MP 1,1-5,1 PSU temperatures

[Tommi14]

Hello,

I just noticed that my PSU temperatures are 41 C and 56 C. These are after a couple of hours on idle. They seem rather high, don't they? Or are these temperatures normal? Specs are: Mac Pro 2009 4.1->5.1, dual X5675, RX580, 970 EVO NVMe. No other disks at the moment.

Thanks!

Tommi

 

[Stex]

I would say yes, they seem on the high side based on my temps. When I have ambient temp reading at about 30C, the highest PSU temps I get are around 40C, usually with about 1-3C difference between PSU sensors. Others should share their temps and thoughts, I am not sure if you have a problem or not, even if they seem above my average temps....

 

[h9826790]

How-To: FIX MAC PRO FAN SPEEDS when using PC (non-EFI) Graphics Cards

The temperature sensors in the Mac Pro are not tuned for things like modern nVIDIA graphics cards and certain SSDs or HDs, and will ramp up the fans tremendously, even though the temperatures are low. The sensitivity range was tuned for the super-cold, old graphics cards the system shipped with...

forums.macrumors.com

 

[KeesMacPro]

Short version: You could set the PSU fan to sensor: supply2 , min at 32C and max 50Celsius.

 

[Tommi14]

Thank you all for your reply! I really appreciate it.

Stex, your setup is quite similar to mine but your PSU temps are lower. Do you use some custom fan controls by MFC etc.?

I read the thread linked by h9826790.

In fact, my PSU temps act very similar to those reported in this post:

https://forums.macrumors.com/thread...-non-efi-graphics-cards.1946672/post-24070574

When idling, the temps are:



After running "yes > /dev/null &" (times 12) in Terminal for 10 min, the temps were:



Notice that the PSMI Supply 2 temp is 8 degrees of C lower that at idle even though the PS fan speed is the same.

If I use the settings proposed by KeesMacPro, I will end up with these temps and fan speeds:



The PSMI Supply 2 temp is 13 degrees of C lower than before but the PS fan speed and the noise are a lot higher.

All the time, the PSU area of the casing and the air coming from the PSU feels quite cool and not hot at all. In contrast, during the "yes"-running the air coming from the lower part of the casing felt noticeable warm.

I might try to dedust the PSU as proposed by h9826790 here:

https://forums.macrumors.com/thread...-non-efi-graphics-cards.1946672/post-24088610

But otherwise, I am not sure what would be the best reaction to the high PSU Supply 2 temp.

I apologise the quality of the screen shots.

 

[Stex]

@Tommi14 -Yes we have similar setups. You probably found some of my old posts in those threads were I got great suggestions from @KeesMacPro and other helpful members.

At this time I am using iStat Menus for fan control and general monitoring. I tend to keep the PS fan rpm quite high to keep the PS temps below 35C. My PS fan is always at a minimum 1050rpm, and lately with warmer weather I push it to 1150rpm. But yes, that fan above 1000rpm is very noisy. In fact, today I finally bought a new Noctua fan to replace the PS fan - a suggested upgrade if you want quieter cMP fans in general.

Regardless of my ambient temps and fan settings, I have never had such wide difference between the two PS sensor readings. Your PS temp difference does not look normal to me. At most, I have 1-3C difference (as previously stated). Your last screenshot has better PS temps for sure, but 1500rpm is very high. I never had to push the PS fan that high even with 30C ambient reading.

Maybe test the higher rpm fan setting for a few days and keep an eye on the numbers to see if there any changes based on use/internal temps etc. I am not sure, maybe others can let us know if they have (or have had) wide difference between PS readings like you...

 

[h9826790]

Thank you all for your reply! I really appreciate it.

Stex, your setup is quite similar to mine but your PSU temps are lower. Do you use some custom fan controls by MFC etc.?

I read the thread linked by h9826790.

