MP 1,1-5,1 MP4,1 -5,1 Upgrade a Tale of Joy and woe

[BlueQube]

"So, I think the SMC still thinks the old chips are in and not letting the new chips run at that hotter temp so the fans kick in sooner, even though the chip temps are well within range."

my first post here so tried to go into some detail:

I have taken an early 2009 mac pro 4,1 tower with 2 x CPU 2.26gh intel E5520's and (flashed) to 5,1

Then added 2 x 3.4gh Westmere Intel Xeon X5690's de lidded new paste and not overtightened.
(www.mac-factory.de)
Then added 48G ram – all good and recognised not mixed run in triple.
1 x 500G nvme m2 to x16 pci slot – boots great
1 x brand new Radeon rx580 Sapphire pulse 8g DDRR5 – (metal compatible) no boot screen as expected, some little dips of the screen to black but all acceptable – recognised and working.
Fresh install of High sierra 10.3.6 then firmware update to 140.0.0.0 via Mojave installer
Then Mojave – all fine and working great.

Reset SMC
Reset the PRam

Issues: the machine boots and runs fine for most things photoshop – DTP – web all good.
Adobe AE, Pr anything that draws a bit of load and the fans creep up more than usual for a mac.

When I put the computer under a load, say rendering with all the cores the fans kick in and get pretty loud, on the motherboard a small red LED flickers in unison with the render then stays solid until the render is done, then goes out when the load is over, the fans still blowhard when the temps of both CPUs are in the acceptable 40c the fans eventually go back to normal quite. (thermal throttling????)
#images attached.

(ok looks like C4d R17 is not compatible with Mojave/high sierra on this build)

Thing is c4d R17 works fine on my other mac pro and laptop on high sierra –
However, the smc and boot rom are now on Mojave ones so that would seem to be the incompatibility to me.

Stopped testing with C4D and used AE and Pr – fans still creep up sooner than they should
Blinking red led not happening unless under very heavy load, but fans still creeping up and getting loud.
Solution so far: MACs fan control, you can turn the fans down and deal with the noise but it’s the red light that needs to be resolved rather than just the fans.
This forum post tell the same story I’m having but not resolved.
https://forums.macrumors.com/threads/red-light-showing-from-processor-memory-board.1718643/

this forum takes of the same chip upgrade but different fan issues.
https://forums.macrumors.com/thread...gh.2006884/page-2?post=23821823#post-23821823


Now luckily, I have another mac pro5, 1 not flashed; A 2010 3.4ghz 12core duel cpu – ie I am building the new one to be pretty much the same as this current one.

The difference being I run a macvids flashed card NVidia gtx 980 4g with boot screen (even though the efi spasms out on boot screen in high sierra she still works a treat.

Now the led flash I get on the flashed 1,4 2009, happens on the 5,1 2010, however it just flashes once or twice and then the fans pick up just a little (hardly any noise and the light goes out- its only very brief.

So thermals right???

Temps at idle: both machines are pretty much the same including fan noise. I’ve looked at the temp’s under load on both machines as a comparison: AE – with openGL being used on the card and effects to push the cpu’s – temps run a little hotter on the new de-lidded cpu’s X5690's but I think they can run hotter up to 125 c

The PCI fan on the flashed mac runs 100RPM faster but to be expected with the RX 580 and the Nvme on the pci.

So, I think the SMC still thinks the old chips are in and not letting the new chips run at that hotter temp so the fans kick in sooner, even though the chip temps are well within range.

So, if the SMC governs the fans on the mac can we fix that?

Firmware Restoration CD (Intel-based Macs)

https://support.apple.com/en-us/HT201692

But this would be no help as apple won’t let you dump the smc. once that particular smc is on your board it isn’t coming off. (!@£$%^&^%$£@) This will work only if your firmware install didn’t work.

Now if it’s all down to the SMC I’m stuck it would seem to having noisy fans.

I’m sure I saw a post saying that if you swap out cpu’s you need to run ASD afterwards
To do a thermal reconfiguration test. Now I’m getting excited again can I fix this?????
I have downloaded ASD file from https://www.loremipsum.info/ia-tools-blog/

for the 2009 (3S132) EFI -OS made the bootable drive as described in the instructions. Tried to run them but they froze (on the nvme drive & a normal bay spinner hd with fresh os instal)

I’m downloading another ASD file 2.5.8 that you burn onto cd so I’ll see if that works?

Well I can’t get the ASD to load, it just freezez on the efi and I only get the apple logo on the ASD os no bar just sits there, doing nothing…….

I’m also going to try AHT but can’t get the usb to Boot

GitHub - upekkha/AppleHardwareTest: Apple Hardware Test Download Links

Apple Hardware Test Download Links. Contribute to upekkha/AppleHardwareTest development by creating an account on GitHub.

github.com

Well I’m getting to the end of my investigative powers so I thought I’d share the experience and hope it helps someone else out, oh and maybe just maybe someone out there can help me.

 

[Macschrauber]

Temperature details are missing while under heavy load.

What about the Northbridge temperatures?

