How to Change Startup Drive when no Startup Screen

[MIKX]

On my Dual Quad 2,93Ghz 4,1>5,1 the CPU heatsink bolts have blue 'thread lock ' paint on thm which impedes finding the correct bolt tension, number of turns etc.

To counter this ( on my spare CPU tray ) I completely 100% shielded the CPU tray leaving only the heatsink bolts exposed and also put tape around the bottom of the bolts . . then I used a new wire brush drill attachment and very carefully removed the thread lock paint on all 8 bolts.

Then with the CPU tray suspended upside I carefully wrapped the top of all 8 bolts with clean cloth strips soaked in a little "Eucalyptus Oil " which emulsifies most adhesives ( I'm an Aussie ).

With the 'thread lock' paint removed I can now more accurately " feel " when the heatsink starts 'feeling' correctly tightened.

NOTE : ' Too loose' is way better than ' too tight'.

It was well worth the effort. I now have a lot more confidence when installing de-lidded CPUs.

Previously, last year I had to buy a very expensive refurbished Dual CPU tray as the original tray's CPU B appeared to have died.

Can't post a pic as both CPUs on this spare CPU tray are in place and working.

 

[gadgetfreak98]

gadgetfreak98

I feel for you.

Suggestions.

0. Of course you have already checked your current power connection to see if ANY
other electrical device will work from it. eg: Desk lamp.
===========================================

1. It " might " be than if you have removed CPU A to inspect the Northbridge that
your subsequent " re-tensioning" of the CPUA heatsink bolts is either over or
under tight.
To check . . . remove CPU B. then experiment with re-tensioning CPU A with only
one pair of identical RAM modules.

1.1. Are your Northbrige heatsink rivets unbroken ?

1.2 Have you cleaned your Northbridge & re-applied Thermal compound recently ?
if so, check to see that none of the thermal compound has spread out onto
surrounding electrical components

2. Remove power cord and remove & re-insert your PCIe M.2 PCIe adapter.

3. Correct SMC & NVRAM re-set =
First, remove the cMP's power cord for a minute = SMC reset.
NVRAM reset = on my 4,1>5.1 starting up cold with P+R+ Apple key + Option key
one cycle takes around 30 ~ 35 seconds = you should hold down the NVRAM keys
for at least 35 seconds. If no chime report here.

4. Remove the cMP side cover and watch for any red LEDs in the CPU tray as you
press the power button. Report results.

5. Remove ALL PCIe cards including GPU. Try a startup - report results here.

I realise that some of the above has already been suggested. I would add them to this list and print it out from whatever PC you are using to read this thread. This will allow you to 'tick off " which techniques you have attempted.

Personally I went through this kind of situation last year and I'd start with #1.2 first and then #1 in the above list first.

Don't panic work slowly methodically. Every problem has a solution.

Best of luck to you

So, it took me a while to have time to dig in, but here’s where I am.

1. I removed CPU B altogether as suggested.

2. The Northbridge heatsink rivets are not broken. I have not ever removed the heatsink. Should I be doing this since (spoiler alert) I’m still not up and running?

3. I re-seated CPU A with new thermal compound again, after a thorough cleaning. To be clear, these are de-lidded CPUs which had been working fine for several months.

4. Did the SMC and NVRAM reset attempt, but no chime, ever after holding down the keys for a long time. Nor am I able to get the caps lock key on the Apple keyboard to light up.

5. I removed all PCIe cards (had originally, per #2 above, re-inserted the PCIe M.2 adapter).

6. I see the one red LED near CPU A light up briefly when I hit the power button. Am I looking at the wrong thing?

7. The main fan speeds up to very loud levels 30 or 40 seconds after hitting the power button.

8. I have tried various attempts to tighten or loosen the heatsink screws, but to no avail.

9. Under #1, it wasn’t clear to me whether you were suggesting the pair of RAM DIMMs be in slots 1 and 2 or in slots 1 and 5 as per another post. I tried both.

10. I also tried the other CPU to confirm there’s not an issue with the one that was in the CPU A slot.

So after all of this, I’m still stuck and no chime. I guess my primary next step questions are:

1. Should I be trying to do something with the Northbridge heat sink? I have ordered replacement clips via eBay, but they won’t arrive for a bit. The current ones, however are not broken, from what I can tell. Should I try to gently remove the clips and clean and re-thermal past?

2. Am I thinking about RAM the wrong way?

3. Is it possible to try CPU and Heatsink B instead of A, or is A always the right one when you are only using a single CPU?

4. Any other suggestions?

Thanks for all the help!

 

[MIKX]

No. 6 That is normal

No. 9 Personally I would put one pair of identical ram modules.
There are tiny LEDs next to each RAM slot, keep an eye on those.

CPU : You could try putting the CPU B chip into CPU A with minimal RAM
and remove CPU B & it's heatsink.

Northbridge :

You can try the " Nut & Bolt " technique but bear in mind that the springs on the rivets are there to allow for expansion/contraction as the Northbridge chip gets very hot,

The above setup works with my original CPU tray.

Last year I ended up buying a re-furbished dual cpu 4,1 tray with 32gb ECC ram. Wasn't cheap but definitely worth it - my 4,1 now performs like a new machine and cooler too as I have put USB fans in to further cool the Northbridge.

Realistically, 4,1 cMPs are now 10 years old, there's no telling what damage has been done to whatever components whether from overheating, dust, accidents etc.

I highly recommend removing your PSU - thoroughly de-dust it but do not open it.

Please PM me re: your progress.

 

[gadgetfreak98]

Tried to PM you per your suggestion. Not sure I did that right tho. Let me know if you didn’t get a message from me separately. Thanks

 

[MIKX]

Replied to your PM.

