MP 1,1-5,1 Mac Pro (4.1/5.1) aftermarket fans WITH automatic SMC control

[KeesMacPro]

I can't bring the fans down with TG Pro even if I set them to "1% always". I turned on "completely override the system" but I get it only slightly below 600rpm like 575rpm.

What's the trick?

The default min rpm @arw posted are on a MP running OpenCore and iStat.
Somewhere there will be a logical explanation for these default min rpm's....
Anyway, I would suggest to read through the beginning of this thread and look for all posts by amedias.
Short version of his thorough investigation (looking for an MP as silent as possible!!!) is that to lower min default fan rpm without cooking something ,one has to know very very well the hardware configuration and the typical load on this specific MP.
Note that the MP 4,1/5,1 is a device released ~12 years ago and designed with the knowledge and hardware that was available at that point in time....

If you feel comfortable and really want to go that path:

GitHub - hholtmann/smcFanControl: Control the fans of every Intel Mac to make it run cooler

Control the fans of every Intel Mac to make it run cooler - hholtmann/smcFanControl

github.com

 

[davorrr]

Thanks!
SmcFanControl goes even lower than TG?

 

[KeesMacPro]

Thanks!
SmcFanControl goes even lower than TG?

NP!
Yes, take a look at the rpm in the posts by amedias.

 

[arw]

The default min rpm @arw posted are on a MP running OpenCore and iStat.
Somewhere there will be a logical explanation for these default min rpm's....

Just tested on a 2009 4,1>5,1 Single CPU, Triple PRAM, Vanilla Mojave with only the latest iStat menus.
After a SMC reset these are the default speeds:

However, once I have clicked anywhere in the fan speed section of iStat and then back to "System Defaults" the speeds go down to this:

 

[Stex]

I will , and post a report as soon as I got a few hours to test it ; )

I'm interested in your results too! All for the perfect thermal conditions for our beloved cMP!

 

[KeesMacPro]

Vanilla Mojave with only the latest iStat menus.
After a SMC reset these are the default speeds:However, once I have clicked anywhere in the fan speed section of iStat and then back to "System Defaults" the speeds go down to this:

So we narrowed it down to iStats Menus !

 

[mnlth0]

Hi all,

I've replaced the PCI fan! Actually, I tried to do it and realized I haven't soldered in many years so I asked a well versed friend to do it... and it worked out great. I can also tell now that PCI fan was the most noisy one! I didn't take any photos while the work was done (so I'll attach a blurry out of focus photo of fan running) but here's a few tips for replacing the PCI fan:

1. This post and one following it will show to what pins one connects Noctua fan and what colors.
2. when dissembling the fan cage, don't forget to remove two hex screws that are on the inside section.
3. when you remove the screws you will need to "jimmy" the case a bit to remove the case cover from part that's the holding fan.
4. IF YOU ARE NOT WELL VERSED WITH SOLDERING, ASK A FRIEND FOR HELP! this is what I did. I tried and realized it would be hard to do it perfectly so I asked a buddy to help and it worked out beautifully. It takes 15 mins for the whole procedure.

Re the discussion about fans and iStat, @KeesMacPro - could this be the reason for issue I reported earlier, when booster fans speeds are increased if I increase speed of intake / exhaust? My issue is that I also removed istat and it still kept happening.

I'm also attaching a screenshot of fan speeds with replaced intake / exhaust and PCI fans. This is still using defaults except the PCI fan being on 1500 for a test, as you can see other fan speeds aren't increased. Noctua PCI fan is loud at 1500.

But the moment I increase intake / exhaust, boosters are automatically increased as well. This happens even if istat isn't installed, whether smcfancontrol or tg pro are running alone. See second screenshot.

Lastly, I feel like booster fans are now the loudest so I will finally order those Noiseblocker blacksilent PR-2 fans! I assume the PR-2 are the ones I order, right?

 

[mnlth0]

@arw I have the same mod for northbridge, except I used NF-A4x20 and not NF-A6x25 FLX. What temps do you see at northbridge? Mine still goes up to 70 sometimes, never below 60.

 

[KeesMacPro]

Re the discussion about fans and iStat, @KeesMacPro - could this be the reason for issue I reported earlier, when booster fans speeds are increased if I increase speed of intake / exhaust? My issue is that I also removed istat and it still kept happening.

