Commander PRO + ML120 unstable RPM regulation

Hi.

I have a problem controlling my 4x ML120 fans with my new Commander PRO.

The RPM regulator is very unstable when using custom curve.

 

For example if I create a simple curve from 0-37 degree == 500rpm and 45 == 2000rpm and I play a game and the water temp stabilizes to 39 degrees requiring ~900rpm from all fans.

All fans will switch between 800 and 1000 rpm every second instead of being stable on 900rpm. (water temp is rock solid during this spin up / spin down cycle)

You can see the fans are increasing and decreasing in speed visually and it is very distracting to listen to.

It is pretty obvious that the regulator in the commander PRO is poorly configured and the feedback from the RPM causes it to adjust up and down every second.

If I used fixed % to set 900RPM all ML120's are rock stable at the RPM value.

 

My motherboards single PWM output which uses pure PWM value control also has no such problem.

It does not use RPM feedback, simply outputs 0 - 100% PWM based on temps.

But I bought the Commander PRO because my motherboard has to control my water pump.

 

Does anyone have any ideas on how i can resolve this?

 

Picture: https://ibb.co/idG7Rk

I have same problem.

 

Fan rpm is suddenly became 0, and fans spinning again. That sounds very annoying...

 

http://i.imgur.com/lOq9GGz.png

Are you sure that the fans are set to your cooler temp rather than the cpu temp? This sort of fluctuation usually occurs when the fans are set to react to the cpu temp.

Are you sure that the fans are set to your cooler temp rather than the cpu temp? This sort of fluctuation usually occurs when the fans are set to react to the cpu temp.

 

Fans are setted to coolant temperature.

This is unusual behavior. I've submitted a bug for this to our bug tracking system. In the meantime, I would suggest trying to alter the fan curve to avoid that 900 RPM dead zone.

 

The higher fans rpm is, the frequent that problem arises. >1400 is major issue rather than around 900.

 

so, pump is more big problem. pump's minimum RPM is much higher.

 

http://i.imgur.com/l65eTA0.png

 

 

 

and suddenly pump's RPM reachs to up to 16K...

 

http://i.imgur.com/03keZBM.png

Thanks for creating a issue in your bug tracking system.

 

And just to clarify.

 

If I set one of my fans to run at 900RPM with 'Fixed RPM' option the fan will switch between 850-950. (as shown in my first picture)

If I set the fan to a fixed % that matches 900PRM with the 'Fixed %' option it is stable.

This should remove any doubt about what temperature is being controlled after since it is not using any temp in these two setups.

I have tried this test with all 4x ML120 Fans.

I normally run all fans at 500RPM in idle and the optimal RPM for gaming is 900RPM in my setup.

 

Current issue list (ranked):

1. Cannot keep the RPM stable on my fans for some ranges (like 900RPM mentioned in this post)

 

2. During operation when the custom fan curve is slowly changing RPM all fans are being turned off for a second once in a while. (LED lights on my ML120 also turns off) I also encounter sometimes that instead of turning OFF it goes full ON --> Max RPM... Scares the **** out of me sometimes since it mounted on the wall 1-2 meters from me :)

 

3. I have a PWM controlled waterpump where the RPM feedback doesn't work. I would have liked to be able to force PWM mode and set a simple custom curve based on % instead of RPM.

This custom curve would also have been an acceptable solution to issue 1)

 

4. I miss some kind of hysteresis setting. I would like my fan speed to be 500rpm and then rise rapidly to 1000rpm at a certain temp.

But if I make that it only runs 1000rpm for 2-3 seconds, then it goes back to 500 rpm due to temp under limit. And 30 seconds later it goes back to 1000rpm for 2 secs ...and so on.

 

I really like the hardware in the Commander PRO! and Corsair Link has improved quite a bit from the last time I used it.

But as when I tried the Commander Mini more than a year ago I will be returning the PRO to the store and get my money back since I expect more from a PWM controller at this price point.

 

Fingers crossed for Commander ÜBER (or whatever the next gen will be called :biggrin:)

Glad you could use my feedback Dustin.

 

Perhaps I will consider the PRO again if I can find it cheaper on black friday.

The 100$ I paid for the one I had was a bit too steep considering the issues i found. (Live in Denmark)

 

But keep improving the Commander, at some point i would love to have one box which can control all 4 fans and my water pump as I want it :)

 

For now I will try and add a temp sensor to my old beagle bone black and see if I can code my own temperature / PWM controller.

 

Cheers

@Corsair Dustin

 

any chance this can be fixed with a firmware or CL update ?

 

I'm thinking of getting a bunch of ml pro fans and the corsair commander pro to control them all (I actually want rgb versions but the ml fans are better performing than the current HD ones) or should I hold off and wait for the next product release in a month or 3 (not asking for you give specifics about upcoming hardware release) I don't want to spend money on the commander and the ml fans if it wont do the job and the fans are useless without a PWM controller (because old motherboard without any pwm headers)

Thanks for creating a issue in your bug tracking system.

