View Issue Details

Jump to NotesJump to History
IDProjectCategoryView StatusDate SubmittedLast Update
0000616SpeedFanFan controlpublic2006-08-22 14:112019-02-07 07:25
ReporterW1NA Assigned Toalfredo  
PrioritynormalSeveritytweakReproducibilityalways
Status acknowledgedResolutionopen 
PlatformXP Professional 
Product Version4.29 
Summary0000616: Speedfan constantly overrides programmed settings.
DescriptionI'm currently running Speedfan on my W1N laptop. Unfortunately, Speedfan refuses to maintain my configuration from time to time, especially after suspending or hibernating the system. I simply set up my chip (ADT7463 on Intel SMBus) to be MANUALLY CONTROLLED and switched off all the other speeds and fans but the cpu fan (Speed01). I've also indicated proper values for the Desired and Warning temperatures as well as the Minimum and Maximum fan speeds. The latter ones are supposed to be automatically variated since I've set up this option, too. To make sure nothing else interferes with this configuration I also tried out all the command line options to disable auxiliary checks such as SMART and so on. However, every time I restore my laptop from suspend or hiberantion Speedfan tends to lose these settings: just a few moments after being restored the cpu fan seems to go into business for oneself though speedfan keeps reporting its speed at only 5% which is the programmed minimum.
Then I need to restart speedfan at least once and change some of the values back and forth until I have speedfan do what I want.

Now, after changing to the new speedfan version (4.29) the behaviour described above even degraded. Not only does the programme show this side effect after every hibernation, it also loses control over the fan just a few minutes after every computer restart.

I also noticed that the fan speed seems to normalise a few moments after exiting speedfan though it is supposed to stay at the maximum level.
I made sure no other software interferes with Speedfan or controlles the ADT7463.
Then I took a look at my BIOS' power management setup. I tested all the given settings such as Automatic, Maximum Performance, Reversed and so on. Only the last setting (Disabled) would give me a proper result with Speedfan. Unfortunately, my CPU speed would constantly remain at 598 MHz despite the fact of having a powerful 2 GHz Centrino engine. That's why I had to give up changing BIOS settings.

The situation has grown pretty weird and I can't see any reasonable solution for this kind of problem, hence I really appreciate every single incoming proposal.

Thank you very much in advance!
TagsNo tags attached.
Motherboard Model
Video Card Model

Activities

alfredo

2006-08-26 11:29

manager   ~0001935

I changed the severity to TWEAK (it might be FEATURE too...). As you indicated, you were forced to allow the BIOS to change fan speeds. On a chip like yours, this should mean that the chip is instructed to do it automatically. In such a case, setting to MANUALLY CONTROLLED should do the trick and allow SpeedFan to take full control of fan speeds. I'm not sure if the issue appears only after resuming or under normal operations too. This is the most important thing for me to know.

W1NA

2006-08-27 15:34

reporter   ~0001950

The problem is I only work with the MANUALLY CONTROLLED setting on. However, it doesn't do the trick. What I don't understand is why Speedfan doesn't show any constant behaviour. It has only happened just a few times so far that Speedfan did everything as expected. By the way, I noticed I had to restart Speedfan several times in a row to achieve that. But let's get back to your question: the issue itself appears both after resuming AND under normal circumstances which I do not comprehend.
Can you see any reason why this happens?

alfredo

2006-08-27 16:33

manager   ~0001951

In its endless fight against usability, manageability and information sharing, Asus does all sort of things. To me it looks like it is hacking its own computers. The AD7463 can change fan speeds without any external intervention. If you properly configured SpeedFan (make sure you properly understood how several temperatures can act on a single fan speed) and you observe what you report, most likely there are two "agents" controlling the fan. There is little that can be done here. If you send me a report using the ACPI TABLES available in the BETA AREA and then send me a short reminder to identify it (don't put it as an attachment) I will try to check it.

W1NA

2006-08-28 03:32

reporter   ~0001955

I absolutely agree on the fact that ASUS virtually deserves to be punished for its bad service. Nevertheless, it is more important to me to try to find a solution or at least some sort of a workaround to circumvent the hacking of their own computers as you called it.
By the way, I do understand how several temperatures may affect the overall progress. That is why I just use only one single temperature source which is the standard CPU indicator in my case. The CPU fan speed is automatically controlled and varied in the range of 5 to 65 per cent. This span turned out to be the optimum for my hardware.
However, the requested report has been set up and sent to you. I really hope you can do the best of it. Let me say I greatly appreciate your work and therefore I would like to thank you for all your efforts!

