Thinkpad Fan Noise Problem: Light at the End of the Tunnel

[troubadix]

btk

@mejdlos: use fancontrol.exe as non_service application, go to task manager [Ctrl Alt Del], choose running processes, right mouse click on fancontrol.exe, increase priority to real time.

Ciao, troubadix

edit 03.19.2008: Today I introduced a new parameter in fancontrol.ini: ProcessPriority=2 (0-5, normal=2), this way you can control process system priority even in service version:

http://staff-www.uni-marburg.de/~schmitzr/tpfc_v042.zip (just beta)

[mejdlos]

btk

@mejdlos: use fancontrol.exe as non_service application, go to task manager [Ctrl Alt Del], choose running processes, right mouse click on fancontrol.exe, increase priority to real time.

Ciao, troubadix

edit 03.19.2008: Today I introduced a new parameter in fancontrol.ini: ProcessPriority=2 (0-5, normal=2), this way you can control process system priority even in service version:

http://staff-www.uni-marburg.de/~schmitzr/tpfc_v042.zip (just beta)

thanks very much. hopefully this will help also to somebody else

[Pascal_TTH]

btk

@mejdlos: use fancontrol.exe as non_service application, go to task manager [Ctrl Alt Del], choose running processes, right mouse click on fancontrol.exe, increase priority to real time.

Ciao, troubadix

edit 03.19.2008: Today I introduced a new parameter in fancontrol.ini: ProcessPriority=2 (0-5, normal=2), this way you can control process system priority even in service version:

http://staff-www.uni-marburg.de/~schmitzr/tpfc_v042.zip (just beta)

Hi troubadix !

You still there and keep this nice software alive. Thanks for your support. I'am gonna test the new one soon...

[RPF]

Hi troubadix / users of tpfancontrol,

I have a T61 and I love this tool! Since the fan is almost always running once you charge the battery it really helps to ged rid of the noise.

The tool and the options are great BUT I'd like to adjust the speed by myself. At the moment on "state 1" the fan blows with about 3000RPM in my configuration starting at 60Degrees.

Well, I would prefer if the fan would always run / start at lets say 50 Degrees but only with about 1500 to 2000 RPM. To me this would be great because 1500 to 2000 RPM makes almost no noise and the T61 does not get so warm which prevents it from higher RPM's and Noise.

What do you think about that? Would it be possible to programm tpfancontrol with this?? The question is how to decrease the RPM's at state 1?

For example: On my HP nc6910p the fan runs almost always, but so slow that you cannot hear it and the notebook stays cool which prevents it from producing to much noise. I'd love my T61 would be as silent as the nc6910p

[troubadix]

@RPF: sorry, but fan levels and rpms are fixed the embedded controller firmware. Some guys here tried to deal with that, but not so luckily:
http://forum.thinkpads.com/viewtopic.php?t=20958

@Pascal_TTH: ...waiting most curiously for your results as you are known as an experienced TPFanControl tester

ciao, troubadix

[RPF]

@RPF: sorry, but fan levels and rpms are fixed the embedded controller firmware. Some guys here tried to deal with that, but not so luckily:
http://forum.thinkpads.com/viewtopic.php?t=20958

Oh, I see...that's sad...would have been great and a powerful improvement.
I really don't understand why all or most Notebook vendors have such huge problems with fan noise, I think they do not spend time on thinking about that and their customer needs.....very bad. Apart from great products they produce, Lenovo should be more customer driven and implement such things in their "ThinkCenter".

But, thanks to such great tools like tpfancotrol this really annoying behavior can be partially mitigated.

Thanks for that

[GPR]

Hi troubadix / users of tpfancontrol,

Well, I would prefer if the fan would always run / start at lets say 50 Degrees but only with about 1500 to 2000 RPM. To me this would be great because 1500 to 2000 RPM makes almost no noise and the T61 does not get so warm which prevents it from higher RPM's and Noise.

Hello Guys,

I was thinking to a simple HW solution to reduce the minimum RPM speed. Does anybody is interested to this?

Please can somebody write a comparison table with the RPM and levels (Eg. Level1=3000RPM, Lever2:....)

I try to work on It...