 

[gadgetfreak98]

So just for the general audience, in case others have ideas too:

I have in various combinations:

1. Tried tons of RAM DIMM configurations
2. Removed the Northbridge heatsink and applied new paste, etc.
3. Tried two difference CPUs in Tray A
4. Mostly removed the PSU and blown out a whole bunch of dust
5. Tried many different tensions on the heatsink bolts in A
6. Not used the CPU or heatsink on B, to isolate that out
7. Removed all PCIe cards and all drives (including optical)
8. Or, only had the GT120 card in
9. Tried re-setting NVRAM and PRAM.

Results:

1. I am getting no chime under any circumstances.
2. Sometimes upon hitting power, I get the fans spinning up to max and loud, sometimes not
3. Sometimes the power switch immediately powers everything off, sometimes I have to hold it down
4. I get no image on my 30” ACD

Welcome any further suggestions!
[doublepost=1557675962][/doublepost]Also, in case it means anything to anyone, the computer seems to turn itself back on sometimes... e.g., if the power cord is plugged in, it will try to power back on and then the fans spin up out of control.

 

[gadgetfreak98]

Well, I was able to come back to this finally late tonight. I found the proper service manual to figure out what I was doing wrong with removing the power supply. I did get it out, blew air thoroughly through it, reinstalled and still at the same place. No chime, no image on screen. So would best guess be power supply replacement? Any other next steps to suggest?

 

[tsialex]

Well, I was able to come back to this finally late tonight. I found the proper service manual to figure out what I was doing wrong with removing the power supply. I did get it out, blew air thoroughly through it, reinstalled and still at the same place. No chime, no image on screen. So would best guess be power supply replacement? Any other next steps to suggest?

Your Mac Pro seems much more a brick than having a dying power supply. What diagnostics led are lit?

 

[gadgetfreak98]

None normally. Per the service manual, the CPUA and CPUB flash momentarily upon being plugged in. When I hit the little white button, the 5V STBY is illuminated. What else should I be doing or looking for?

Oddly, the fan keeps spinning up to full blast shortly after turning the power back on, fwiw.

And also, I did again clean the Northbridge heatsink (and replaced one of the connectors), reseated the CPUA and have tried a whole bunch of different tension levels on the bolts (making sure not to get too tight). Just in case something was slightly off there. But nothing has changed.

 

[tsialex]

None normally. Per the service manual, the CPUA and CPUB flash momentarily upon being plugged in. When I hit the little white button, the 5V STBY is illuminated. What else should I be doing or looking for?

Oddly, the fan keeps spinning up to full blast shortly after turning the power back on, fwiw.

And also, I did again clean the Northbridge heatsink (and replaced one of the connectors), reseated the CPUA and have tried a whole bunch of different tension levels on the bolts (making sure not to get too tight). Just in case something was slightly off there. But nothing has changed.

If EFI_DONE led don't light when you press the diagnostics button, you have a brick.

Check page 30 of the Service Manual:

 

[gadgetfreak98]

But which one is EFI_DONE? I see it draws reference to that, but on p. 29 of the service manual, I don’t see one labeled as such.
[doublepost=1558242500][/doublepost]

If EFI_DONE led don't light when you press the diagnostics button, you have a brick.

Check page 30 of the Service Manual:

View attachment 837863

Sorry, meant to quote you:

But which one is EFI_DONE? I see it draws reference to that, but on p. 29 of the service manual, I don’t see one labeled as such.

 

[tsialex]

But which one is EFI_DONE? I see it draws reference to that, but on p. 29 of the service manual, I don’t see one labeled as such.
[doublepost=1558242500][/doublepost]

Sorry, meant to quote you:

But which one is EFI_DONE? I see it draws reference to that, but on p. 29 of the service manual, I don’t see one labeled as such.

 

[gadgetfreak98]

View attachment 837865

Is this what you would expect?
[doublepost=1558243651][/doublepost]sorry for the poor picture... And didn’t realize that those lights were there too, as my finger was initially blocking them

 

[tsialex]

View attachment 837871

Is this what you would expect?
[doublepost=1558243651][/doublepost]sorry for the poor picture... And didn’t realize that those lights were there too, as my finger was initially blocking them

The first one should be lit too, SYS_PG.

Not your case but for future reference, if you don’t have an Apple original GPU, GPU O.K. keeps off.

 

[gadgetfreak98]

The first one should be lit too, SYS_PG.

So I am doing this without any CPU installed at this point, following the diagnostics in the service guide. Would that mean the SYS_PG should not illuminate? According to the guide, at this point:

Expected behavior: • PSU PWROK (green), 5V STBY (amber), EFI DONE (green), GPU OK (green) LEDs illuminated

I don’t see them suggesting SYS_PG. I do have the 4 noted above lighting up. So two questions:

1. If no CPU, nor RAM is installed is it ok that SYS_PG is not illuminated?
2. If NOT ok, any suggestions as to what that indicates?

Thanks!

 

[tsialex]

You have to test with CPU tray and with a Xeon installed in socket A plus memory installed in one or more of sockets 1,2,3,4.

With this config SYS_PG, EFI_DONE and GPU_OK should be lit.

 

[gadgetfreak98]

Ok, I’m going to try by putting one of my old CPUs back in... maybe, just maybe, the new ones are the problem???
[doublepost=1558245702][/doublepost]So now, with a new original CPUA in, I do NOT have the EFI_DONE light, but SYS_PG is on. So does this mean something maybe up with the CPU A heatsink or something like that? I tried it once slightly more tightened, and once slightly more loosened.