No, I dont think this is related.

I will finally order those Noiseblocker blacksilent PR-2 fans! I assume the PR-2 are the ones I order, right?

Yes, Blacksilent Pro PR-2 are the ones that I planned to test, so I'd test them first before disassembling heatsinks, cutting cables, soldering etc etc etc

 

[mnlth0]

Good point, thanks! Just to be sure, how would you test them before disassembling?

 

[KeesMacPro]

how would you test them before disassembling?

If you made an extension cable for the PSU,Intake or Exhaust fan , you can connect the fan to that cable for testing.
Obviously this is for a short test , as one of the fans will be temporarily disconnected , so watch Temps carefully!
I'd test if it runs stable at 800 rpm, 1000 rpm and see if it runs at 2500 (according to the specs).

If you made extension cables for all 3 120mm fans, I'd disconnect the Exhaust because it has the least direct impact on Temps.

 

[mnlth0]

Excellent, that's a great tip! Thanks

 

[arw]

@arw I have the same mod for northbridge, except I used NF-A4x20 and not NF-A6x25 FLX. What temps do you see at northbridge? Mine still goes up to 70 sometimes, never below 60.

Because of its size, I knew the 60mm fan can only be placed at a semi-ideal place/angle regarding the Northbridge but I thought at an inaudible speed it could outperform the 40mm anyway.
I wanted it attached to the CPU cage to keep the CPU tray's easy accessibility.
It was however a real PITA to mount it to the cage without transferring any vibrations.
Config:

• In/Out NF-S12B redux-1200 with removed fan guards (makes a lot of difference under load)
• Stock Boosters (dual x5690; fresh thermal paste for CPU B only, then I ran out)
• Stock Northbridge/Heatsink paste
• NF-A6x25 FLX at absolut minimum

I let Cinebench run for an hour (pic1) and let it cooldown afterwards (pic2).
I'll leave it like that and wait how it changes in summer.

Edit: I just realized, these values say nothing about the fan without having a reference before installing it. I will catch up on that later.

 

[arw]

I'm interested in your results too! All for the perfect thermal conditions for our beloved cMP!

I just remembered the values before removing the fan guards (or however the may be called):
- For the NF-S12B redux-1200 as Intake and Exhaust the removal led to decreased speeds of 850 rpm from around 1050 rpm after 1 hour of Cinebench for dual x5690 in system controlled environment. That's a 20% reduction in speed and more than doubles the headroom to rpm max.
The (subjective) acoustical impression is by far greater than these values suggest. After all it was the noise at high loads that led me to try this in the first place.

 

[avro707]

Because of its size, I knew the 60mm fan can only be placed at a semi-ideal place/angle regarding the Northbridge but I thought at an inaudible speed it could outperform the 40mm anyway.
I wanted it attached to the CPU cage to keep the CPU tray's easy accessibility.
It was however a real PITA to mount it to the cage without transferring any vibrations.
Config:

• In/Out NF-S12B redux-1200 with removed fan guards (makes a lot of difference under load)
• Stock Boosters (dual x5690; fresh thermal paste for CPU B only, then I ran out)
• Stock Northbridge/Heatsink paste
• NF-A6x25 FLX at absolut minimum

I let Cinebench run for an hour (pic1) and let it cooldown afterwards (pic2).
I'll leave it like that and wait how it changes in summer.

I'm using the smaller 40mm Noctua fan put over the Northbridge but I mounted it from the CPU heatsink, so it's a bit more fussy to take out the CPU tray. My Northbridge temps usually stay below 60ºC. At the moment I have 58ºC, in a state fairly similar to your cool down. I have the bits and pieces to try and make the northbridge heatsink larger which might have a similar effect but no time to try and do this. I'd also prefer to use a spare NB heatsink rather than my good ones.

 

[KeesMacPro]

I‘m eager to read your results, should you try it.

Just ran some tests to see what happens when the fanguards are removed.
I recorded the fan noise with an AKG condenser mic placed in front of the MP at a distance of 25cm.
Tests were done with Intake and Exhaust both equally set to different fixed rpm on a MP 4,1>5,1 DUAL X5690 running Mojave.

