so have always resorted to buying the CR2032.

That’s really interesting info you’ve provided @tsialex - really appreciate your willingness to share your knowledge and learning so much from you. Nga mihi nui (thank you)

 

[Snow Tiger]

Just bought another cMP 5,1 Server (Mid 2012) Quad-Core (single CPU). When I first plugged the power cable in it started automatically before I even hit the power button. All good, so I upgraded the firmware from the original 7F to 89, then popped a RX 580 in and upgraded to 144.0.0.0.0, but didn't install Mojave.I then put the 5770 back in because I'm running High Sierra (note: this is not the same cMP as in my sig). All went well, until I installed High Sierra (10.13.6 - latest version) on one of the internal spare hard drives. When setting up the initial User account, I set my home country and the machine immediately powered off.

And then there was no sign of life. The power button would not activate. Dead. No fans gently whirring. No hard drive clatter. Nothing! When I plugged the power cable in I could see the top drag LED flash once. If I pushed the diag button in the 5V STBY LED shines orange while I hold it down.

I removed the CPU tray and checked for dust bunnies, checked that the Northbridge rivets look intact, and checked for no bent pins on the backplane. I removed and reseated all 4x 2GB ram modules. I removed GPU (there are no other PCIe cards), and all the HDDs. I removed and inspected the PSU - nothing out of place - fairly clean. I removed the PRAM battery - no change - still dead. I put the battery back in - still no change. No matter what I did the cMP is simply dead, except for some signs of life with the diagnosis LEDs when I plug the power cord in (flashes red) or hit the diag button (5V STBY shines amber).

This morning I decided to replace the PRAM battery after reading that as a suggestion on a thread in here. Already depressed by the disappointment of a dead cMP, I pressed the power button not expecting anything, and woe and behold it POSTs and it's running again. I've since done three more shutdowns/restarts and no issues at all that I can tell. See iStat Menus Sensors snapshot below.

Couple of questions:
1. I see many references about leaving the cMP to sit for an hour or more to rest the PSU if it fails to start up. Do the observations above suggest a falling PSU?
2. Could a replacement, brand new CR2032 battery make all the difference? I mean, could a failed/dead battery even stop the cMP from POSTing and powering up? Do the settings for country get saved to PRAM, and could a dead/faulty battery therefore cause issues?
3. What other thoughts/troubleshooting do you suggest I do to test my cMP?
4. What else do you think might be going on?

View attachment 895235

Well , when you had your PSU out , I would have blasted all the dust out of it with a compressor with 60 PSI at 6 inches . You'll need to let it dry for 12 to 24 hours afterwards , though . Dust in the PSUs can cause a lot of problems and I cannot even tell you the number of times just a simple tear down and cleaning of all the Mac's components can bring a dead or unstable Mac back online after rebuilding .

Alex is right about ten year old PSUs possibly having bad capacitors . Although it must be pretty rare . I've actually never seen this with these PSUs and I've torn them down .

Out of curiosity , what are the diagnostic LEDs when you push the on-board button at start up ?

Upper bank :

OVTMP CPUA
OVTMP CPUB
PSU PWROK
5V STBY
3.3V ( No light installed )
1.5V ( No light installed )
1.05V ( No light installed )


Lower bank :
SYS PG
PLT RS ( No light installed )
X1 PG ( No light installed )
EFI DONE
GPU OK
5V ( No light installed )
RSM RST ( No light installed )

During normal operation , the PSU PWROK is green , the 5V STBY is amber , the SYS PG is green and the EFI DONE is green .

 

[Snow Tiger]

PSU dying causes this type of weird behaviour, happened to me and after some investigation I found two bulged capacitors on the PSU PCB.

Open the PSU, take a look for any leaking or bulging caps, clean it and then move it to your main Mac. Use it for some days and see if it's working correctly. Test your new Mac Pro with a know working PSU. This is what I do to track weird power problems.

Alex , should the OP worry about the PSU caps discharging into him accidentally ? This actually happened to a tech friend of mine with cMP PSUs . Knocked him flat off his feet .

I have professional discharge tools in my shop , but there is no provision to discharge the caps in these PSUs since the component leads are covered with plastic rather solidly .

I just let the caps auto discharge after around 24 hours . These caps are designed to auto-discharge , right ? I've never gotten zapped by them , thank heavens !

To others reading this , once a PSU has been removed from a System , it should auto-discharge its caps as a safety feature . They recharge every time you plug the power cord back into the PSU .