In fact, my PSU temps act very similar to those reported in this post:

https://forums.macrumors.com/thread...-non-efi-graphics-cards.1946672/post-24070574

When idling, the temps are:

View attachment 1782260

After running "yes > /dev/null &" (times 12) in Terminal for 10 min, the temps were:

View attachment 1782261

Notice that the PSMI Supply 2 temp is 8 degrees of C lower that at idle even though the PS fan speed is the same.

If I use the settings proposed by KeesMacPro, I will end up with these temps and fan speeds:

View attachment 1782264

The PSMI Supply 2 temp is 13 degrees of C lower than before but the PS fan speed and the noise are a lot higher.

All the time, the PSU area of the casing and the air coming from the PSU feels quite cool and not hot at all. In contrast, during the "yes"-running the air coming from the lower part of the casing felt noticeable warm.

I might try to dedust the PSU as proposed by h9826790 here:

https://forums.macrumors.com/thread...-non-efi-graphics-cards.1946672/post-24088610

But otherwise, I am not sure what would be the best reaction to the high PSU Supply 2 temp.

I apologise the quality of the screen shots.

@Tommi14 -Yes we have similar setups. You probably found some of my old posts in those threads were I got great suggestions from @KeesMacPro and other helpful members.

At this time I am using iStat Menus for fan control and general monitoring. I tend to keep the PS fan rpm quite high to keep the PS temps below 35C. My PS fan is always at a minimum 1050rpm, and lately with warmer weather I push it to 1150rpm. But yes, that fan above 1000rpm is very noisy. In fact, today I finally bought a new Noctua fan to replace the PS fan - a suggested upgrade if you want quieter cMP fans in general.

Regardless of my ambient temps and fan settings, I have never had such wide difference between the two PS sensor readings. Your PS temp difference does not look normal to me. At most, I have 1-3C difference (as previously stated). Your last screenshot has better PS temps for sure, but 1500rpm is very high. I never had to push the PS fan that high even with 30C ambient reading.

Maybe test the higher rpm fan setting for a few days and keep an eye on the numbers to see if there any changes based on use/internal temps etc. I am not sure, maybe others can let us know if they have (or have had) wide difference between PS readings like you...

The Delta PSU runs cooler, also much lower temperature between sensor 1 and 2 (than the AcBel PSU)

 

[Stex]

The Delta PSU runs cooler, also much lower temperature between sensor 1 and 2 (than the AcBel PSU)

I would think @Tommi14 and I have the same PS based on the specs of our systems...? 4,1-->5,1? Or do you think we have different PS?

 

[h9826790]

I would think @Tommi14 and I have the same PS based on the specs of our systems...? 4,1-->5,1? Or do you think we have different PS?

Can be different

 

[Stex]

Can be different

Got it. Thanks Martin.

@Tommi14 - for your reference I took this screenshot just now with 30C ambient, not really at idle though, system is doing medium to low load I guess:

 

[MIKX]

I replaced my 4,1>5,1 cMP's PSU fan with a Noctua fan and their fan speed controller.
Total temp control + fan silence.

cMP 4,1 & 5,1 120mm Noctua PSU fan with Noctua NA-FC1 PWN speed controller- both working.

I have replaced my cMP 4,1>5,1 Apple standard 120mm PSU fan with the Noctua NF A12x25 120mm PWN fan and have also installed The Noctua NF-FC1 PWN Fan speed controller. I am happy to report that both the Noctua fan and the fan speed controller are working perfectly. EDIT : I cleaned up the...

forums.macrumors.com

 

[Tommi14]

@Stex, thank you for your reply and advices! I really appreciate it. Your PSU temps look definitely lower compared to mine.

@h9826790, I guess there is no way to find out the PSU brand other than taking it out of the case, is there?

@owbp must have a PSU from the same manufacturer as I because his PSU temps are very close to mine and they act exactly as mine during idle, under a load and when increasing the fan speed.

https://forums.macrumors.com/thread...-non-efi-graphics-cards.1946672/post-24070574

 

[h9826790]

@h9826790, I guess there is no way to find out the PSU brand other than taking it out of the case, is there?

