Thinkpads Forum

@danda821: start the taskmanager -> create a new task (search for program-execution) -> choose "at computer/system start". First that's all. Additionally you can change advanced settings.

For practical purposes:

• settings -> end of task -> disable timer
• energy-settings -> disable all

greets
kw

I follow you procedure. Ctrl-Alt-Del -> task manager -> New Task -> Browse -> ...\fancontrol.exe -> Ok. But there is no option for "at compter/system start". Thank you.

Sorry I just have a german windows and don't know exactly the available english options. First I meant taskscheduler not manager. There must be something like "at windows/system start" even that what fantool shall do.

Guys, just place a shortcut to tpfancontrol in user' s autostart folder in startmenu. You also have to make sure that "StartMinimized=1" and "Active=2" are set in the .INI file.

In german: Mach im Autostart Ordner (Start, Programme, Autostart) einfach eine neue Verknüpfung zu Fancontrol. Ausserdem musst Du in der .INI Datei "StartMinimized=1" und "Active=2" setzen.

@vpn-user: too simple in order to work. I've tried this alike a registry entry (hklm\...\run). No way. With the taskscheduler it work's.

Thats exactly the way I do it and it works as intended. After logon, the program sits in the systray and does the job when put into user' s autostart folder.

Of course the program is not running while the logon screen is displayed. It runs that way only when logged on.

That' s why I think running it as a system service would be a good idea. System services run all the time when the machine is up. You don' t have to care for user priviliges to run it (Of course only Admins are able to install it) and it would restart automatically if crashed.

Question from Marcus: "@Aristotle: When starting, does the program show your FanBeep settings correctly in the Log window? Also, do other values (longer, different frequency) work? "

Answer from Aristotle:
The log shows "FanBeep= 0 0" in both 0.15 and 0.17, and yet I get a beep in 0.17. I think I get the beep when the fan turns off, not when it turns on.

Best,
Aris

@vpn-user: I don't know the reason therefor. I've tried both ways a several times using correct ini-settings. The heck with it I have a solution.

Hi,

I just can't run v017, it displays something like "missing or can't load WINIO" etc.

But v011 works well.

Why?


Best,

lywsp
2668-T43 ch8, WinXP Pro Sim Chinese.

Have you extracted *all* of the files in the *.zip to the same folder and are running the program from there? That would be the cause of that error.

christopher_wolf wrote:Have you extracted *all* of the files in the *.zip to the same folder and are running the program from there? That would be the cause of that error.

Yes,I have, and the file WINIO.SYS is in the same folder.

No problems here using v017 on a T43 2668F7G. Works like a dream... Only wish now is to be able to control the slight pulsing on the lowest fan level. This fan level would be inaudible if it weren'f for the buzz every 5 seconds that sounds like a bee has got stuck inside the computer.

@geoffrey: did you try to choose a zero value far beneath your lowest temperature level so fantool has to work contiuously on level 1? I don't have this pulsing problem but without fan the temperature increases too fast. Meantime I'm accustomed to this.

New version.

See README.TXT in the 2nd post of this thread for changes.

Current Version 0.18 (download the "Windows Binaries" archive):
https://sourceforge.net/project/showfil ... _id=153962

Source Code (look for the "Source" archives [use "View all project files"]) here:
https://sourceforge.net/project/showfil ... _id=153962
(MS Visual Studio 6.0 project)


As always:

README.TXT updated in 2nd post of this thread (version changes, tested models, some adjustments etc.)

Please read the README.TXT before running the program for the first time. This is a bare bones tools with no bells or whistles (you'll need to modify the ini file to be able to access the function to turn chang the fan).

Also see the links section on the readme.txt for tech background.

If you want to discuss temperature settings, sensor locations etc. please use this thread: http://forum.thinkpads.com/viewtopic.php?t=17733

Questions, suggestions, and reports of successful tests on various untested models (see readme) should be posted here.

Enjoy


Markus

Markus, it looks like you might have packaged .17 in your current .18 Zip distribution.

Please try again now (file should now be v018b.zip). Sorry for the error.