Don't use a CR2032 with a Mac Pro, the heat from the GPU will make a CR2032 fail within 6 months.

I hope this is not true . I've been replacing the old factory BR2032 batteries with brand new CR2032 since I haven't noticed any PCIe slot thermal issues with my Builds . I've been doing this operation for years with my high end Builds and not one of my clients has reported a RTC related issue , so I think the batteries are OK to use .

I just replaced the battery in one of my Builds with a "CR" version . It is absolutely jammed packed with high end components that could really heat things up if I didn't use MFC with customized settings based on a lot of different thermal sensor readings , including some from an IR thermometer gun .

 

[JedNZ]

More great info. Thanks for that. So it's not working at all, but if I plug in the power lead I get the brief flash of red on either the OVTMP CPUA or OVTMP CPUB - not sure which one because it's so fast. It's only a single CPU machine. And if I then hold down the drag button I get an amber LED on the 5V STBY. No other LEDs show. Unfortunately I didn't look to see what LEDs were lit up when it was actually running.

I'll go and clean out the PSU (it looks very clean anyway, so maybe the guy that sold to me gave it a clean) after I first test whether my working cMP PSU works in this dead cMP. I've bought some compressed air duster in a can. Is that okay to use? What other precautions do I need to take? Can I use a static free brush to help loosen up any heavy dust?

 

[Snow Tiger]

More great info. Thanks for that. So it's not working at all, but if I plug in the power lead I get the brief flash of red on either the OVTMP CPUA or OVTMP CPUB - not sure which one because it's so fast. It's only a single CPU machine. And if I then hold down the drag button I get an amber LED on the 5V STBY. No other LEDs show. Unfortunately I didn't look to see what LEDs were lit up when it was actually running.

I'll go and clean out the PSU (it looks very clean anyway, so maybe the guy that sold to me gave it a clean) after I first test whether my working cMP PSU works in this dead cMP. I've bought some compressed air duster in a can. Is that okay to use? What other precautions do I need to take? Can I use a static free brush to help loosen up any heavy dust?

I wouldn't worry about any LEDs that briefly flash at System start up , System shut down , plugging in or removal of the power cord . That's normal .

I use clean bristle paint brushes to clean components . Never had an issue with static discharge in 15 years .

 

[tsialex]

Alex , should the OP worry about the PSU caps discharging into him accidentally ? This actually happened to a tech friend of mine with cMP PSUs . Knocked him flat off his feet .

I have professional discharge tools in my shop , but there is no provision to discharge the caps in these PSUs since the component leads are covered with plastic rather solidly .

I just let the caps auto discharge after around 24 hours . These caps are designed to auto-discharge , right ? I've never gotten zapped by them , thank heavens !



I hope this is not true . I've been replacing the old factory BR2032 batteries with brand new CR2032 since I haven't noticed any PCIe slot thermal issues with my Builds . I've been doing this operation for years with my high end Builds and not one of my clients has reported a RTC related issue , so I think the batteries are OK to use .

I just replaced the battery in one of my Builds with a "CR" version . It is absolutely jammed packed with high end components that could really heat things up if I didn't use MFC with customized settings based on a lot of different thermal sensor readings , including some from an IR thermometer gun .

Eletrolitic capacitors can hold hazardous charges for a long time, you always have to take care with them. MP5,1 PSUs don’t expose the back of the PCB, so you can clean it up more or less safely.

About BR2032 vs CR2032, if wasn’t needed why Apple uses it?

A BR2032 costs 5x more. BR2032 are appropriate for long duration of slow amperage discharge and have a chemistry adequate for higher temperatures, a must below GPUs. CR2032 are appropriate for fast discharges with relative high amperage in comparison, the chemistry formulation have slow tolerance of high temperatures.

 

[Snow Tiger]

About BR2032 vs CR2032, if wasn’t needed why Apple uses it?

Ahhhhh , because Apple has a long history of manufacturing Systems that are quiet as opposed to properly cooled ! So they use the higher quality batteries to mitigate the resultant higher temps .

All the components in my Builds I can test are within safe thermal operating ranges , so maybe that's why I can get away with using the CR versions of the batteries . I keep component load thermals as low as possible for enhanced durability .

 

[JedNZ]

The PSU looks very clean. I didn’t open it completely. It’s now “drying out”, so I popped the PSU from my other cMP in and no dice. Won’t startup - no sign of life. So not the PSU then?

