Lost CPU Fan Speed Sensor Reading

RE: Lost CPU Fan Spped Sensor Reading

Your right - I never looked there. On 4.20 it is at the end of the list (I have not checked the other versions - I am just running 4.20 at the moment).

OK - 4.24 works the same as 4.20

Evidently I have a "funny" board that does not report correct sensor values until I "poke" it (i. e. get it to run at a higher multiplier than 5X). After that it seems to do OK (in that I can see speed fan and HWINFO both winding down after stopping prime95 - HWINFO still hangs after a bit though).

Thanks - I think I can live with the fan speeds being reported as they are. I just have to remember to "poke" the CPU first (hmm - I will have to see if my underclock at 5X is actually working as well as I expected).

Good show - thanks for your help.
 
RE: Lost CPU Fan Spped Sensor Reading

No problem :) I glad it's resolved now, since it drove me crazy not to be able to find such issue :) I thought it was some serious flaw in HWiNFO...

As for the sensor order - when you see the CPU fan at the bottom of window, go into "Configure Sensors" / Layout and click "Restore Original Order". That will then position the CPU fan value where it should be and HWiNFO will remember this setting.

I think the problem with your board or fan config is that the fan sometimes runs at a speed which is out of the range which can be measured by the fan speed sensor.
 
The fan speed disappeared from my sensor list after going with the latest beta update - I reverted back to an earlier version and now have fan speed in the sensor list.

HP Elitebook 8460p
 
1004ron said:
The fan speed disappeared from my sensor list after going with the latest beta update - I reverted back to an earlier version and now have fan speed in the sensor list.

HP Elitebook 8460p

What's the sensor name under which you see that fan speed and what's the last HWiNFO version that works ?
 
Martin said:
1004ron said:
The fan speed disappeared from my sensor list after going with the latest beta update - I reverted back to an earlier version and now have fan speed in the sensor list.

HP Elitebook 8460p

What's the sensor name under which you see that fan speed and what's the last HWiNFO version that works ?

v5.06-2640 works

Sensor name is "Fan"

Edit: Tried v5.10-2700 and it's showing fan speed, so it's just the latest beta that's not working for me.
 

Attachments

1004ron said:
Martin said:
1004ron said:
The fan speed disappeared from my sensor list after going with the latest beta update - I reverted back to an earlier version and now have fan speed in the sensor list.

HP Elitebook 8460p

What's the sensor name under which you see that fan speed and what's the last HWiNFO version that works ?

v5.06-2640 works

Sensor name is "Fan"

Edit: Tried v5.10-2700 and it's showing fan speed, so it's just the latest beta that's not working for me.

OK, but what's the sensor heading name? Can you please attach the HWiNFO Debug File from the latest Beta?
 
I noticed a similar behavior with HWiNFO, and it appears to depend on whether the fan speed was 0 rpm when HWiNFO launched. I have an external fan cabled to motherboard fan header Sys 2, and HWiNFO usually displays it okay. But I run software -- a combination of Rem0o's FanControl app, which is like SpeedFan, plus a custom .bat file that I wrote -- that controls the speed of the external fan based on the S.M.A.R.T. temperature of my external hard drive, and when the hard drive temperature is 32C or cooler the software keeps the fan speed at 0 rpm. I can make HWiNFO display or not display the external fan sensor by manually choosing the fan speed prior to launching HWiNFO.

The solution seems straightforward: Modify HWiNFO so it won't assume 0 rpm at launch means no fan is present. In other words, display the fan sensor even if it was 0 rpm. (Unless the user configures HWiNFO to hide it, which would make sense if no fan is attached.)

Unimportant details in case someone is curious: I installed multiple copies of FanControl, each configured to set a different constant fan speed, and each associated with its own Windows Task Scheduler task. The tasks have no Task Scheduler automatic triggers but can be run on demand. My .bat file uses smartctl.exe to read the external drive's temperature, and the .bat stops & starts the tasks so that only one copy of FanControl at a time will be running... the copy that sets the fan speed that's appropriate to the external drive's temperature. The external hard drive is a 5400 rpm WD Red that's often 32C or cooler. The external drive is attached with a SATA-to-USB adapter, and FanControl doesn't know how to do USB pass-through of S.M.A.R.T. commands, but smartctl.exe does.
 