 

And just to clarify.

 

If I set one of my fans to run at 900RPM with 'Fixed RPM' option the fan will switch between 850-950. (as shown in my first picture)

If I set the fan to a fixed % that matches 900PRM with the 'Fixed %' option it is stable.

This should remove any doubt about what temperature is being controlled after since it is not using any temp in these two setups.

I have tried this test with all 4x ML120 Fans.

I normally run all fans at 500RPM in idle and the optimal RPM for gaming is 900RPM in my setup.

 

Current issue list (ranked):

1. Cannot keep the RPM stable on my fans for some ranges (like 900RPM mentioned in this post)

 

2. During operation when the custom fan curve is slowly changing RPM all fans are being turned off for a second once in a while. (LED lights on my ML120 also turns off) I also encounter sometimes that instead of turning OFF it goes full ON --> Max RPM... Scares the **** out of me sometimes since it mounted on the wall 1-2 meters from me :)

 

3. I have a PWM controlled waterpump where the RPM feedback doesn't work. I would have liked to be able to force PWM mode and set a simple custom curve based on % instead of RPM.

This custom curve would also have been an acceptable solution to issue 1)

 

4. I miss some kind of hysteresis setting. I would like my fan speed to be 500rpm and then rise rapidly to 1000rpm at a certain temp.

But if I make that it only runs 1000rpm for 2-3 seconds, then it goes back to 500 rpm due to temp under limit. And 30 seconds later it goes back to 1000rpm for 2 secs ...and so on.

 

I really like the hardware in the Commander PRO! and Corsair Link has improved quite a bit from the last time I used it.

But as when I tried the Commander Mini more than a year ago I will be returning the PRO to the store and get my money back since I expect more from a PWM controller at this price point.

 

Fingers crossed for Commander ÜBER (or whatever the next gen will be called :biggrin:)

 

Same issues here.

Using H100i v2 (firmware 2.8.0.0) and Commander PRO (firmware 0.2.136, bootloader 0.3).

Fans connected to H100i are fine, but those connected to Commander are insane. Randomly drop to 0 and accelerate to max, and I couldn't find any pattern with that.

And when I set all fans at 800, one of the fan stays at ~1135, while others vary between 780~820.

I like to add to the issue since I spent some time finding out why my fans would stop and start at around 750rpm. I found that if I set a fix rpm of 768 my fans would just stop but any other rpm is fine. I even made a video of the issue [ame=https://youtu.be/9aqUuvQl56E]HERE[/ame].

I found that if I set a fix rpm of 768 my fans would just stop but any other rpm is fine.

1. Good find :praise:, I have reproduced the issue using SIV and found that 512 RPM + 1024 RPM are also wonky. Please will you see if you get the same effect?
   
2. When I set 768 RPM then the fans don't even start :eek:.
   
3. When I set 769 RPM for a 10,000 RPM fan the speed is about 1,687 RPM :(:
   
4. I also just noticed 512 RPM + 1024 RPM also both end up as 0 RPM :o:.
   
5. All of #2 + #3 + #4 must be CLCP 0.2.136 firmware issues :o:.
   
6. Had I developed the CLCP firmware to me it would be obvious that such speeds as 256 + 512 + 768 + ... should be validated and I am at a loss as to how the Corsair validation process failed to validate these values :sigh!:.
   
7. I suspect any time the low 8 bits are zero (the low byte is zero) things may go wonky :!:.
   
8. I don't have the CLCP hardware datasheet (Corsair please can I have it + the CLCP protocol specification so I can look into implementing a work around?), but at a guess the firmware engineer missed some aspect of how to setup the fan controller.
   
9. I would also still like the Corsair RGB DIMM datasheet and given SIV can already control these on X399 systems (CL 4.8.2.1 can't) the last reason Corsair specified when declining to provide this was and still is inappropriate so I ask Corsair to reconsider.

So just so you guys know, we're aware of this bug, but because of the issues we've seen with H100i/H80i and the current version of C-Link, this has been backburnered to get more pressing problems fixed.

1. This has even happened before with the H100i LED settings and members needing to use such as 255,1,1 rather than 255,0,0 because of a similar firmware issue, see http://forum.corsair.com/forums/showthread.php?p=709734 :rant2:.
   
2. Is the issue still "backburnered" after almost three years? I feel it should have been fixed by now and hope a CLCP fix takes rather less time.
   
3. I wonder if it has not been fixed as Corsair don't have the firmware source, see http://forum.corsair.com/forums/showthread.php?p=793290.
   