The definitive answer is on the sticker of the PSU, which you need to take it out to read.

 

[Stex]

@h9826790, I guess there is no way to find out the PSU brand other than taking it out of the case, is there?

Oh well, sounds like you will have to open up your PS to fully confirm your PS brand @Tommi14. FYI, I will be installing a new PS fan in the next 2 days, so will post here my PS model with some pics of the label (it might help future readers to see a documented example on how the two brands differ on the temp readings etc).

 

[Tommi14]

Oh well, sounds like you will have to open up your PS to fully confirm your PS brand @Tommi14.

Yes, I am not going to do that, at least not now. I don't need to know that information so much.

 

[TECK]

Got it. Thanks Martin.

Your Northbridge Tdiode runs cool like a spring morning breeze, mines is at 76 degrees. Might need to clean and replace the paste.

 

[h9826790]

Your Northbridge Tdiode runs cool like a spring morning breeze, mines is at 76 degrees. Might need to clean and replace the paste.

View attachment 1783286

No, your thermal paste is fine, all you need is just spin up the fans

 

[Tommi14]

Is it possible that an SMC reset or a NVRAM reset could correct odd (too slow) fan behave or high temperature readings?

I have never done the resets (at least not the NVRAM reset), even after the 4.1->5.1 upgrade and the CPU upgrade.

 

[h9826790]

Is it possible that an SMC reset or a NVRAM reset could correct odd (too slow) fan behave or high temperature readings?

I have never done the resets (at least not the NVRAM reset), even after the 4.1->5.1 upgrade and the CPU upgrade.

It's usually the other way around. The fans may run at unnecessarily high speed, and SMC reset may able to bring them back to normal idle.

The cMP's fans are programmed to run at very low speed whenever possible. This is how Apple choose between "noise" and "cooling".

 

[Tommi14]

Ok, thanks.

I am just trying to understand what are the possible reasons for the PSU temperature readings on my Mac Pro being so different (a lot higher) compared to the temperature readings reported here from other Mac Pros with more or less similar components and settings.

And, at the same time, why the temperature readings with this Mac Pro

https://forums.macrumors.com/thread...-non-efi-graphics-cards.1946672/post-24070574

act almost exactly as with mine?

 

[Stex]

Ok, thanks.

I am just trying to understand what are the possible reasons for the PSU temperature readings on my Mac Pro being so different

There are several options worth testing to understand more about your temps and potential issues, as suggested by previous comments/replies to you:

1- Find out what PSU brand you have (as @h9826790 pointed out this could be the main issue at play and he has ample experience to suggest so!)
2- Do a deep clean of the PSU, which happens to be easier and convenient after you took the PSU out to check the brand. Using compressed air and cleaning the fan carefully will help in most cases.
3- Program your own fan settings to reduce temps then monitor those new settings over 1 week or so to see if there are any improvements/changes or new issues popping up ((as @KeesMacPro suggested and he has lots of experience with cMP thermal issues to suggest so!)

All of the above involve time and effort, but it is very likely that after doing the above you will have a better sense of 1) whether it was just a combination of your PSU brand+collected dirt OR 2) a faulty sensor.

 

[KeesMacPro]

I am just trying to understand what are the possible reasons for the PSU temperature readings on my Mac Pro being so different (a lot higher) compared to the temperature readings reported here from other Mac Pros with more or less similar components and settings.

And, at the same time, why the temperature readings with this Mac Pro

https://forums.macrumors.com/thread...-non-efi-graphics-cards.1946672/post-24070574

act almost exactly as with mine?

