Possible to control Toshiba laptop fan speed with RW everythin?

What the title of the thread asks. If it matters, specifically a Toshiba Satellite L40-18L. I'm aware there aren't any "normal" programs that can control this computer's fan speed, like Speedfan and others, but since RW is a special type of program, I'm wondering if there happened to be a person that managed to find what numbers you have to change in order to get fan control in the Embedded Controller section.

Also, are there any significant dangers in fiddling with the Embedded Controller values, aside from accidentally finding the correct numbers by random chance and turning off the fan, not realising it and overheating the CPU?

 

1. You have to hack/find out the acpi_fan_control settings in the bios

Get a BIOS Editor and perhaps you will see the string for the fan control option. Extract the BIOS with one of Andyp´s BIOS Tools. Open the string with a Text editor and perhaps you will see the settings for the fan speed etc..

2. You will have to redifine the fan control settings in the bios and flash the modified bios.

I dont know how to do this...

3. Get a Laptop service CD and hope that there is a; test fan; option

I did this with my FSC T3010D because the fan was working ``unnormal´´ and i had got this maintance CD from my friend wich contained a fanH.bat (for high speed) and a fanL.bat (low speed) and a fanO.bat (fan off).
I just copied them on my desktop and used them when necessary .

 

in reply to the first post..


-------------------------------------------------------------------------------------------------------
fanBOOST v1.00 enables manual control of the fan of the cpu-cooler on toshiba satellite L750 + ????

available at: Dev-Host - The Ultimate Free File Hosting / File Sharing Service

commandlineswitches:

none trayicon control
-a1 set auto mode
-a2 set auto BOOST
-b1 set BOOST 1
-b2 set BOOST 2
-b3 set BOOST 3


for the utility to function you need to download and install rw-read & write utility freely available at RWEverything | Read & Write Everything

the sourcecode is included to easily adapt for evt. variation in adressing among different models.


ON THE ISSUE..

on toshiba-laptops the fan is solely controlled by the embedded controller (ec), undefined in the acpi-table and thus not recognised by a wide range of available fan control software. and it's not as easy to obtain insight into the dynamics of the control system, as the ec-firmware ain't readily available for decompiling. hence you have to observe how the ec-registers behave and identify the significant bytes. you can do that as well as change the registers with rw-read & write utility.

so to control the speed of the fan you can address registers in the ec, changing the mode of operation.
on my toshiba satellite L750 the default operation is a four step thermally toggled automatic regulation with rpms ranging from 0 to 4200, the final step kicking in above 70c, the bottom below 40c.
when you are on the final step it is possible to boost the speed to 5100 rpm by setting the value 0x04 in byte 0x55 in the ec-register.
the fan will speed down again below the downward threshold to step 2 at 65c.

you can also get constant boost speed by setting the value 0x07 or choose slightly less noisy overspeeds by 0x0b or 0x09.

an inverted tachoreading can be read at word 0x5c + 0x5d and the celsius temperature at byte 0x58.


of course there may be no simple linear relationship between raises in fanspeed and cooling ability.
in my experience, though, the boost speed enhances the cooling effect enough to keep both the graphics core (which is located at the far end of the heatpipe) and the ordinary processor cores well below the point of intels internal processor thermal throttling at 100 C playing current games.

in a 3x6 min full cpuload throttlestop bench test running at 2200mhz dissipating 29-33w (rising with temperature) the auto cooling goes 93c while fanBOOST level 3 goes 85c for the max core temp.
on top of a laptop cooler auto mode goes 91c while boosted goes 82c.
so fanBOOST lowers the temperature difference to the surroundings (24c) by 8c while the laptop cooler gives a 2c drop.

the laptop cooler basically supplies fresh air for the intakes while fanBOOST increases the flow through the radiator.

during the benchmarks the temperature drop across the processor was 6c. from these temperatures basic cooling theory tells an 11% effective rise in airflow between the auto mode and the fastest boost speed. and a corresponding rise in cooling power.
the rise in real airflow from the centrifugal fan ought to be of the order of the rise in speed of 21%. the difference from the observed is probably mostly due to the temperature drop along the heatpipe between processor and radiator.
in a superficial way along the same line the effective rise in cooling power due to the laptop cooler can be estimated to 3%.
but the laptop cooler is quite a bit more quiet though impractical.


i don't know why toshiba doesn't make the fanboost available to common users. the fans may be worn down quicker of course. what toshiba has done is to implement an extra quarantining throttling system, limiting the multiplier to 18x (of 22x stock, 31x max turbo) and seriously degrading the processor effect far below the tdp of 45W until restart.
this nagging limitation can however be eliminated by throttlestop ( Downloads), leaving it up to the processor internals to regulate the power and multiplier capabilities. it seems like throttlestop is a mere necessity if you wanna run a laptop processor somewhere near the rated performance. it can also be used to limit the cpu-multipliers to leave sufficient power headroom for the integrated gpu. a multiplier in the range of 23x - 26x is good for gaming and other gpu-intensive tasks on my i7-2670qm processor.

