Fan RPM deviations when using RPM mode?

[tomasz1986]

I'm wondering if deviations when it comes to the actual RPM that the fan is running at, and the RPM set in its curve are normal.

For example, I have the RPM set to 300 in the curve as follows:

However, the fan's actual RPM is ~260:

The fan itself has been calibrated too:

Other fans also show differences in their numbers, e.g. 440 RPM instead of 460 RPM, and 450 RPM instead of 460 RPM, but those seem less drastic.

[Rem0o]

Yes it's completely normal. Calibration curves are not perfect, and they will change with time. If your dust filter gets clogged up, your fans will spin slower. Depending which other fans are spinning near a specific fan will also affect this. Your calibration itself has a margin of uncertainty. The algorithm then that uses that curve also has a 32 rpm error. All of this combined will result in offsets.

You may redo the automatic calibration from time to time if the values deviates too much.

[tomasz1986]

Thank you for a quick reply and the explanation! I need to say though that these deviations exist even right after calibrating the fans. I have repeated the calibration process several times, but there is no difference no matter how many times I do it.

One thing that I'm still not sure about is that in the calibration table, 300 RPM is matched to 20% speed, and this is more or less correct, because if I set the fan speed to 20% manually, the RPM does fluctuate around 300 RPM. However, when using the curve set to the same 300 RPM, the actual percentage ends up being around 17% (as seen in the screenshot).

[Rem0o]

That's probably the error/ uncertainty in the algorithm itself reading the curve, which has a 32 RPM error, which in turn makes it really really fast. Since you are running at such low speeds, that error is relatively big.