Comparing a few MP4,1/5,1 with different hardware setups , but all running Mojave natively and BootROM 144.0.0.0.0 ,
I deducted a few things:

- there are 2 (main) power rails and afaik PS1 =5Volts and PS2=12Volts
PS1 Temp is relatively constant regardless CPU and/or GPU load
PS2 Temp depends mainly on the GPU powerdraw (logically because GPU power=12V)

- the Temp of the GPU and the design ( open vs reference) has its influence on the Temp PSU too
Apparently the T PCI ambient has more influence on the PSU T than I thought..
A Sapphire Pulse RX580 (=open) results in a higher T PSU than e.g. a GTX780TI (=reference) , curiously because the GTX draws a lot more power than a RX580....

- In one of my single CPU 4,1>5,1 i noticed that PS1 was constant at 37Celsius regardless T ambient, load etc etc
PS 2 stayed around the same T at idle and increased depending on the (GPU) load
This was a ACBell PSU and I couldnt find any faulty sensor or leaking capacitor, but replaced it for another known working and dust free ACBell PSU anyway.
Result: exactly the same T readings

After several SMC and NVRAM resets, I installed another known working LogicBoard.
Result: exactly the same T readings, so the SMC is working correctly, unless the CPU board is causing this phenomenon, which i doubt....

During this testing, as mentioned by @Tommi14 i noticed that the air coming out doesnt "feel" hot , regardless the T readings of the PSU.


So an ACBell PSU runs hotter/ shows higher readings than a Delta PSU , as confirmed by @h9826790
Even compared to a DUAL CPU with higher load, but with a Delta PSU....

Another factor is that filling up the optical bay with e.g. 2 x Superdrive , reduces the incoming airflow to the PSU significantly, resulting in higher overall PSU T as well .....

 

[Tommi14]

Thank you @KeesMacPro, @Stex and @h9826790 for your advices and effort.

I finally got the time to take the PSU out. It was easier than I thought. The brand of the PSU was AcBel. The PSU did not look particularly dusty, but I blew it thoroughly with CompuCleaner Xpert anyway.

And the result: no effect on the PSU temp readings. Below are the idle temp readings after the cleaning:



So, the PSU temp readings are still high.

I also tried what will happen if I replace RX 580 with GT 120: no effect either. Below are the idle temp readings with GT 120:



The PSU temp readings are the same with RX 580 and GT 120, but the PCI and PS fan speeds are slightly higher with RX 580.

So, any idea what could be deduced from this? At least it seems that the high PSU temp readings are not related to the GPU.

As a side note: with GT 120 there was a temp reading for "PCIE slot 1", which I think is related to GT 120 and it was around 50 C. This also is quite high, isn't it?

Tommi

 

[KeesMacPro]

The PSU temp readings are the same with RX 580 and GT 120, but the PCI and PS fan speeds are slightly higher with RX 580.

The PCI and PS fan , when controlled by the SMC (=auto) , seem to respond a lot to the PCI power draw and not only the PSU and/ or PCI Temp readings. The RX580 draws more power than a GT120 .

As a side note: with GT 120 there was a temp reading for "PCIE slot 1", which I think is related to GT 120 and it was around 50 C. This also is quite high, isn't it?

Yes, the GT120 shows up as "PCIE slot x".
The GT120 is a very basic GPU , about 1 decade old , and easy to heat up with graphic demands normal for today's standard.
When I install a GT120 for e.g. testing I always read Temps of 50C (at idle) and upwards.
So, besides the quite poor performance, the GT120 runs hot in general...

 

[Tommi14]

I have made some progress in lowering the PSU temps in my Mac Pro.

Firstly, I set in Macs Fan Control for PS a sensor-based value: PSMI Supply AC/DC Supply 2, min 35 C, max 75 C.

Secondly, I put hard drives in bays 3 and 4.

These two changes lowered the PSU Supply 2 temp from 56 C to 40 C at idle.

It seems that for some reason having hard drives in bays 3 and 4 has significant affect to the PSU Supply 2 temp. The temp is a lot higher if the bays are empty.