Fan RPM anomylous readings

[GoogleCodeExporter]

Sometimes I get an unrealistic high RPM reading for one of the fans and that 
becomes the Max value until I reset Max.  It would be nice if polled RPM sample 
beyond 25% change from previous polled sample would not update the displayed 
value, but two sequential polls within new range would cause it to update.

Original issue reported on code.google.com by old...@ymail.com on 8 Jun 2011 at 6:06

[GoogleCodeExporter]

I agree, but I'd suggest more than 2 successive readings.

I've recently seen more motherboards that report HIGHLY anomalous data from 
time-to-time...even in the BIOS Settings (usu. "health" or similar menu) 
displays.

These anomalous readings can be both high or low, and can cause FALSE ALARMS 
unless the monitoring programs "massage" the data to eliminate them.

BTW, on one motherboard, I was able to eliminate this problem with the CPU fan 
RPM report/alarm by changing the CPU fan from using PWM control to 
voltage-level control (a capability luckily provided by a user-setting in a 
BIOS Settings menu). I would PREFER to run the CPU fan using PWM, but such is 
life.

Original comment by transgen...@gmail.com on 12 Sep 2011 at 6:12

[GoogleCodeExporter]

I should add that the anomalous readings persisted through tests with different 
CPU fans running in PWM control mode. This strongly suggests to me that this is 
a motherboard firmware and/or hardware problem. It may be common to a specific 
firmware, or specific integrated circuit manufacturer's chip that is shared by 
various motherboard manufacturers (e.g. Gigabyte, ASUS, etc). I suggest there 
is likely a timing issue that interferes with the sensor reading update code 
that causes false RPM data to be acquired by the firmware.

Original comment by transgen...@gmail.com on 12 Sep 2011 at 6:29