I removed the CPU heatsink to check the Northbridge rivets properly. All looking good. Good travel - no signs of perishing. I cleaned the thermal paste off and cleaned the CPU and the heatsink with Arcticlean and prep. Cleaned the cpu tray with my anti static brush, put it back together. Still no sign of life. So not the CPU or CPU tray then?

My next step is to remove the backplane and check for any signs of damage and check all cables are seated correctly.

 

[MIKX]

Might help but first remove all PCIe cards except for your (supposedly working ) GPU.

 

[Snow Tiger]

Have you tried swapping in the CPU Tray of your factory MP 4,1 into your factory 5,1 ? The system fans will spin to max and be very loud due to SMC version mismatched components being used in the same System , but its not dangerous in the short run . It'll give us an idea if your factory 5,1 Tray is the culprit .

 

[krakman]

The PSU looks very clean. I didn’t open it completely. It’s now “drying out”, so I popped the PSU from my other cMP in and no dice. Won’t startup - no sign of life. So not the PSU then?

I removed the CPU heatsink to check the Northbridge rivets properly. All looking good. Good travel - no signs of perishing. I cleaned the thermal paste off and cleaned the CPU and the heatsink with Arcticlean and prep. Cleaned the cpu tray with my anti static brush, put it back together. Still no sign of life. So not the CPU or CPU tray then?

My next step is to remove the backplane and check for any signs of damage and check all cables are seated correctly.

Welcome to the world of mystery!

There are several threads in this forum, all with the same symptoms. Infact it happened to me after i took three 2009 mac pros out of storage. Each one died within 15 minutes of each other. Occasionally they will restart but then hard shut down after a short period of time. All my faulty macs show an amber 5v led. Changing tge PSU does not fix the problem.

The closest solution/suggestion we have is possible capacitor leakage or condensation causing corrosion on the back of the back plane. If it is condensation/corrosion causing the problem then you will need to remove the whole back plane and clean it with pcb cleaner. If it is capacitor leakage
You need to find which capacitors has died and solder in a replacement.

You might find getting a replacement backplane is the simplest solution

 

[JedNZ]

Yep tried my 4,1 tray. Power button does nothing so no fans or other signs of life.

 

[Snow Tiger]

Yep tried my 4,1 tray. Power button does nothing so no fans or other signs of life.

Try bringing your Mac into a more minimal configuration . Remove all SSDs, HDDs , PCIe Cards , Optical Drives . See if that will help boot your Mac . If that doesn't work , remove the internal Bluetooth and WiFi cards .

You should also remove all connected devices , like keyboards , mice , USB hubs , etc . A healthy Mac can start up in this configuration , although all you'll see is the white power light and system fans purring away .

Something has failed and is aborting your start up .

If all of this doesn't help , then I guess you have a failed logic board and you should replace it .

 

[JedNZ]

I've pulled the GPU (it's the only PCIe card), Wifi card, BT card, Optical drive bays, no USB peripherals attached (no mice, no keyboard, no USB sticks) and with just a boot drive installed. No go - no white LED on the power button - no whirring of fans or hard drive chatter - the GPU fan doesn't spin up either. No LEDs on the CPU tray light up either. I've tried with other working boot drives - still no go. Nothing happens when I press the power button. I have tried all of these things with two other power cables, which are confirmed working on my other cMP. I can't even reset NVRAM (Cmd+Opt+P+R). I also tried with and without the PRAM battery (which is brand new - albeit a CR2032 because I cannot source a BR2032).

The 5,1 CPU tray works in my 4,1>5,1 - I get the immediate rush of fans when I start up. All ram is registered too. System Profiler shows the 4,1 backplane
1.39f5 and CPU tray 1.39f11 mismatch. So I know the CPU tray isn't the problem.

I removed the backplane, inspected it and cleaned it. No obvious signs of damage. I also inspected the front panel (power button, front USB and FW ports) - again no obvious damage. I did a closer inspection of the CPU tray pins on the backplane - no obvious damage - no bent pins, no dust bunnies. I tried to start it with by jumping the two bottom-right jumpers on the drag area, but no sign of life. I tried this with the front port cable disconnected and connected - same result both times. Honestly, I've removed and re-installed the CPU cage and tray numerous times trying different options.