 

sorry wrong address.
fanBOOST is now available at https://www.asuswebstorage.com/navigate/s/CABA2106684445629032B5F6CAC793EE4

by the way if you experience any troubles with the program you might have to reset the ec. to do that just shut down, remove the powercord and the laptop battery for a few minutes.

in that case you have to locate the proper register bytes for your specific model.
to do that open the embedded controller tab in rw and look for a byte that goes 01, 02, 03 in correspondance with the speed steps of the auto fan mode. that will be the byte to set the speed mode - which on my machine is 0x55. then change the address in the sourcecode and compile your own version with autoit v3.

:hi2:

 

Hi - thanks for your work

I want to lower fan speed for laptop ASUS U38n - most important while it is in idle (it has an AMD APU integrated graphics, but two fans to high level in idle)
I'm investigating what register address in embedded controller using RW-everything, I can mange to change the speed between different modes - however the setting reverts directly to the original setting within a few seconds (I think the revert is made my ACPI windows driver that continuously controls this ). Why is this not the case for you guys?


Another option I'll guess would be to modify the BIOS, I'd downloaded AMI bios tool and investigated that for this ASUS model U38n, these settings seems to be more deeply hidden than for example the Intel corresponding ASUS model UX31A. I can only find the bios string, but not easy to figure out what to change...

 

Hi snowstorm81

I'm not familiar with asus laptops, and i've concentrated on raising the fanspeed to obtain the full rated performance for for my processor.
You've probably already checked out the following thread: http://forum.notebookreview.com/asus/705656-fan-control-asus-prime-ux31-ux31a-ux32a-ux32vd.html
As far as i've read on asus laptops it is a matter of extracting, decompiling, modifying, recompiling and reinjecting the acpi-table into the registry.

On my toshiba the acpi-table does not contain anything about the fancontroller. I think it's because they wanted to make it harder to manipulate this piece of hardware.
toshiba laptops has a driver called tvalz.sys that is able to read the fanspeed and some powerreadings from the ec, but that is about all it can do.

Via rw it is however possible to lock the present state of the auto fan control also at the bottom level if I write 0x06 to the 0x55 ec-register byte.
And it is possible to set a permanent low speed (2500 rpm) by writing 0x17 or 0x15 (3050rpm). I didn't include it in my program because I didn't want people to fry their chips, even though there is an internal protection system in the processor in case of overheat.
So you better have a good hardwaremonitor like hwinfo ( HWiNFO, HWiNFO32 & HWiNFO64 - Hardware Information and Analysis Tools) if you plan to make it more silent.

If you have located the right register byte it's a matter of trial and error to identify the significant values to write. There's no more than 256 possibilities for that. Have a nice day
Alot of them will not do anything useful, or the ec will revert them again, but as you can see from the ones I discovered there is not necessarily an understandable pattern in what values are good.

I must advice you to only try a modified bios as the very last resort.
It can lockup during flashing, leaving you with a dead motherboard, as most laptop bios chips are not socketed.

But it is funny to overcome the various limitations imposed by the oem's bios and bloatware.
Cause when they advertise with a given processor, we don't like to see that it is for instance thermally limited due to an inadequate cooler.

Good luck.

 

Hello,

Does anyone know if it is possible to lock in a constant fan speed on a Toshiba Satellite Pro U500? I followed the instructions above and used RWEverything to muck around in the EC and identify the significant bytes -- they are 97 for regular speeds and 98 for 'overdrive' very fast speeds. When I manually change the contents of these bytes in RWE the fan speeds up or slows down as expected, but no matter what I set either byte to, the system always changes the setting back to the default for the current temperature within seconds.

My system overheats when playing newer games, so that I have to be on the lookout and quit playing every 20 minutes or so and give the system a couple minutes to cool down, which needless to say is very annoying, not to mention running the ragged edge of 100°C can't be good for the hardware. I tried recompiling fanBOOST with my byte addresses (eg. "WEC16 0X98 0x01" to set byte 98 to 01, i.e. level 1 overspeed), it compiled ok and runs normally, but fails to do anything with the fan speed, not even for the two seconds that manual change in RWE works. I'm not a programmer and this is kinda above my level (btw, what does the byte address format "0x-someting-something" mean? :-/ the bytes in RWE only have 2-digit addresses, 00 to FF... Why always the "0x"?), so maybe I'm doing something wrong. But if I understand it correctly, fanBOOST does not prevent the system from changing the EC setting back -- it needs a setting that will stay locked in permanently to work, so it's essentially just a GUI for what you can do manually in RWE just as well, right?

So, given that I can't program, what can I do to lock in the speed setting I need permanently? (by permanently, I mean keep the system from changing it arbitrarily; I would still need to be able to shift it back down manually, both to spare the hardware and to spare myself the noise)

Thank you!

 

I'm trying to find the fan EC bytes on my brand new Toshiba Z30T-A, but RWeverything half-hangs and eventually shows "FF" on all the bytes.
I'm running on the stock Windows 7 installation delivered on the machine. Did you make any alterations to be able to open the EC window? (in my case, other tabs seem to work better)

Thanks
Mike

 

The same goes for me , the temps climb very fast and in the end I end up getting a system shutdown. I request somebody to help me find the address for the X775 Qosmio fan. Because quitting very often is so frustrating.