I attached some screenshots but if TLDR:
Noise is indeed a bit lower ~ 1,5 dB , which doesnt seem much, but something very interesting appears as stated by @arw :
My Intake runs in this season (Winter) at ~ 950 rpm at idle, but after removing the fanguard it's now running at ~ 750 - 800 rpm (the fan is controlled by T NB diode) apparently because of improved airflow.
So in real life fan noise is reduced much more than 1,5dB.


My conclusion: excellent idea !

Edit: post edited.

 

[arw]

Wow, that is some excellent investigation.
I‘m really glad to been able to contribute something to this great thread.
Regarding the noise improvement: When you don‘t compare the sound levels at fixed fan speed values but fixed temperatures, it better explains my subjective impression.
Just to elaborate what I mean, assuming these two lead to the same temperatures: -8 dB at 1200 rpm with fanguard versus -11,5 dB at 900 rpm without. Depending on the evaluation algorithm, this may mean more than halving the perceived loudness. (+3 dB(A) is equal to double the volume).
On a side note, the Exhaust fan benefited the most from the removal as there, because of the orientation, the Noctua blades were moving the closest to the fanguards.

 

[KeesMacPro]

When you don‘t compare the sound levels at fixed fan speed values but fixed temperatures, it better explains my subjective impression.

I fully agree , and indirectly I did confirm your impression.
It might be complicated though to get exact data testing fixed Temps...
Besides there are many parameters which may differ for every case e.g. hardware setup , ambient T, type of load etc..

I liked your idea a lot and I guess lot of MP users would not consider to mod original hardware.
I'm always in for an outside the box idea !

Edited post #341

 

[Pressure]

Just ran some tests to see what happens when the fanguards are removed.
I recorded the fan noise with an AKG condenser mic placed in front of the MP at a distance of 25cm.
Tests were done with Intake and Exhaust both equally set to different fixed rpm on a MP 4,1>5,1 DUAL X5690 running Mojave.

I attached some screenshots but if TLDR:
Noise is indeed a bit lower ~ 1,5 dB , which is not much, but something very interesting appears as stated by @arw :
Apparently because of improved airflow my Intake runs in this season (Winter) at ~ 950 rpm at idle, but after removing the fanguard it's now running at ~ 750 - 800 rpm (the fan is controlled by the NB diode).


My conclusion: excellent idea !

The decibel scale is logarithmic so 1.5 dB is a lot, especially for background noise.

A 3 dB increase is twice the sound energy and a 10 dB increase means the sound energy is increased by a factor of 10.

 

[novaestudios]

Wow! Just WOW! Thank you all so much for this thread! Especially @amedias and @KeesMacPro but all who added to this thread/topic with such great and practical info. I own a Mac Pro 5,1 2010 edition since it came out in 2010. Use it as my main studio computer, and love it! I have upgraded my double CPU, RAM, PCIe NVME, etc, so it is running Big Sur etc. Love it that we can still get such an 'old' machine up to par.

I LOVE that there is such a committed MP enthousiast community. Hope to be using my MP for many years to come.

And since we just entered another 4 weeks of lockdown here in NL, I want to do something nice this Christmas holiday, and I decided to mod my Mac Pro more. Starting with swapping out as much fans as possible, trying to get my MP as silent as possible.

Still reading through this long thread, trying to decide what my upgrade path will be with which fans.
So if I am not mistaken, we can swap the CPU fans, but it's better not to do that? @KeesMacPro?

Ok, reading on!

 

[KeesMacPro]

Wow! Just WOW! Thank you all so much for this thread! Especially @amedias and @KeesMacPro but all who added to this thread/topic with such great and practical info. I own a Mac Pro 5,1 2010 edition since it came out in 2010. Use it as my main studio computer, and love it! I have upgraded my double CPU, RAM, PCIe NVME, etc, so it is running Big Sur etc. Love it that we can still get such an 'old' machine up to par.

I LOVE that there is such a committed MP enthousiast community. Hope to be using my MP for many years to come.

And since we just entered another 4 weeks of lockdown here in NL, I want to do something nice this Christmas holiday, and I decided to mod my Mac Pro more. Starting with swapping out as much fans as possible, trying to get my MP as silent as possible.