Anyways, I've put it all back together again and have tried two other CPU trays (albeit 4,1>5,1 - single and dual-CPU that are both working in other machines), plus I've tried a variety of different ram configurations - the ram works fine in the other machines. Still no go.

I'm at my wits end with this one. When I press and hold the diag button it still shows an amber LED for the 5V STBY. So this says the PSU is working and that the backplane is getting power. But I'm thinking it must be a backplane issue.

Any thoughts?

 

[Snow Tiger]

I've pulled the GPU (it's the only PCIe card), Wifi card, BT card, Optical drive bays, no USB peripherals attached (no mice, no keyboard, no USB sticks) and with just a boot drive installed. No go - no white LED on the power button - no whirring of fans or hard drive chatter - the GPU fan doesn't spin up either. No LEDs on the CPU tray light up either. I've tried with other working boot drives - still no go. Nothing happens when I press the power button. I have tried all of these things with two other power cables, which are confirmed working on my other cMP. I can't even reset NVRAM (Cmd+Opt+P+R). I also tried with and without the PRAM battery (which is brand new - albeit a CR2032 because I cannot source a BR2032).

The 5,1 CPU tray works in my 4,1>5,1 - I get the immediate rush of fans when I start up. All ram is registered too. System Profiler shows the 4,1 backplane
1.39f5 and CPU tray 1.39f11 mismatch. So I know the CPU tray isn't the problem.

I removed the backplane, inspected it and cleaned it. No obvious signs of damage. I also inspected the front panel (power button, front USB and FW ports) - again no obvious damage. I did a closer inspection of the CPU tray pins on the backplane - no obvious damage - no bent pins, no dust bunnies. I tried to start it with by jumping the two bottom-right jumpers on the drag area, but no sign of life. I tried this with the front port cable disconnected and connected - same result both times. Honestly, I've removed and re-installed the CPU cage and tray numerous times trying different options.

Anyways, I've put it all back together again and have tried two other CPU trays (albeit 4,1>5,1 - single and dual-CPU that are both working in other machines), plus I've tried a variety of different ram configurations - the ram works fine in the other machines. Still no go.

I'm at my wits end with this one. When I press and hold the diag button it still shows an amber LED for the 5V STBY. So this says the PSU is working and that the backplane is getting power. But I'm thinking it must be a backplane issue.

Any thoughts?

well , you could remove the boot drive ... we're looking for a component that is aborting your start up .

without actually accessing your Mac , I can think of only removing two other components . The front panel to logic board cable ( which also includes the power button function ) . The power button cable can also be disconnected separately . You'd need to start up the Mac by jumping the solder pads on the logic board .

Maybe removing the logic board and examining if there was a screw or some other metal object between the chassis and logic board shorting your Mac .

 

[Nermal]

I also tried with and without the PRAM battery (which is brand new - albeit a CR2032 because I cannot source a BR2032).

Looks like PB has them:

Buy the Panasonic BR-2032/BN 200mAh, 3V COIN Carbon Monofluoride Button Coin... ( BR-2032/BN ) online

pbtech.co.nz "Panasonic BR-2032/BN 200mAh, 3V COIN Carbon Monofluoride Button Coin type lithium batteries 20mm Diameter

www.pbtech.co.nz

 

[tommy chen]

your backplane has gone

 

[krakman]

Maybe removing the logic board and examining if there was a screw or some other metal object between the chassis and logic board shorting your Mac .

Or corrosion shorting out the pins on some component.

The other possibility is component failure.....

Maybe the Reed switch has failed in the PSU?. You can put the PSU from the faulty machine into another to test and eliminate this possibility.

Maybe a capacitor has died on the backplane, although you would need a diagnostic tool to measure the capacitance of each of the caps on the board to find which one it is.

I'm not an expert but my research into this problem has highlighted a common theme in that the most common fault in electronic devices is capacitor failure.

