Controls do not follow the Curves settings

[Karmalakas]

Describe the bug
With the auto-calibration feature, there was a bug introduced (I believe). Meaning - at least v184 v185 and v186 has this issue.

In the screenshot you can see, that based on the first curve, first control should also be at 0%, but almost always it keeps ~50%. No clue where these values are coming from.

Is there a log.txt file next to FanControl.exe with recent date entries?
Last log entry is days old and log itself is very short

Relevant hardware specs and setup
MSI MPG Z790 EDGE TI MAX WIFI
Gigabyte nVidia RTX 4070Ti

All fans in BIOS are set to full speed and there's no other MSI software installed. As mentioned, it worked fine with older versions.

Also was thinking, that it maybe might be similar to this report, but it's a bit confusing, because can't see curves there, so not sure

[Rem0o]

Your graph is not in RPM mode, so the calibration feature is not at work. How far back version wise do you need to go for it to work as expected?

[Karmalakas]

Same happened with Auto curve. Actually that's why I tried switching to a Graph - thought maybe it'll work.
Does it really have to be in RPM mode? I mean Graph shows percentage, right? I assume control can pick it up, can't it?

I would need to install previous versions I suppose to test which one still works. I would like to avoid that so much, but if it's necessary, I might try and find some time

[TiRoZh]

Am experiencing the same issue on the same board (Edge Ti Max WiFi). All fan speeds seem to be incorrect. I'm not sure if it's the actual reporting of the RPM that's incorrect or the actual speeds being set.

To illustrate via example, I have a series of 3x P12 Max fans connected to one header. If I manually set the calibration to increments of 165RPM (to a max of 3300RPM at 100%) at 5% intervals, the reported RPM never dips below ~45% (specifically, the reported RPM seems to never drop below ~1500 RPM), despite my 'speed' sensor reporting the fan is running at an expected of 30% (which, according to the manually set calibration, should be 990RPM). If you set a manual control and set it to 50%, it will plop the fan at ~2100 RPM (~64% of the 'calibrated' max of 3300). From my testing, this issue continues whether 'Smart Fan' is enabled or disabled in BIOS across any sensor (all, some, or none).

I have noted, however, that the noise level of the set speeds seem consistent with whatever I set in BIOS. That is, if I were to set my fans at a consistent 45% in BIOS, they are audibly louder than when Fan Control is claiming they are running at ~1500RPM (despite reporting 30%). If I manually set my fans to run at 30% in BIOS, they are approximately as loud as the aforementioned FanControl setting. This is much easier to notice for me, because the scaling of loudness for the P12 Max is pretty... well, noticeable.

I have read reports of HWInfo causing conflicts with LHM, as well as MSI services doing similar things. I have tested with one, and both, of these disabled upon boot and this issue persists. I don't think this is a software conflict.

All this to say, it almost seems as if the settings are actually being applied correctly (i.e. the reported 30%), but the RPM displayed is what is inaccurate. Given that this problem seems consistently reported exclusively across MSI boards, I assume it's something related specifically to MSI's reporting of fan speeds. I feel confident in saying this, because even if I set my fans to run at a consistent 30% in BIOS (again, 990RPM), even HWInfo will report them running at ~1500RPM (again, ~45%) -- with or without FanControl running.

I will also note that I find it interesting that, when Smart Fans are disabled in MSI BIOS, it gives a very explicit warning that manually setting your RPM to consistently stay anything below 45% may cause issues. They specifically say 45%. Pretty interesting that the reported fan speeds also never dip below ~45%, wouldn't you say?

This issue persists as far back as v179 across both NET 8.0 and NET 7.0 versions from what I've tested (and I would imagine it continues to do this in prior builds given the changelogs).

[Rem0o]

@TiRoZh

I have a series of 3x P12 Max fans connected to one header

This may cause issues. Which of the 3 fans does it feed the header with the RPM sense wire? One? The 3 signals mixed all in one wire?

If I manually set the calibration to increments of 165RPM (to a max of 3300RPM at 100%) at 5% intervals

Why do you do a manual calibration? Does it differ from what the auto calibration does? Do you have a start%/stop% set on your control ? To be clear, you can't "invent" or "force" a calibration, you "read" a calibration from what the fan does, not the other way arround.

[TiRoZh]

@Rem0o

This may cause issues. Which of the 3 fans does it feed the header with the RPM sense wire? One? The 3 signals mixed all in one wire?

It receives one signal (specifically from the uppermost fan). It uses a 3x PWM splitter, receiving speed signals from only one connection.

Why do you do a manual calibration? Does it differ from what the auto calibration does? Do you have a start%/stop% set on your control ? To be clear, you can't "invent" or "force" a calibration, you "read" a calibration from what the fan does, not the other way arround.

Yes, I do find that the auto calibration is very inaccurate for me. Even when calibrating a single fan connected to a header, I find it will set repeating RPM values for various ranges (e.g. an NF-A14 iPPC-3000 will be calibrated to 1800 RPM for 50%, 60%, 70%, and 80%). In an attempt to avoid this (since there doesn't appear to be a way to manually remove an auto-generated calibration AFAIK), I manually set my calibrations to each fan's specs (like what I have listed w/ the P12 Max). I set a start of 5%, and a stop at 0%.

From my understanding (you'd probably know better than me, though), this allows the fan to stop when it hits 0% (which is set to 0RPM anyway, and it never actually hits), and start at 5% (in the case of the P12 Max, at 165RPM, although it's never actually set to go this low). The curve is then set to a graph (mapped by %, not RPM). I would assume, then, when the graph says "Hey, you're at 40c so let's set to 30%', it would then be mapped to the manually set RPM value at 30%. I also have 'Force Apply' enabled as it recommended to fix the issue with something interfering with the curve being set (although I do find it has not changed anything whether enabled or disabled).

Let me know if I've been interpreting this wrong!

I have attached screenshots in case my explanation is unclear.

[TiRoZh]

Adding an update:

Upon further inspection, it seems that MSI boards will default 100% RPM to a whopping total of 15000 (not 1500, but 15000) RPM regardless of fan. The scales for which percentages are set are completely off. 100% RPM will refer to the maximum speed of the fan (despite the scale displaying 15000), but setting my fans to 35% (whether through BIOS or through FanControl) will achieve the equivalent of 50% (i.e. 1650 RPM for a 3300 max. RPM on a P12 Max), and a setting of 30% for whatever reason correlates to an actual value of ~45% (i.e. ~1500 RPM for the aforementioned P12 Max).

I have no idea why this happens, if it's a setting in BIOS, if it's just MSI's mistake, or if there's an issue with the fans themselves. This is consistent across all fans connected to any fan header on the board though (whether individual, daisy-chained, across different fan models, etc).