Still reading through this long thread, trying to decide what my upgrade path will be with which fans.
So if I am not mistaken, we can swap the CPU fans, but it's better not to do that? @KeesMacPro?

Ok, reading on!

Thx Nova, much appreciated and glad to be of some help!

The main reason for replacing the OEM Delta fans (as you probably noticed) is to get a lower noise level.
It's virtually impossible to find fans with such an extended rpm range as the Deltas (and the Deltas are excellent quality too).

Finding fans that run fine at min rpm or in a higher rpm is not difficult, but both at the same time appears impossible.
So by analyzing the Temps and rpm range for every individual case , one can choose a fan with a sufficient rpm range when under heavy load.

Anyway, to answer your question : yes, it is technically possible (of course) but there are no adequate fans for the DUAL CPU tested and reported here until now.
I've burned my fingers a couple of times buying fans that were not 100% working in the MP , so 3 weeks ago I send an email to Blacknoise requesting more detailed data of some fans , but unfortunately never got a response...
Otherwise, perhaps I would have reported a new positive find here.

Generally the OEM CPU fans in the DUAL are quite silent though, and in case the replaced Intake/Exhaust fans might run on their max RPM at heavy load , the CPU fans will function as a safety net because of there extended rpm range.

Hope to have answered your question , enjoy modding the one and only Mac Pro!

 

[Minimum91]

Came back to report on NF-A8 FLX for Booster fan on the single CPU model. Finally received it yesterday and installed it right away. It doesn't work well - it has the pulsing problem, where on default SMC settings it pulses all the time whilst on low temps and you can audibly hear it. It does not stop rotating, but it's slows down and then boostes back up. Moreover, it is actually quite a bit hummier on low speeds vs the NF-R8 which I found surprising. It does move air better though, when using custom configurations via software. Eventually after several hours of testing, I removed it without hesitation as I could definitely hear audible difference even when walking past my room in the apartment.

Bonus treat : Wanting the machine as silent as possible I went on to change the stock shroud of the RX580 Pulse in this machine to two NF-A9x14 Noctuas. It worked out really well and is now twice as silent as with stock fans. Unfortunately I don't have any measurement equipment to provide real data.

 

[KeesMacPro]

Came back to report on NF-A8 FLX for Booster fan on the single CPU model. Finally received it yesterday and installed it right away. It doesn't work well - it has the pulsing problem, where on default SMC settings it pulses all the time whilst on low temps and you can audibly hear it. It does not stop rotating, but it's slows down and then boostes back up. Moreover, it is actually quite a bit hummier on low speeds vs the NF-R8 which I found surprising. It does move air better though, when using custom configurations via software. Eventually after several hours of testing, I removed it without hesitation as I could definitely hear audible difference even when walking past my room in the apartment.

Bonus treat : Wanting the machine as silent as possible I went on to change the stock shroud of the RX580 Pulse in this machine to two NF-A9x14 Noctuas. It worked out really well and is now twice as silent as with stock fans. Unfortunately I don't have any measurement equipment to provide real data.

Sorry to hear the NF-A 8 FLX is pulsing on the SMC min rpm.
Good to report here as this will help others though ?

Wow, nice idea to mod the Sapphire Pulse , although at low graphics load at 0 rpm, the original fans are easily audible ( some midrange howling wolf at heavy load).

The Noctuas look quite flat is this a 15mm version?

 

[Minimum91]

Sorry to hear the NF-A 8 FLX is pulsing on the SMC min rpm.
Good to report here as this will help others though ?

Wow, nice idea to mod the Sapphire Pulse , although at low graphics load at 0 rpm, the original fans are easily audible ( some midrange howling wolf at heavy load).

The Noctuas look quite flat is this a 15mm version?

At lows speeds the A9x14 are basically inaudible. Yep, these are 14mm. I chose these versus the other recommendations I found online so that I might be able to add another PCI card in the future if I need to. With the fans attached the card, it measures just about the same height as in original form.

 

[novaestudios]

This is a fantastic thread! I just started doing the pin swap on my 120mm Noctua NF-A12x25 ULN fans. Can someone confirm whether or not I have it right? TIA

View attachment 995726

How do you pry out the busses out of this connector? I'm trying and trying, but they won't come out. Any tips?