MP 1,1-5,1 Mac Pro 4.1 -< 5.1 X5690, fans NOT spinning up?

[goodfidelity]

I just mounted a pair of delidded X5690 in a 2009 MB dual CPU board.

Machine boots perfect, runs perfect and fans operate nice and silent.

However upon using the stress test YES in terminal, i can reach CPU temp of 84 degrees (then i cancel test: KILL ALL YES) and still the fans are NOT increasing the RPMs.

So:
- I resetted EMC, powercord out and hold power button 10 sec
- I downloaded istatmenus and macfan control to monitor temps
- I reseated the CPU
- I Put pressure on the connectors for the fans (however fans run fine, and RPM shows fine)
- I can control the fans perfectly with mac fans control, they go up and down, and temperature is moving accordingly on the CPUs
- I tried in SIERRA and MOJAVE OSX on different disks, same result.
- When machine boots up, fans do the usual increase in RPM and then settle down.

For some reason the fans BOOST A and BOOST B + intake and exhaust does not increase the RPM when the CPUs are obvioslu overloaded. I cant think that its normally allowed with 70++ degrees (C) on the CPU and still the fans are not gearing up.

I noticed northbridge is 70+ degrees also, and when looking at the board i can se its cooling fins / cooler mounted with a pair of screws and not plastic rivets with spring pressure. This might have something to do with the northbridge temperature i assume, as the heatsink might be unevenly attached due to lack of springs..

However:

Nothing i did can make the fans increase RPM without me manually using the Mac Fans control.

Anyone seen this before? I have done multiple upgrades of these machines, lidded and delidded, but never seen this problem before. Usually i just reset all the stuff and they run again. This one does still not control the fans, even if everything is obviosly working as it should. Just not the controlling.

Can this be related to firmware or something? Its running 144.0.0, can this be updated to verify that this is not the issue if there was corruption during upgrade by previous owner.

//GF

 

[goodfidelity]

CPU temp is: A=80 / B=74 degree CELCIUS.

Fans are sitting at:
PCI 1200 RPM
PS 768 RPM
EXHAUST 600 RPM
INTAKE 600 RPM
BOOST A 1114 RPM
BOOST B 1113 RPM

Northbrigde dieote gives 70 degree, and northbridge heatsingk shows 57 degree. Maybe those are normal? I know it gets hot.

 

[tsialex]

I just mounted a pair of delidded X5690 in a 2009 MB dual CPU board.

Machine boots perfect, runs perfect and fans operate nice and silent.

However upon using the stress test YES in terminal, i can reach CPU temp of 84 degrees (then i cancel test: KILL ALL YES) and still the fans are NOT increasing the RPMs.

So:
- I resetted EMC, powercord out and hold power button 10 sec
- I downloaded istatmenus and macfan control to monitor temps
- I reseated the CPU
- I Put pressure on the connectors for the fans (however fans run fine, and RPM shows fine)
- I can control the fans perfectly with mac fans control, they go up and down, and temperature is moving accordingly on the CPUs
- I tried in SIERRA and MOJAVE OSX on different disks, same result.
- When machine boots up, fans do the usual increase in RPM and then settle down.

For some reason the fans BOOST A and BOOST B + intake and exhaust does not increase the RPM when the CPUs are obvioslu overloaded. I cant think that its normally allowed with 70++ degrees (C) on the CPU and still the fans are not gearing up.

I noticed northbridge is 70+ degrees also, and when looking at the board i can se its cooling fins / cooler mounted with a pair of screws and not plastic rivets with spring pressure. This might have something to do with the northbridge temperature i assume, as the heatsink might be unevenly attached due to lack of springs..

However:

Nothing i did can make the fans increase RPM without me manually using the Mac Fans control.

Anyone seen this before? I have done multiple upgrades of these machines, lidded and delidded, but never seen this problem before. Usually i just reset all the stuff and they run again. This one does still not control the fans, even if everything is obviosly working as it should. Just not the controlling.

Can this be related to firmware or something? Its running 144.0.0, can this be updated to verify that this is not the issue if there was corruption during upgrade by previous owner.

//GF

Dump the BootROM, save it securely then flash 10.14.6 MP51.fd and check if anything changes.

While it’s the SMC firmware that manages the fans, the hardware descriptor inside the NVRAM volume define the sensors resolution for the each one of the sensors. Earlier made early-2009 with base_16 and base_17 hardware descriptors are notorious for phantom sensors and sensor problems.

If the behavior changes when flashed with MP51.fd, you’ll need a BootROM reconstruction service.

 

[goodfidelity]

thanks Alex, ill give it a try tonight!


//GF

 

[goodfidelity]

I tried to put the blank rom with rom tool, and after reboot i noticed still was original serial number.

Does that mean the flash was not successful?

I did as pr. instructions and used the csrutil disable, and then rebooted machine into the rom-reprogramming mode by holding the power key at startup befor initializing the flash.

Still it does not look like the ROM file was updated right?

//GF

 

[tsialex]

MP51.fd is unserialized, so, you didn't flashed it. You probably did not enter Firmware Programming Mode correctly.

 

[goodfidelity]

do i just rename them .rom before flashing? found some different ones.

//GF

 

[tsialex]

do i just rename them .rom before flashing?

Correct.

found some different ones.

//GF

Be sure to use Apple 10.14.6 full installer, do not get this from a 3rd party and always check the crc32 before anything.

 

[goodfidelity]

Thanks!

Yeah, i used the ROM you provided some year ago

Problem is still there. Since it CPU's report the temperature correctly. Is there any reason to believe problem is with the CPU daughter board or can it still be processors?

From what i can see, there has been work done on this CPU board, as the Northbridge chip has screws and not nylon retainers. Also i noticed what i think looks like residue from burned material on the copper part of heat-sink. This might indicate that this machine has overheated before, and just got new CPU's as a simple fix before it was re-sold.

Maybe there is a problem with the SMC firmware on the CPU board?
Is there any way to obtain a known working copy of this SMC firmware?

I know one guy did a reflash of this, as i read that in another thread.. to use in 2010 machine the board from 2009 machine. Is it located on the daughterboard or on the mainboard?

//GF

 

[tsialex]

Thanks!

Yeah, i used the ROM you provided some year ago

Problem is still there. Since it CPU's report the temperature correctly. Is there any reason to believe problem is with the CPU daughter board or can it still be processors?

You can always install the CPU tray on another Mac Pro with the same SMC release and check if the behavior changes.

From what i can see, there has been work done on this CPU board, as the Northbridge chip has screws and not nylon retainers. Also i noticed what i think looks like residue from burned material on the copper part of heat-sink. This might indicate that this machine has overheated before, and just got new CPU's as a simple fix before it was re-sold.

Maybe there is a problem with the SMC firmware on the CPU board?

Possible, while the SMC microcontroller is extremely resilient, failures happen.

Is there any way to obtain a known working copy of this SMC firmware?

Nope, not possible to dump since it's a secure microcontroller that don't even have a way to the firmware be accessed once stored (read the Renesas H8S datasheet for details) and Apple never issued an update like back with the MacPro2,1 (that permitted users to fully upgrade MacPro1,1 to 2,1).

I know one guy did a reflash of this, as i read that in another thread.. to use in 2010 machine the board from 2009 machine. Is it located on the daughterboard or on the mainboard?

//GF

You are mistaken, it's not possible to do it without the SMC firmware code that no one outside Apple have.

Several people REPLACED the SMC microcontroller from another CPU tray, there is one thread about it somewhere.