09-26-2020 01:46 PM - edited 09-26-2020 01:47 PM
The Z640 rear fans are running 2500RPM and they are very loud. I measured 80dB
Here are my troubleshooting so far:
- The fans are ok, PWM works
- The motherboard is ok, PWM regulation works.
- I can set the speed to maximum in bios, so PWM works.
- The ambient sensor is good, 25 Celsius.
- CPUs are cool, 52 Celsius.
- I tried CMOS clear.
- I tired fresh Win10Pro install.
- Installed the earliest and latest BIOS, same RPM.
- Tried CMOS clear with fresh Win10Pro, it seemed like the fans were quiet for some time but when install finished, they still ramped to 2500RPM.
- When 100% CPU utilization, the fans don't spin up? I get 800RPM at 100% and 80 Celsius.
I googled everywhere and I can't find nothing else to try.
The rear fans are the same as the CPU0 fan, so the rear fans are able to run in 400RPM.
I noticed there is power connection for Memory Fan Baffle on motherboard and case has attachment holes. So perhaps the BIOS thinks the Memory Fan Baffle is missing and tries to compensate by ramping up the rear fans.
To conclude, it seems like some simple BIOS bug where the fan curve is messed up for some reason
Here is the "normal" HP Performance Advisor. The configuration warning is about missing some memory modules.
Solved! Go to Solution.
09-27-2020 08:28 AM - edited 09-27-2020 03:05 PM
I've seen a similar thing in the past on a HP workstation PWM fan and this is an easy fix if it is the same issue. As you know those rear case fans on the Z640 (and the Z620) are dual 92mm fans with the two sets of 4 wires bonded together along the cable so that at the motherboard plug end there are 6 instead of 8 total wires. I don't know if those two workstations used exactly the same dual-fan kit. A number of eBay sellers seem to think so. Take a look at your dual-fan plug end and make sure you end up with the same plug end if you go this route. Here is the situation:
The HP PWM fans always are fed +12VDC, and if you take away their PWM control wire they'll run at full speed. Their wiring (from pin 1 to pin 4) is the standard PWM order... Ground, +12VDC, rotor RPM signal back to the motherboard, PWM control from motherboard out to the rotor's control board. If anything in the system disrupts the PWM control signal the default thus is to run at full 12V speed. This can happen from a break in the PWM control wire to a failure of a PWM control circuit board component. That is what I saw in the past.
If you're lucky the issue is at the fan end of things rather than at the motherboard end, and getting a replacement pair of those fans with their special 6-pin fan plug end will fix your issue. I'm seeing them on eBay for roughly $15.00 USD for the pair. I would not hesitate to buy a pair used... they tend to last a very long time. Let us know how things turn out, and here is some pics from the two part numbers I found (Assembly P/N 644315-001 and Spares P/N 683764-001). You may find a later part number by searching HP's parts system specifically for your Z640. The dual-fans are 92mm, not 80, and a Z640 I checked here was using two Nidec fans, but HP also used Delta ones in these units:
09-28-2020 01:01 AM - edited 09-28-2020 01:28 AM
@SDHthanks for the reply, I like your investigative work.
I do own two Z620 and I swapped out the Rear Fan Assembly and the result is the same. I know the fans are wired correctly and the PWM is working because changing the fan idle will speed up the fans and when the PC boots, the fans are on slow at first, then ramp up after 5 seconds into boot, (so the motherboard do control the fan speeds for sure).
Now if I had already Z620 fans on my Z640, when I purchased it, that would mean the original Z640 fans have different PWM-to-RPM relations. But Googling and searching in eBay, there seems not to be such different unit for Z640. It's just reusing the Z620 fans.
The problem is actually more serious, the fans don't speed up when I put the workload on 100%. The RPM stays the same no mater what is the CPU temperature. So my problem is not the rear fans, but the fan system in whole.
Further reading it seems like there is a known flaw with the Z640 fan control that is tied to BIOS updates:
1. Fix a suitable fan speed in BIOS and you may have to live with a noisier system.
2. Refresh the BIOS from the original to the latest version in proper sequence without skipping versions. Remember to clear CMOS after every upgrade. Proceed to further upgrade the BIOS only if fan rpm changes to CPU load as intended. I managed to get my Z820 with the liquid cooler to adopt a proper fan control strategy instead of using a very high rpm all the time. I suspect that the fan control settings in BIOS gets corrupted somehow when the BIOS is upgraded and following this procedure keeps the workstation working and myself happy and sane.
3. Jerry rig a Noctua fan controller to both or each CPU for manual fan control. Careful of the fan pinouts.
4. Repair the motherboard if you have access to board level schematics.
5. Purchase a replacement motherboard.
So the only solution would be to get a motherboard replacement to fix fan control.
I do really hope some HP engineer will read this and find the bug in Z640 fan control and can patch this!
EDIT: I stumbled on some good tip here. The fans are controlled by Intel ME separate chip, not by BIOS or OS. And in fact I found drivers for Intel Management Engine that I have not installed. Will try later and report back.
09-29-2020 02:04 AM
So I installed all drivers (Intel ME and BIOS) and rebooted but nothing changed. All fans are stuck in some limbo RPM and will not regulate to the computer temperature. I still think the problem is somewhere with the firmware and not hardware.
09-29-2020 02:31 PM - edited 09-29-2020 02:33 PM
So I managed to fix the fan problem. I'm using ASUS HYPER M.2 X16 CARD V2, a quad NVMe card, and did not have bifurcation turned to 4x4x4x4 (I'm using a single NVMe drive now so figured I'll leave it on 16x). When the PCIe slot was in 16x mode the rear fans were running faster. I discovered this by removing PCIe cards one-by-one.
If you have same problems and try to solve this, I did swap the "Boot Block Recovery" and "ME/AMT flash override" jumpers on, did the corrupt BIOS update from USB stick and made a clean OS install. Might have not needed it as I did change the bifurcation setting later, but it might be something that helped.
Thanks to @SDH for thinking along