2006-08-28 03:33

 

ACPI-Report.txt (74,171 bytes)   

alfredo

2006-08-29 13:41

manager   ~0001957

It looks like that ACPI is changing fan speeds on your computer. Unfortunately, it is not relying on built-in capabilities from your hardware monitoring chip. This means that, yes, you have two agents trying to control fan speeds. Unless you manage to disable one of them (guess which one? :-)), I think there is little that can be done on the safe side.

W1NA

2006-09-01 13:15

reporter   ~0001964

Hmm. I've been experimenting with the ACPI under XP ever since you told me about it. What I basically found out is the fact that you must not try to disable the main ACPI driver. Otherwise you'd have to face the same consequences I had to suffer: the system ends up failing to start properly and all you can see then is a lot of nice blue screens. Lucky me, I had made a backup prior to playing around with my OS.
Another problem is the fact that Windows reports several additional ACPI drivers which are responsible for different parts of the computer such as the notebook lid, the standby button and so on. I managed to fool the system by installing nonsense drivers (e.g. IBM Dummy Device) instead. Unfortunately, this approach didn't do the trick. From what I can tell it's not enough to circumvent these sub-drivers since they all rely on the main ACPI system driver. However, the latter one is as stubborn as a mule and I therefore failed to deactivate it safely.
Let me ask you a few questions on this. Do you see a way to learn from the ACPI TABLES report I sent to you which agent exactly is trying to control the fan besides Speedfan? If it's the ACPI driver (let's assume it for a moment), are you aware of any tricks that might help me disabling it safely?
Is there anything to be done on Speedfan's side to prevent this? I mean, I noticed Speedfan simply loses control over the fan from time to time under the circumstances I've already described. And the thing is Speedfan doesn't even seem to have a clue that it happens since it keeps reporting a wrong fan speed percentage. Is it possible to decrease the checking interval to make sure Speedfan regains control faster? Or could you even implement some sort of a field similar to the one that is used for setting the fan speed percentage? With such a field on my side I could probably adapt Speedfan to my situation helping it in winning the race against ACPI. That would be great!

alfredo

2006-09-01 14:15

manager   ~0001965

The only reliable way to disable ACPI intervention would be to modify the ACPI tables. Adding hacks to SpeedFan so that it fiercely combat any other fan controller is something I'm unwilling to add. I wouldn't be surprised if Asus would end up combatting harder and harder.

W1NA

2006-09-01 19:41

reporter   ~0001967

If modifying the ACPI tables is supposed to be the only reliable way to get rid of ACPI then how exactly can it be done?
By the way, I wasn't asking you to commence a war against ASUS. What I really need is a way to slightly alter your perception of Speedfan's timer interval. I mean, what is so bad about adding another control field to Speedfan? The main point is not the combatting any other fan controller as you said. Speedfan just turned out to be a little bit too slow compared to the ACPI driver on my system. Besides that I would like to call your attention to the fact that ASUS doesn't seem to be (at least completely) responsible for my special problem. Although there are some ASUS drivers installed I'm quite certain they aren't to blame for my concern. Unfortunately, their removal didn't do the trick at all.
As a matter of fact, the only delinquent I see (unless I'm completely wrong) seems to be once more Microsoft.

Issue History

Date Modified Username Field Change
2006-08-22 14:11 W1NA New Issue
2006-08-26 11:29 alfredo Note Added: 0001935
2006-08-26 11:29 alfredo Severity major => tweak
2006-08-26 11:29 alfredo Status assigned => acknowledged
2006-08-27 15:34 W1NA Note Added: 0001950
2006-08-27 16:33 alfredo Note Added: 0001951
2006-08-28 03:32 W1NA Note Added: 0001955
2006-08-28 03:33 W1NA File Added: ACPI-Report.txt
2006-08-29 13:41 alfredo Note Added: 0001957
2006-09-01 13:15 W1NA Note Added: 0001964
2006-09-01 14:15 alfredo Note Added: 0001965
2006-09-01 19:41 W1NA Note Added: 0001967