 

[BlueQube]

I've booted her up and run tests again, images below.
Maybe I missed this I just don't know but clearly cpu A diode from pceci is at 99c so that's the culprit after all. It would seem it's either the chip or the heatsink or paste.

So next action to swap the chips over and see if it the chip ?
if the led flashing and the temps move over it's the chip
if it doesn't it could be the heatsink on cpu A?

god I was losing my mind Thank's for making me do those checks again Macschrauber

just reading through this a bit confusing though ??
Definitions of sensors for classic Mac pro display

Definitions of sensors for classic Mac pro display - CRYSTALIDEA Forums

 

[Macschrauber]

Those differences under load are way huge.
Seems the cooling is not effective.

those are delidded processors?
ist the cooling surface polished and degreased?
are the heatsinks degreased around where they make contact with paste?
used quality paste, at least Arctic MX4?
is the glue cut evenly?
are the black plastic spacers mounted?

 

[tsialex]

"So, I think the SMC still thinks the old chips are in and not letting the new chips run at that hotter temp so the fans kick in sooner, even though the chip temps are well within range."

my first post here so tried to go into some detail:

I have taken an early 2009 mac pro 4,1 tower with 2 x CPU 2.26gh intel E5520's and (flashed) to 5,1

Then added 2 x 3.4gh Westmere Intel Xeon X5690's de lidded new paste and not overtightened.
(www.mac-factory.de)
Then added 48G ram – all good and recognised not mixed run in triple.
1 x 500G nvme m2 to x16 pci slot – boots great
1 x brand new Radeon rx580 Sapphire pulse 8g DDRR5 – (metal compatible) no boot screen as expected, some little dips of the screen to black but all acceptable – recognised and working.
Fresh install of High sierra 10.3.6 then firmware update to 140.0.0.0 via Mojave installer
Then Mojave – all fine and working great.

Reset SMC
Reset the PRam

Issues: the machine boots and runs fine for most things photoshop – DTP – web all good.
Adobe AE, Pr anything that draws a bit of load and the fans creep up more than usual for a mac.

When I put the computer under a load, say rendering with all the cores the fans kick in and get pretty loud, on the motherboard a small red LED flickers in unison with the render then stays solid until the render is done, then goes out when the load is over, the fans still blowhard when the temps of both CPUs are in the acceptable 40c the fans eventually go back to normal quite. (thermal throttling????)
#images attached.

(ok looks like C4d R17 is not compatible with Mojave/high sierra on this build)

Thing is c4d R17 works fine on my other mac pro and laptop on high sierra –
However, the smc and boot rom are now on Mojave ones so that would seem to be the incompatibility to me.

Stopped testing with C4D and used AE and Pr – fans still creep up sooner than they should
Blinking red led not happening unless under very heavy load, but fans still creeping up and getting loud.
Solution so far: MACs fan control, you can turn the fans down and deal with the noise but it’s the red light that needs to be resolved rather than just the fans.
This forum post tell the same story I’m having but not resolved.
https://forums.macrumors.com/threads/red-light-showing-from-processor-memory-board.1718643/

this forum takes of the same chip upgrade but different fan issues.
https://forums.macrumors.com/thread...gh.2006884/page-2?post=23821823#post-23821823


Now luckily, I have another mac pro5, 1 not flashed; A 2010 3.4ghz 12core duel cpu – ie I am building the new one to be pretty much the same as this current one.

The difference being I run a macvids flashed card NVidia gtx 980 4g with boot screen (even though the efi spasms out on boot screen in high sierra she still works a treat.

Now the led flash I get on the flashed 1,4 2009, happens on the 5,1 2010, however it just flashes once or twice and then the fans pick up just a little (hardly any noise and the light goes out- its only very brief.

So thermals right???

Temps at idle: both machines are pretty much the same including fan noise. I’ve looked at the temp’s under load on both machines as a comparison: AE – with openGL being used on the card and effects to push the cpu’s – temps run a little hotter on the new de-lidded cpu’s X5690's but I think they can run hotter up to 125 c

The PCI fan on the flashed mac runs 100RPM faster but to be expected with the RX 580 and the Nvme on the pci.

So, I think the SMC still thinks the old chips are in and not letting the new chips run at that hotter temp so the fans kick in sooner, even though the chip temps are well within range.

So, if the SMC governs the fans on the mac can we fix that?

Firmware Restoration CD (Intel-based Macs)

https://support.apple.com/en-us/HT201692

But this would be no help as apple won’t let you dump the smc. once that particular smc is on your board it isn’t coming off. (!@£$%^&^%$£@) This will work only if your firmware install didn’t work.

Now if it’s all down to the SMC I’m stuck it would seem to having noisy fans.

I’m sure I saw a post saying that if you swap out cpu’s you need to run ASD afterwards
To do a thermal reconfiguration test. Now I’m getting excited again can I fix this?????
I have downloaded ASD file from https://www.loremipsum.info/ia-tools-blog/

for the 2009 (3S132) EFI -OS made the bootable drive as described in the instructions. Tried to run them but they froze (on the nvme drive & a normal bay spinner hd with fresh os instal)

I’m downloading another ASD file 2.5.8 that you burn onto cd so I’ll see if that works?