Markus

v018b works great for me...no more beeps! -Aris

geoffrey wrote:No problems here using v017 on a T43 2668F7G. Works like a dream... Only wish now is to be able to control the slight pulsing on the lowest fan level. This fan level would be inaudible if it weren'f for the buzz every 5 seconds that sounds like a bee has got stuck inside the computer.

Hi Shimodax,

I guess that's a one more call to solve the pulsing problem. The problem actually is more noticeable when runing fan at level 1 or 2 than others.

I believe you have access/control to all the other bits of that register, so twisting the bit for the disengaged mode shouldn't be a big problem, right?

Frankly, I'm not sure will that simple algorithm work or how better it turns out to be, but it is worth a try. Applying this algorihtm at fan level 1-2 should be harmless because you certainly will run fan at level 3+ when the temp. hits like 55C.

@gzt

I'll see if I can make this an experimental feature once all the other stuff works (if 0.18 proves to be solid).

It's probably not too hard but not unlikely to introduce new problems (handling of failiure to set the fan, interactions with other software etc.).



Markus

Finally! Kudos to Markus!

Some sensors do not give signal on my T40 (details in signature).

CPU 36°C (0x78)
APS 44°C (0x79)
PCM 34°C (0x7a)
GPU 41°C (0x7b)
BAT 33°C (0x7c)
BAT 29°C (0x7e)
BUS 0°C (0xc0)
PCI 0°C (0xc1)
PWR 0°C (0xc2)
XC3 0°C (0xc3)

yeah, I agree, the "can't read status error" could be a headache. So far v0.18 runs well except gives such message. I run nhc at background.

Anyway, if all the problems are sloved you should ask IBM to pay you.

Why the NHC writer won't tell you the way they read sensors? If we know it we may have algorihtm avoid the conflict.

...

I also believe the waiting time (5s) after a read failure perhaps too long. In case there are consecutive reading /writing errors, like 5 errors, should we wait 25 seconds to react? That should be fine if we don't run intensive stuff, but could be potential disaster in some cases.

Did you try using a loop with small time steps(e.g, 0.1 sec.) after a reading/writting error for about 3 seconds? Will that work better? Since the conflicts is not that frequent the loop overhead shouldn't be a problem.

I might be totally wrong without digging into your code. ignore this if that's the case.

Some more analysis. The program works fine on my T40 in spite some sensors are absent/not read.

speed setting vs fan speed :

1,2 2900-3000 rpm
3,4,5 3500-3700 rpm
6,7 4000-4100 rpm

Accordingly the following table should be better for smart mode on T40 than default:

Level=47 0
Level=50 2
Level=55 4
Level=60 7
Level=70 128

Finally would be nice to fix continuous recording of a warning message in System log in event viewer:

The embedded controller (EC) hardware returned data when none was requested. This may indicate that the BIOS is incorectly trying to access the EC without syncronizing with the OS. The data is being ignored.

George

geobel wrote:Some more analysis. The program works fine on my T40 in spite some sensors are absent/not read.

speed setting vs fan speed :

1,2 2900-3000 rpm
3,4,5 3500-3700 rpm
6,7 4000-4100 rpm

Accordingly the following table should be better for smart mode on T40 than default:

Level=47 0
Level=50 2
Level=55 4
Level=60 7
Level=70 128

Finally would be nice to fix continuous recording of a warning message in System log in event viewer:

The embedded controller (EC) hardware returned data when none was requested. This may indicate that the BIOS is incorectly trying to access the EC without syncronizing with the OS. The data is being ignored.

