MP 1,1-5,1 [Solved, RAM overheating]Booster fan full speed while other fans idle

[shuotong]

Hi all,
I recently have This wired behavior as title suggests, my booster fun running at 4000rpm where all other four (intake exhaust PCI PSU) runs at their idle speed under full load CPU stress test, my CPU t-diode shows a temperature of 55C which is way too low for booster to run at full speed.
The machine is a 4,1 -> 5,1 single processor with X5690 upgrade.
The SMC reset does not help, it still goes to full speed when under load. Without the load everything is fine though, what concerns me is why all other fans are not speeding up even with graphics card under load, the PCI fan used to go to 1500rpm, intake and exhaust will goes to 1200 or so.
I searched the form and some suggested to use smcfancontrol to reset the fan management, that does not work well either, right now I’m out of my wisdom of what’s the next thing to try. Everything was very normal until today it suddenly speed up for no reason.
[automerge]1576391677[/automerge]
According to iStat menu there’s no unusual temperature behavior apart from the CPU is running much cooler due to the insane fan speed.

 

[h9826790]

Please at least post a full picture of the iStat menu readings.

Otherwise, we can only assume your computer is "normal" because that's what you are saying (except the high booster fan speed), and nothing we can help.

 

[MIKX]

Also, it would be a good idea to write a description = "SIGNATURE" of your cMP + GPU brand + Bootrom version etc.

As an example, my SIGNATURE is at the bottom of this post.

99% of of the talented Mac Pro Thread members cannot read minds.

 

[shuotong]

This is probably rare but I'll post my understanding here.
The source of this problem leads to "overheating" RECC memory, I realized the booster fan will speed up in case the DIMM reading get way too high, from my experience the booster fan will accelerate once the DIMM temperature reading is about to reach 75. This of course is not going to cool the memory down so the fan will simply gradually goes to max. The operating temperature of these memory sticks are minus something to 95 degrees, so within their normal temperature range, but obviously the feedback loop of the Mac Pro are designed using factory memory.
I encounter this problem when running memory intensive work, everything goes back to normal if I use only two sticks (regardless of which two) in slot 1,3 or 2,4 so I give them enough space to cool down. I believe the reason is quite simple, they're not full size DIMM and very old (manufactured in 2012), probably manufactured using 45nm or 65nm technology. Running CPU stress test gives me no problem, all three fans accelerate simultaneously. At first I thought it was CPU running too hot and I gave the CPU new thermal pastes, which makes it worse and the two case fan never accelerate because the CPU is even cooler.
If my assumption is right, that's would be a flaw in the fan speed feedback loop. The obvious way to cool down memory is definitely not using booster fan.
Using mac fan control to set the intake/exhaust fan running at 1000rpm will solve the problem. But since these memory sticks are still in their return window so I'll probably return them and get some newer stuff.

[automerge]1576516060[/automerge]

This is my istat menu sensor readings under CPU stress test. I was quite surprised when I replace the AS5 thermal paste with Thermal Grizzly kryonaut, I only changed the thermal paste and nothing else, radiator was cleaned regularly. My machine runs silent after the thermal paste change, the biggest noise source now is my hard drives.
I did not saved a screen shot before, but it'll update it here once I can simulate the environment again.