 

[Macschrauber]

Ive posted about the violet 220 uf caps of the backplane leaking. Got one board with same symptoms, changed the caps but no go. Cleaned the area with isopropyl of course. Maybe take the whole board in a isopro bath and brush it wildly nothing to lose there

 

[Snow Tiger]

Ive posted about the violet 220 uf caps of the backplane leaking. Got one board with same symptoms, changed the caps but no go. Cleaned the area with isopropyl of course. Maybe take the whole board in a isopro bath and brush it widly nothing to lose there

I think what would be a better choice here would be an ultrasonic cleaner tub . But the ones big enough ( 30 L ) for a logic board can be expensive . You can use distilled water instead of expensive detergent fluids . You also might want to buy a brand name product so it lasts more than 60 days . Skymen is a real high quality manufacturer :

Quality Industrial Ultrasonic Cleaner, Automatic Ultrasonic Cleaner FACTORY, Industrial Ultrasonic Cleaner manufacturer, Automatic Ultrasonic Cleaner from China

China leading provider of Industrial Ultrasonic Cleaner and Automatic Ultrasonic Cleaner, Skymen Technology Corporation Limited is Automatic Ultrasonic Cleaner FACTORY.

www.skymenultrasonic.com

Mechanical 2l 3l 4l 6l 10l 15l 22l 30l Ultrasonic Cleaner Heated Timer Industry | eBay

Ultrasonic cleaning is based on the cavitation effect caused by high frequency ultrasonic wave vibration signals in fluid. As a result, it is more convenient and efficient than hand wash, steam cleaners, high pressure water jet cleaning or other machines.

www.ebay.com

 

[JedNZ]

Looks like PB has them:

Buy the Panasonic BR-2032/BN 200mAh, 3V COIN Carbon Monofluoride Button Coin... ( BR-2032/BN ) online

pbtech.co.nz "Panasonic BR-2032/BN 200mAh, 3V COIN Carbon Monofluoride Button Coin type lithium batteries 20mm Diameter

www.pbtech.co.nz

Didn’t even think of them. I’ll order from them direct - Not too expensive either.

 

[tsialex]

I've pulled the GPU (it's the only PCIe card), Wifi card, BT card, Optical drive bays, no USB peripherals attached (no mice, no keyboard, no USB sticks) and with just a boot drive installed. No go - no white LED on the power button - no whirring of fans or hard drive chatter - the GPU fan doesn't spin up either. No LEDs on the CPU tray light up either. I've tried with other working boot drives - still no go. Nothing happens when I press the power button. I have tried all of these things with two other power cables, which are confirmed working on my other cMP. I can't even reset NVRAM (Cmd+Opt+P+R). I also tried with and without the PRAM battery (which is brand new - albeit a CR2032 because I cannot source a BR2032).

The 5,1 CPU tray works in my 4,1>5,1 - I get the immediate rush of fans when I start up. All ram is registered too. System Profiler shows the 4,1 backplane
1.39f5 and CPU tray 1.39f11 mismatch. So I know the CPU tray isn't the problem.

I removed the backplane, inspected it and cleaned it. No obvious signs of damage. I also inspected the front panel (power button, front USB and FW ports) - again no obvious damage. I did a closer inspection of the CPU tray pins on the backplane - no obvious damage - no bent pins, no dust bunnies. I tried to start it with by jumping the two bottom-right jumpers on the drag area, but no sign of life. I tried this with the front port cable disconnected and connected - same result both times. Honestly, I've removed and re-installed the CPU cage and tray numerous times trying different options.

Anyways, I've put it all back together again and have tried two other CPU trays (albeit 4,1>5,1 - single and dual-CPU that are both working in other machines), plus I've tried a variety of different ram configurations - the ram works fine in the other machines. Still no go.

I'm at my wits end with this one. When I press and hold the diag button it still shows an amber LED for the 5V STBY. So this says the PSU is working and that the backplane is getting power. But I'm thinking it must be a backplane issue.

Any thoughts?

What's the status of EFIDONE LED when you press the DIAG button? Install an Apple OEM GPU and then press the diag button while you power up. If EFIDONE is off, you probably have brick and not a really defective backplane.

Since you did 3 2 firmware upgrades before things went south, I'm thinking that is most probable that you have a corrupted SPI than a defective backplane. If you know your way around a hot air rework station, you can check if it's a corrupted SPI flash or not easily.

 

[JedNZ]

Where’s the EFIDONE LED? Do I need to have powered up to do this? Only the 5V STBY LED lights up when I press the diag button.

Im sure I only did the two firmware upgrades, from 7F to 89, then 89 to 144. I did System Profiler snapshots showing they worked.

i have no experience with hot air rework stations. I’ve only got a basic soldering iron and only have basic soldering skills only.
[automerge]1582400487[/automerge]
And I was using a OEM Radeon HD 5770 1GB with High Sierra. I only installed my RX 580 just to do the 144 upgrade, then put the 5770 back in. Maybe I should try putting the RX 580 in and seeing if it will boot with that.