Well I can’t get the ASD to load, it just freezez on the efi and I only get the apple logo on the ASD os no bar just sits there, doing nothing…….

I’m also going to try AHT but can’t get the usb to Boot

GitHub - upekkha/AppleHardwareTest: Apple Hardware Test Download Links

Apple Hardware Test Download Links. Contribute to upekkha/AppleHardwareTest development by creating an account on GitHub.

github.com

Well I’m getting to the end of my investigative powers so I thought I’d share the experience and hope it helps someone else out, oh and maybe just maybe someone out there can help me.

1) Nehalem processors are not supported anymore, support ended with El Capitan, and won't work correctly after 10.14.3. A lot of software nowadays require at least Westmere Xeons to work correctly and some even require newer processors that are incompatible with MP4,1/5,1.
2) MP4,1 or MP5,1 SMC firmware are not user or technician upgradeable, only Apple can do that.
3) It's the BootROM firmware that the Firmware Recovery CD have.
4) Once you flash a MP4,1 with MP5,1 firmware, you have to use the MP5,1 AHT and ASD, versions for MP4,1 will fail. MP5,1 ASD is 3S149.

 

[flygbuss]

I've booted her up and run tests again, images below.
Maybe I missed this I just don't know but clearly cpu A diode from pceci is at 99c so that's the culprit after all. It would seem it's either the chip or the heatsink or paste.

So next action to swap the chips over and see if it the chip ?
if the led flashing and the temps move over it's the chip
if it doesn't it could be the heatsink on cpu A?

god I was losing my mind Thank's for making me do those checks again Macschrauber

just reading through this a bit confusing though ??
Definitions of sensors for classic Mac pro display

Definitions of sensors for classic Mac pro display - CRYSTALIDEA Forums

Swap CPU A and B to see if one of them is the culprit. If so, contact Mac Factory. I bought several CPUs and other stuff from them and they've been always fair.

 

[BlueQube]

Swap CPU A and B to see if one of them is the culprit. If so, contact Mac Factory. I bought several CPUs and other stuff from them and they've been always fair.

Thank you flygbuss that's my next job , just waiting for thermal paste to arrive and good to know about macfactory might have to contact them if i can track this issue down to the chip.
[automerge]1580141288[/automerge]

1) Nehalem processors are not supported anymore, support ended with El Capitan, and won't work correctly after 10.14.3. A lot of software nowadays require at least Westmere Xeons to work correctly and some even require newer processors that are incompatible with MP4,1/5,1.
2) MP4,1 or MP5,1 SMC firmware are not user or technician upgradeable, only Apple can do that.
3) It's the BootROM firmware that the Firmware Recovery CD have.
4) Once you flash a MP4,1 with MP5,1 firmware, you have to use the MP5,1 AHT and ASD, versions for MP4,1 will fail. MP5,1 ASD is 3S149.

Thank you tsialex, I have 5,1 version so will try that (022-4831.dmg) and thanks for the heads up with the other chips.

 

[BlueQube]

Those differences under load are way huge.
Seems the cooling is not effective.

those are delidded processors?
ist the cooling surface polished and degreased?
are the heatsinks degreased around where they make contact with paste?
used quality paste, at least Arctic MX4?
is the glue cut evenly?
are the black plastic spacers mounted?

yes delidded
the surface was cleaned and greased, some light scratches on the heat sink plate.?
using artic 5
glue? not sure what you mean?
yes the black spacers are mounted

going to swap over the chips today so will take shots for more #ref

 

[BlueQube]

Ok the job is done and its all working a dream!!!!

So I took the liberty to make sure as Macschrauber said to polish and clean the heatsinks nice and clean. I used ice 5 thermal paste and proper cleaner to shine them up.

swapped CPU a for CPU b - booted up no luck - cleared the smc and she booted.

used cinabench to put it under load and hay presto it's quiet as a mouse. No temp issues all in the green same as my other machine set-up.

So after 3 rounds of thermal paste, it just shows you to use the right stuff. I also applied it as a line and not a round pea-size.


Thank you all for helping me sort this out, much appreciated.

 

[flygbuss]

Ok the job is done and its all working a dream!!!!

So I took the liberty to make sure as Macschrauber said to polish and clean the heatsinks nice and clean. I used ice 5 thermal paste and proper cleaner to shine them up.

swapped CPU a for CPU b - booted up no luck - cleared the smc and she booted.

used cinabench to put it under load and hay presto it's quiet as a mouse. No temp issues all in the green same as my other machine set-up.

So after 3 rounds of thermal paste, it just shows you to use the right stuff. I also applied it as a line and not a round pea-size.


Thank you all for helping me sort this out, much appreciated.

Congrats! Enjoy your 'new' system.

 

[Macschrauber]

glue? not sure what you mean?

if the spacers are used it could make too much height. So no contact to the heat sink.