George

Thank you very much for all your dedicated work, Markus

Before the release of your Thinkpad T43 Fan Control I have tried to control the fan bye manipulating NHC and IBM Power Utility by changing between selfmade profiles
http://forum.thinkpads.com/viewtopic.ph ... 183#109183

But when using low CPU settings /"Battery max" I would get a compensatory rise in fan RPM, and changing to "max.performance" made the fan RPM go down, but only for so long - Because the sensors eventually triggers the fan when the temp goes up because og CPU rise.
With your tool however, I can keep the CPU/temp down AND conrtrol the fan at the same time.
I have realised that no fan makes my T43p get warm very fast, even with no wirelesss and min CPU . So I have realised that I have to accept some degree of fan activity , in my case level 2, much about the same audible level as my HD (60Gb 7200rpm, default settings with NCH). After manipulating the The Fan Control for some days now, this is my best compromise settings, much the same as yours , geobel:

Hopefully , this may be of help for other T43/T43 owners with the same level of fan noise acceptance :

(When just surfing the net / not in need of a high CPU , wireless on and running iTunes / airTunes ).

In this setting the GPU/PCI is the fan triggers, most often the GPU, and in my experience it is regularly about 2°C higher than the other sensors.

Level=48 0
Level=50 2
Level=55 5
Level=60 7
Level=70 128

FanBeep= 60 10

At the moment my readings are :

( Power setting Max Battery, Brightness level 6/7 , AC, battery charging level 99 %, USB 1 :cooling pad-fans off, USB 2: Maxtor HD 7200 RPM , room temp 22°C )

Fan speed : 0x02 (Fan Level 2, Non Bios), Highest Tp 49°C

3055-3085 RPM

CPU 47°C (0x78)
APS 45°C (0x79)
PCM 36°C (0x7a)
GPU 49°C (0x7b)
BAT 34°C (0x7c)
BAT 31°C (0x7e)
BUS 44°C (0xc0)
PCI 48°C (0xc1)
PWR 46°C (0xc2)


When the fan goes to level 5 (temp > 55) I regularly change to Manual (level 7; 4055 rpm ) and also start the external cooling pad, to get the temp/fan down very fast.
I find this preferable to waiting for a long time for the fan to cool the T43 p down at level 5 or so..

Again Markus, warm (or rather cool- ) thoughts...

T43p 2668-F8G ( 2.0GHz , 2 GB RAM , ATI V3200 128MB , 60GB 7200 RPM , 15" UXGA )

Program stops along with "Fn+F3", display off.
How to avoid that?

I just want to report that using version .18 of the program after running nearly 18 hours it crashed and left the fans off.

geobel wrote: Finally would be nice to fix continuous recording of a warning message in System log in event viewer:

The embedded controller (EC) hardware returned data when none was requested. This may indicate that the BIOS is incorectly trying to access the EC without syncronizing with the OS. The data is being ignored.

George

The system log file will reach its max size in a few seconds with the warning message GEOBEL mentioned above,after that,it can't record other events at all.
Will the next version fix the bug?

lywsp wrote: The system log file will reach its max size in a few seconds with the warning message GEOBEL mentioned above,after that,it can't record other events at all.
Will the next version fix the bug?

Well, system will always be able to record events. The problem is that any recorded event is quickly overwritten by the continuously recorded warning message.

George

Sorry about the crash ... I though I fixed one reason for that but obviously there's still another quirk somewhere in the program. We'll have to see if there's a pattern in these, meanwhile the only resolution would be to not use the program when the notebook is running unattended.

Regarding the log. The program does write a logfile to disk (with everything that's also shown in the log part of the window ... the file fancontrol.log).

Are you getting so many errors that these quickly scroll out of the window? The log window has a buffer of 100 lines, technically these should be enough for hours of operation before stuff scrolls out (it will still be in the log file though).


Markus

Shimodax wrote: Regarding the log. The program does write a logfile to disk (with everything that's also shown in the log part of the window ... the file fancontrol.log).

Are you getting so many errors that these quickly scroll out of the window? The log window has a buffer of 100 lines, technically these should be enough for hours of operation before stuff scrolls out (it will still be in the log file though).
Markus

No, we are talking about Windows System log. Can be seen via Event viewer (Control Panel/ Administrative tools/Event viewer)

George

geobel wrote:No, we are talking about Windows System log. Can be seen via Event viewer (Control Panel/ Administrative tools/Event viewer)

Yep, got that, but stuff isn't written to the system log so easily (not in matter of a few lines of code anyway), so that is something I don't want to do at the moment.


Markus