This is intentional behavior, because when a fan reports 0 RPM it's not possible to determine if it's momentary not spinning or not connected at all.
And since sensor chips often feature several fan inputs, showing a 0 RPM fan always would result in listing a bunch of fans which are not connected.
 
It's imprecise to say HWiNFO would always list a bunch of fans that aren't connected. As I wrote, the user can hide those sensors to stop them from being listed.

I believe it's undesirable for HWiNFO to behave as though fans are definitely not connected when that's false. HWiNFO fails to monitor something that it knows how to monitor, that the user wants monitored. There are alternatives better than pretending the fan doesn't exist.

One alternative would be to display fans that are initially 0 rpm in a different color by default, perhaps gray, that would mean HWiNFO is uncertain whether fans are connected. The user could move them to the list of hidden sensors (appropriate when no fan is connected) or change the color (appropriate when a fan is connected). Or the user could simply leave them gray. While not hidden, the rpm (which might not remain zero) should of course be displayed. If in fact no fan is attached, the efficiency cost (the cpu time wasted polling the sensor(s) and displaying 0 rpm) is surely negligible compared to the total number of sensors being monitored, and the waste will presumably go to zero if/when the user hides them.

Alternatively, HWiNFO could by default include fans that are initially 0 rpm in the list of hidden sensors. This would allow the user to unhide them. HWiNFO should remember the user's choice to unhide them, and keep them unhidden until the user chooses to hide them again.

Another alternative would be to provide checkboxes in one of the configuration tabs -- one checkbox per fan header -- by which the user could specify whether HWiNFO should display the fan if it was 0 rpm when launched. Or a single checkbox to specify whether HWiNFO should display all fans that were 0 rpm when launched.
 
I agree that the current solution is not convenient, I will put this change on my list.
 
Thanks!

When HWiNFO chooses not to display the 0 rpm fan sensor, an undesirable side effect is that it also forgets the layout position where I had placed the fan sensor (next to the sensors of the external drive that the external fan cools). I assume a beneficial side effect of the new 0 rpm feature will be that HWiNFO won't forget the fan sensor layout position(s).

That will make the "layout rearranging bug" less painful. (When I try to move a sensor with drag&drop or with the up & down arrow buttons, it often causes an unexpected sensor to move or moves the sensor to an unexpected position in the order. Then, trying to correct the unexpected result often has another unexpected result, and clicking the Restore Original Order button becomes the lesser of evils.)
 
Next Beta build of HWiNFO will include a new option in sensor settings - "Show all fans".

That sounds good, but what will happen if the user selects the new "Show all fans" option and also manually configures HWiNFO to hide a fan sensor that has no fan attached? Will HWiNFO remember that the user hid that sensor, and override the "show all fans" option for that sensor? I think this would make sense.

Also, for fan sensors that aren't in the main list of (unhidden) sensors because "Show all fans" isn't selected and they were 0 rpm, how about including them in the list of hidden sensors? This could provide another way for the user to unhide the one(s) that actually have a fan attached.
 
With "Show all fans" enabled, all fans including the stopped ones will be treated the same.
 
With "Show all fans" enabled, all fans including the stopped ones will be treated the same.

I hope you didn't mean the literal interpretation of "all fans will be treated the same," because the literal interpretation would be too extreme. Manual hiding of individual fans should take precedence over the "show all fans" setting. (And of course, the manual hiding should be "sticky": HWiNFO should remember which fans were manually hidden.)
 
Last edited:
The fans feature in the new beta looks great, thanks! I tested whether manual hiding takes precedence, and whether manual hiding is sticky, and the beta passed both tests.

I manually hid my unused fan sensor, and the fan sensor for my "often 0 rpm" external fan is displayed as desired.

You might consider changing the label of the checkbox to "Show 0 rpm fan sensors" to make it clearer what the option is for.

You might consider reversing the sense of the checkbox from show to hide ("Hide 0 rpm fan sensors") and have them displayed by default. I assume we can only guess whether most users would prefer to hide or show them. And I think a good rule of thumb is to avoid hiding things from people until they choose to have them hidden, because people might not realize there's more than meets the eye, and might assume what they see is what they got.
 
Back
Top