[infinus]

I can see that RPM levels can't be set through the software. It would be kind of nice to have the ability to "remap" the levels however so you can have out of order fan speeds.

Example:

Colume 1 is tpfan control Level Colume 2 is EC Level
1 128
2 128
3 128
4 6
5 7
6 64
7 64

This way the program would work properly for how I like to use it, where the BIOS is in control until the laptop gets really hot and then the program takes over and bumps the fan speeds up higher then the BIOS will.

[Paul Unger]

This way the program would work properly for how I like to use it, where the BIOS is in control until the laptop gets really hot and then the program takes over and bumps the fan speeds up higher then the BIOS will.

And at the other end of the spectrum are those of us who would like to run the fan at, say, 1000 rpm @ level1 (rather than the 'factory' 3000 rpm). I look forward to seeing what you figure out, GPR! Keep us posted.

[GPR]

And at the other end of the spectrum are those of us who would like to run the fan at, say, 1000 rpm @ level1 (rather than the 'factory' 3000 rpm). I look forward to seeing what you figure out, GPR! Keep us posted.

Well.. I have a simple question for you/everybody...

Has really sense to reduce the fan to a noise-less RPM and force it to run continuosly?

The answer should be yes if a const RPM should keep the temp under the next level for the general purpose use.. (office word processing, internet......) What's your Opinion about??

Assuming you say Yes, the simplest HW possibility I see is to reduce the fan speed in a linear mode.. so if level1=3000 and level7=4000 you can understand that, if I fix the Level1=2000 the level7=3000!!

I see 2 problems on this solution:

1. If to reduce the FAN speed is so easy.. why does anybody do that before? As troubadix said me.. there is a reaction control over the tachometer wire.. so it's possible to have an EC fan error.. If this happens ---> GAME OVER!!

2. If I solve the .1, this HW solution should disengaged itself when the Fan temp reaches 65/70 degree... I do not want a bug process fries me the CPU.. so It's safe to give back the original capabilities to the Fan.

A good electronic Engineer (and not one like me..) should easely project a smart interface to do that... But there is not enought space.. and a complex circuit needs to be made in SMD components.. not easy to self made...

So I'll try to work over a simple solution...

If you do not hear any news about me.. is because a f@cking process fried my TP while testing my interface!!

Stay tuned.. cu, bye

[Paul Unger]

Well.. I have a simple question for you/everybody...

Has really sense to reduce the fan to a noise-less RPM and force it to run continuosly?

The answer should be yes if a const RPM should keep the temp under the next level for the general purpose use.. (office word processing, internet......) What's your Opinion about??

If I understand you correctly, the sense (for me) of running the fan at a low speed continuously is to combat a 'pulsing' fan, that is, a fan that constantly switches on/off at a preset threshold. A constant quiet noise is more tolerable than a sporadic harsh noise. Where I work in Solomon Islands the ambient temperature is quite hot and anything I can do to keep my computer temperature down (and the fan not on/off-ing all the time) is welcome. I hope that makes sense!

PS I hope you don't fry your computer!

[troubadix]

Hi!

as there are no reported issues til now I put version 0.42 to the link in my signature. v0.41 is still available at:
http://staff-www.uni-marburg.de/~schmitzr/tpfc_v041.zip

Ciao, troubadix

[GPR]

Hi All,
I measured the FAN Voltage supply connected the speed levels.
This is the 1st step to modify the Fan speed.
I made this analysis on a T40, so my levels should differ from T60's.
So, levels may differ as differs the RPM speeds related to a Thnkpad model, but all models uses the same FAN, so the specifications are the same (Vin max= 5V).

LEVELS . . . Vin . . . Vtach
Level_1-2 [3.65 V] [1.86 V]
Level_3-5 [4.11 V] [1.95 V]
Level_6-7 [4.48 V] [2.04 V]
level_64_ [5.01 V] [2.35 V]

You can notice that T40 has four active levels only (64 is the disengaged).

The fan PINOUT is with the central wire (BLU) as Ground, the Red is the Vin and the Orange is the Vtach.

I think the easiest way to solve the noise at level 1 is to reduce a little the Vin supply. (A little means sone hundred of RPM..)
Easy to make, really cheap and not too much dangerous for the system, because the fan loses few RPM only.

This solution is to be verified if it's possible, because there is a reaction back of the Vtach and the CE should rise an error.
This because the Vtach changes when the Vin changes.

When I have the time to disassembly my T40 again I'll try this easy solution, inserting a diode in the supply line.

Stay Tuned!

Bye
-=GiP=-

[GPR]

Hi everybody..
finally.. I made it! I modified the FAN power Supply over my T40!

The solution used is the simpliest possible.. I introduced in the Vin line (the RED wire) a simply DIODE... and.. it works!! No CE error or other problems found.

I used a Diode type 1N4007, really common. But if you need a 20% efficiency more, you can use a Diode type 1N4148.

The 1N4007 reduces the FAN power supply by 0.8V (1.1V if you uses the 1N4148 diode). It seems too much, but...

..You should consider that the FAN Controller Embedded (CE) comphensate this modification (because of the reading back the Vtach) and inreases the Fan supply around +0.3V/+0.4V more (see the table in my previous Post).
So, inserting a Diode you can effectively reduce the Vin about 0.4V (0.6V is you use a 1N4148 diode).

Here attached an image that shows how to make it. (to see it you should have the http:8080 port opened. If you cannot see it, write here your email and I'll send to you).



Now the Level1 RPM is really noiseless!


I wait for your comments....

Enjoy

[mejdlos]

btk

@mejdlos: use fancontrol.exe as non_service application, go to task manager [Ctrl Alt Del], choose running processes, right mouse click on fancontrol.exe, increase priority to real time.

Ciao, troubadix

edit 03.19.2008: Today I introduced a new parameter in fancontrol.ini: ProcessPriority=2 (0-5, normal=2), this way you can control process system priority even in service version:

http://staff-www.uni-marburg.de/~schmitzr/tpfc_v042.zip (just beta)

hi,

finaly, I got the the writing another paper and having chance to test your advise with process priority and a new version of fancontrol. Unfortunaly, I have to say that it does not work. Ms Word and Endnote still cause non-responding fancontrol even if the priority is set to realtime. It works until I enter some citations from endnote to word, then, the problem occurs.

another issue: from time to time, my computer experience some kind of serious error which is expressed by terribly slow function of harddrive or some bus. all other functions are affected: terribly slow hibernation, shutting down, loading the system, copping via usb ports etc. the only thing that helps is system restore, then computer works well but gradually slows down (but not terribly) through the time and finaly this bug occurs. I tried to look to event log but nothing obvious was there. But after the system restore, the only changes are in kernel system folder.

could this be linked to fancontrol? do you have any idea what's going on? some advice? would be gratefull for help. second issue is more important.

thanks

Lukas

[troubadix]

Hi,

I never made this experience with TPFanControl nor heard of it before, but:

if your computer slows down in reaction, go to taskmanager (Ctrl-Alt-Del in XP, right click on taskbar in Vista) go to processes and sort them by cpu usage by clicking on top the column. This way you can find out what kind of process is eating the resources. Wonder what it is

Ciao, troubadix

[mejdlos]

Hi,

I never made this experience with TPFanControl nor heard of it before, but:

if your computer slows down in reaction, go to taskmanager (Ctrl-Alt-Del in XP, right click on taskbar in Vista) go to processes and sort them by cpu usage by clicking on top the column. This way you can find out what kind of process is eating the resources. Wonder what it is

Ciao, troubadix

Hi,

regarding processes, the only process eating resources is winword.exe. if the MSword is running without the endnote there is no problem - cpu usage virtually zero. If they run together - no problem until a insert citations from endnote to the word - then, cpu usage 99 % almost all the time (including cpu temperature constantly around 70°C even with fan under the bios control). Looks like maintaining the connection between those programs needs a lot of resources. question is why other programs take priority over the word (if switched between them) without any problems and TPfancontol not even if set to realtime?

If I could I would get rid of those two programs, but endnote is really useful.

thanks

Lukas

[troubadix]

Hi,

I guess this is not an issue of TPFanControl, it's only linked to the matter that TPFanControl is using WS_EX_TOOLWINDOW :

WS_EX_TOOLWINDOW Creates a tool window, which is a window intended to be used as a floating toolbar. A tool window has a title bar that is shorter than a normal title bar, and the window title is drawn using a smaller font. A tool window does not appear in the task bar or in the window that appears when the user presses ALT+TAB.

I compiled a special version for you with ordinary window to check, just unzip and replace old fancontrol.exe in C:/tpfancontrol by new one.

http://staff-www.uni-marburg.de/~schmit ... ejdlos.zip

I guess it's an issue of endnote, but I know nothing but the name about it.

Ciao, troubadix

[mejdlos]

Hi,

I guess this is not an issue of TPFanControl, it's only linked to the matter that TPFanControl is using WS_EX_TOOLWINDOW :

WS_EX_TOOLWINDOW Creates a tool window, which is a window intended to be used as a floating toolbar. A tool window has a title bar that is shorter than a normal title bar, and the window title is drawn using a smaller font. A tool window does not appear in the task bar or in the window that appears when the user presses ALT+TAB.

I compiled a special version for you with ordinary window to check, just unzip and replace old fancontrol.exe in C:/tpfancontrol by new one.

http://staff-www.uni-marburg.de/~schmit ... ejdlos.zip

I guess it's an issue of endnote, but I know nothing but the name about it.

Ciao, troubadix

Hi,

thank you for your effort but this solution does not work as well. still delays, even if priority to realtime, chrashing while hitting closing the program window etc.

But, i don't want to bother you since I'm the only one who has this problem. I did a little research and there are some (even if just only two) other citation managers probably capable to replace the endnote. I don't know them well, probably will be worse (since they are free) but It pisses me off to have a such a badly written software (and don't talk of its selling price) eating everything it can.

I hope that your effort with process priority and ordinary toolwindow will be usefull for other users in different problems.

thanks again

Lukas

[xiong]

Hi everybody..
finally.. I made it! I modified the FAN power Supply over my T40!

Enjoy

Your work is very interesting. I'd like share some experience with you.

I have an X60, which also has a noisy fan. I ever planed to modify the hardware since the maker would not provide the EC fix (I bet it is NOT so difficult). First, I measured the voltage applied to the fan by oscilloscope. No surprise, I saw a PWM waveform. The average voltage might be the same as what you have measured. I know there should be some feedback to control the rpm, so I guess that simply inserting a resister or diode would not reduce the rpm effectively. Now I am quite surprised you can reduce the rpm that way. I have not done any experiments because I have to use that X60 all the time. I am just wondering...

1. After you insert the diode, what is the real rpm? What about the system reading, is it the same as before?

2. Have you ever tried to eliminate the "pulsing noise"?

Thanks!

[GPR]

I am just wondering...

Hello Xiong...

I was also surprised because of the back reaction of the tachometer. Infact I noticed that the average voltage was increased by the EC (after inserting the diode) but the delta-V was around half volt reduced to the Fan. So i noticed a little benefit.

Your words make me a dubt... I was sure to have measured aroud 0.5V less than before... When I have time, I open it again..
(surely not before 20 days, I'm leaving tomorrow for a work in China)

I measured the voltage using a Digital Multimeter because of my own oscilloscope is still in my ex girl-friend home!!
So no way to see the waveform.. But I can imagine how the EC Controls the fan speed.. like a Dimmer, working over the average voltage.

If you have time, try to test it.

Another solution should be to "pull-Up" the average of the tacometer, using a little and fast ceramic capacitor...
The CE automatically decreases the pulses to reach the desired back reaction voltage.

what do you think about?

Keep Me informed..
Bye

[xiong]

Hi GPR,

I appreciate your reply.

Just wondering if anybody understands the mechanism the maker is using. It is unknown for me and seems not good. I once worked on some temp stabilization circuit. I feel it is not very difficult to get a cooling system with the fan running constantly and as slow as possible. It can also change the rpm continuously according to the temp.

I also purpose a circuit to amplify the Vtach that can cheat EC like you said. But I do not know how a cap can effect Vtach. Did you mean apply a cap to Vin in parallel which acts as a low pass filter?

I bet any additional circuit is unnecessary if the maker could fix their buggy EC firmware.

Enjoy your work and life in China. We can discuss a little more later.

[GPR]

Hi Xiong,
Can you analyze the Vtach wafeform and post an image (simply make a PIC to the display.. or draw it if you know how it is...)?
If possible 2 levels Level1 and level7 or 64.

If the Vin supply is a pulse waveform.. alro the reaction back should be...
I think whit a simple RC circuit, a capacitor and a resistor between Ground and Vtach pins can Low-Pass the waveform and increase the Voltage average.
This should induce the EC to reduce the Vin.

I'm just arrived in Shanghai.. everything is ok..
bye.. 再见.

[Pascal_TTH]

Hi GPR and Troubadix !

Very nice mod GPR. I will read more about and I plan to do it on my R61. This laptop is very cool with fan at speed 1 but it runs at 2900 rpm and I found it a bit too loud.

Troubadix, I'am gonna run the last version of TPfancontrol and give some feedback.

Do you think it's possible to make TPfancontrol with installing the two other programs ? I mean like older TPfancontrol, when you can copy/past the folder without any install ?

I do some test, it seems not to work but I'can find if it's due to tpfancontrol.exe or to missing registry key.

To tell the true, I'am tring to make it run without installing. I know it's perhaps not allowed by the lisense of the io driver... But for my own use...

I can tell you that Thinkpadfan control V0.25svc runs fine with Santa Rosa. I use it since my last posts.

[troubadix]

Hi Pascal,

the TVicport port driver comes as a service and needs the registry key, you will find if installed by setup at:

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\TVicPort (for 32bit systems)
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\TVicPort64 (for 64bit systems)

Just .dll in windows/system and .sys in windows/system32/drivers is not sufficient at all

But once exported it, it will work I think if you import the key to registry and put the files to the proper place

I changed to TVicport for the 64bit support and I will do an update someday with only one setup.exe.

If you like to keep everything under control, you will still work well with earlier versions.

Thx4feedback and for support

Ciao, troubadix

[Pascal_TTH]

Thank you ! I read out all the dev you make for TPfancontrol works with Windows Vista. It's a very nice job as usaual.


I'am gonna make a .reg and a .bat file for an old school install. It will be more easy and faster (in my specific use).


I noticed that TPfancontrol don't read the same temp then Rivatuner for the GPU. I suppose Rivatuner reads the thermal sensor from the Quadro while TPfancontrol reads the Lenovo IC temp.

Running both on the T61, RivaTuner show 51°C and TPfancontrol 43°C. So, for the first time I use the offset option. It works fine to activated the fan but the sensor show the temperature without the offset value. It seems nobody repport this.

[troubadix]

Hi!

So, for the first time I use the offset option. It works fine to activated the fan but the sensor show the temperature without the offset value. It seems nobody repport this.

Yes, of course, you're right but when I introduced the offset possibility I was thinking of guys with T60s and hot ATI gpus which would like to have chilled down their graphic chips a bit They might like it better to see the 'native' high temp in the display. But for your needs...I 'invented' a new parameter ShowBiasedTemps=0/1 . In the icon the 'switch' temp will always be shown, but in main window, now with v0.43beta it's your choice

In total there are three more new parameters:

Code: Select all

BluetoothEDR=0   //set to 1 for bluetooth with enhanced data rate
NoWaitMessage=0  //set to 1 to avoid sleep message
ShowBiasedTemps=0 // set to 1 to show calculated temps for sensors

so download v0.43 just .exe & .ini (TVicPort has to be 'installed' ):

http://staff-www.uni-marburg.de/~schmit ... 43beta.zip

Have a test and fun

Ciao, troubadix

[Pascal_TTH]

Dear troubadix, your are really the best ! Gonna try this ! But later, too much to do now...

[emkaes]

I have lenovo x300 and my fan on level1 have 1900rpm, on level2 5200rpm... What I can do to make on level2 2500-3000rpm ??

[Pascal_TTH]

Hello emkaes,

It seems that Lenovo introduced a new version of the Thinkpad IC.

Have you try higher value ? Perhaps the X300 ony have two speeds. It's not impossible. The CPU in an ULV so his TDP is low. Aslo the X300 uses an IGP with didn't heat a lot.