4. Do Corsair have the CLCP firmware source?
   
5. Companies such as Intel and AMD have publically available lists of known issues and expect Corsair should also provide these.
   
6. I suspect it would be a rather long list and wonder, "Why don't Corsair provide this?" :naughty:.

1. Good find :praise:, I have reproduced the issue using SIV and found that 512 RPM + 1024 RPM are also wonky. Please will you see if you get the same effect?
   
2. When I set 768 RPM then the fans don't even start :eek:.
   
3. When I set 769 RPM for a 10,000 RPM fan the speed is about 1,687 RPM :(:
   
4. I also just noticed 512 RPM + 1024 RPM also both end up as 0 RPM :o:.
   
5. All of #2 + #3 + #4 must be CLCP 0.2.136 firmware issues :o:.
   
6. Had I developed the CLCP firmware to me it would be obvious that such speeds as 256 + 512 + 768 + ... should be validated and I am at a loss as to how the Corsair validation process failed to validate these values :sigh!:.
   
7. I suspect any time the low 8 bits are zero (the low byte is zero) things may go wonky :!:.
   
8. I don't have the CLCP hardware datasheet (Corsair please can I have it + the CLCP protocol specification?), but at a guess the firmware engineer missed some aspect of how to setup the fan controller.
   
9. I would also still like the Corsair RGB DIMM datasheet and given SIV can already control these on X399 systems (CL 4.8.2.1 can't) the last reason Corsair specified when declining to provide this was and still is inappropriate so I ask Corsair to reconsider.

 

Just tried 512 and 1024 rpm, the fans stop on them too.

 

Come on Corsair let the guy have his data sheets since his software does more for me then your own like for example being able to group my fans or just have them as an led strip and also Siv can control my pump fully since it's not restricted to 4000rpm while the pump goes to 4800rpm. Thanks red-ray for that mate .

it's not restricted to 4000rpm while the pump goes to 4800rpm.

 

What is rather strange is that the CLCP firmware restricts RPM mode to 4000 RPM, but > 4000 RPM can be used for a Custom Curve. Note the Set 4001 RPM error 0x12 for CLCP Fan 6 (CLCP-L 10K-6) below.

 

I guess you must use a Custom Curve, do you?

 

Corsair have failed to provide a publically available specification for what the CLCP firmware should be capable of and feel they should.

 

What is rather strange is that the CLCP firmware restricts RPM mode to 4000 RPM, but > 4000 RPM can be used for a Custom Curve. Note the Set 4001 RPM error 0x12 for CLCP Fan 6 (CLCP-L 10K-6) below.

 

I guess you must use a Custom Curve, do you?

 

As usual Corsair have failed to provide a publically available specification for what the CLCP firmware should be capable of and feel they should.

 

Ya I use custom curve based on water temps.

 

Hello,

 

I have a question relating to this issue: Are the 0 RPM drops at values such as 512, 768, etc. happening even when your custom curve does not specify these numbers?

 

For example, two set values on your custom curve are 750 and 800: would the fans stop spinning, then spin back up during the transition between these two set points as the fans pass 768 RPM?

 

It seems to me that if this does happen, then this is a major issue, and if not then it is rather easy to avoid these power of 2 numbers, at least until there is a fix.

 

 

 

By the way, @Corsair Dustin, really nice job responding to people on these forums. On all the major issues I've found you're there providing insight. The same couldn't be said a few months ago. I really hope you can help get these issues across to the software team.

I have a question relating to this issue: Are the 0 RPM drops at values such as 512, 768, etc. happen even when your custom curve does not specify these numbers?

 

I think that is does and this is why the fans keep stopping.

 

Below you can see CLCP-L HD-1 uses a Custom Curve and should be spinning at 768, but it's zero.

 

OK, I used numbers to make it easy to show the effect, but expect it will always happen.

 

Well, that is a serious problem. I truly hope that is fixed soon as I'm expecting my CLCP today :p

Well, that is a serious problem. I truly hope that is fixed soon as I'm expecting my CLCP today :p

 

Yes, same here. I probably won't bother to install it until this is fixed as I'd actually like to have my fans work at least as well as the motherboard+temperature sensor configuration that I am currently using.

Corsair Dustin has been working tirelessly with our dev team making sure that you guys' findings are passed on. He's on PTO at the moment and will be back sometime this week.

 

If you're having problems with the fan's rpm using the custom curve, I'd suggest that you use the pre-set profile in C Link for now while we evaluate this odd behavior. The moment we sort this out, we'll share the update with you guys right away.

 

Many, many thanks, Greybeard (and of course Dustin)! I really appreciate it. I ended up installing my CLCP, and it is very nice. The only thing that is a problem is this "fan stop" issue - if it didn't have this it would be wonderful!

 

Please let the dev team know that they will totally be my heroes when they fix this :biggrin: :biggrin: