I've noted dynamic switching bug also with Power4 Gear utility. Happened never before updating to SP2.
But, as I earlier mentioned, fan issue has never appeared when running on battery.
-
maybe the notebook receives too much power/energy and releases this by letting the fan run? Do your fan problems occur when the battery is charging, or only when its full?
i'm no expert but could that be it? -
i think that this "fan problem" is coused by fast rising temperature of cpu. I have noticed that this occure when I open few aplication at the same time.
The cpu load rise, the temperature too. When we use laptop on battery with win xp, max clock is 800 mhz so the cpu canot warm enough to "fan problem" occure....
thats my opinion off course. -
That's interesting Ganid. Maybe we could then fine tune the frequency threshold with Speedswitch and see where fan problems start to appear. At 2133Mhz they obviously do, now we need to see what happens at lower clock speeds like 1500 and 800Mhz.
-
I'll say it again... Check out my post http://www.notebookreview.com/forums/topic.asp?TOPIC_ID=9894 about CrystalCPUID (page 1 near the bottom).
CrystalCPUID is a much better tool for tunning the CPU frequeny + voltage (lover voltages + lover frequeny = lover temp). []
-
Exacty, i'am using CrystalCPUID with Quickfire's settings, IT IS working for the third day in my laptop.
-
An update about the error/conflict concerning the I/O.. My brother recently bought an Asus laptop, an A4700G and guess what, he has the same conflict concerning the I/O... Maybe all Asus laptops have this problem, so this problem might not be related to the fanissues.
-
Hello,
maybe this thread http://www.michna.com/kb/WxACPIEC.htm is somehow connected to our problem.... -
Whilst configuring CrystalCPUID appears to resolve the fan issues I'm concerned that some long term damage to the CPU etc may occur. I'm not tech savvy and wouldn't know what voltages or refresh rates to use, other than those specified in Quickfires message.
What is interesting is that this program actually works, has anyone else had any experience with the above program? -
"Hello,
maybe this thread http://www.michna.com/kb/WxACPIEC.htm is somehow connected to our problem...."
I tried that, but my A2D didn't boot with tuned acpiec.sys file. Underclockin with CrystalCPUID software seems to be only relief for this issue so far... -
Have been watching the forum since entering a post on page 4. I have just had my a2d in for repair under warranty for the same issues of fan alway's on - heating etc. They changed the motherboard. It's come back with a similiar problem still. The fan is always on when the power adapter is in. As soon as i boot up, the temp is over 54 degrees (fan operating temp)question: How could a cpu be at 54 degrees when i have just booted it up? If I pull the power adapter out it drops 10 degrees and the fan stops because it is now registering around 41degrees. I don't think this is a hardware problem. And i don't think the temp is right either. i think for some reason the pc 'thinks' it's that temp and is acting accordingly and that the power supply - voltage - has something to do with it.
-
Yes, AC power supply has definitely something to do with it.
I repeat (again) that this fan problem has never appeared when I'm running on battery. AC power supply is generating heat to battery as it is recharging and that heat affects to mainboard also.
My notebook visited also service point and they changed my mainboard too. Without any improvement. I just noticed few weeks ago that I have now older BIOS version, 0205. I thought that I had the latest one, as I had when I got this piece of ..ehhmmm.. equipment. I'm too lazy to update my BIOS, because it wouldn't help anything.
BTW, it is very interesting to see that Asus M5N owners are having very same issues with their notebooks. ASUS has two notebook model with serious heat / fan issues. I doubt that the rest of Asus notebooks have more or less these problems too ... -
<blockquote id='quote'> quote:<hr height='1' noshade id='quote'>Originally posted by pasta
Last edited by a moderator: May 8, 2015 -
I personnally think the problem lies in how the hardware temp monitored is reported to the softwares. It seems quite strange that the temperature of the CPU drops from 50 to 40 in less than 1 second, only because you unplug the AC...
Also, I'm currently running the IBM computing grid, which is a distributed project, blah blah, anyway, when I'm running this, the temperature goes above 70 (around 78-80) in less than 10 seconds. If I close the program, the temperature goes back to 50-55 in less than 10 seconds too (I'm not using powergear, just using the dynamic cpu control that comes with windows).
And last but not least, something I noticed 2 days ago, but I wanted to be sure before posting it... I had the fan stuck problem earlier today, and so the temperature was around 40°C with the fan on (theorically, it should goes off at 50, and on at 55). No matter what, the fan would not stop. I launched the ibm computing grid, to make the temperature rise above 50, and I closed the program as soon as the temperature was around 70. The temperature then dropped down to 50°C, and magically, the fan stopped...
That led me to think (a quite simple reasoning, but sometimes stupid reasonings are what works best) that the fan stops when the temperature is equal to 50°C. And so, if the temperature goes from 51 to 49, the fan won't notice it should stop, and will run forever...
Anyway, this trick does the thing for me, try it and report? -
Not strange at all. We're talking about a reduction from 2133MHz @ 1.65V to 400MHz @ 1.3V. I just tried it. Plugged in, fan running: 55C -> Unplug -> fan stops -> wait a few seconds for temp reading to update -> 46C.
-
I just tested my notebook with Prime95 benchmark and torture tests. I had the fan jam almost immediately when I booted my computer. Then I tried those tests, and tempereture rose from ~31 C to ~60C. When short benchmark test ended, temp was about 56 C, then it dropped to about 46 C in two seconds. Fan didn't stop at all during temperature changes.
-
"That led me to think (a quite simple reasoning, but sometimes stupid reasonings are what works best) that the fan stops when the temperature is equal to 50°C. And so, if the temperature goes from 51 to 49, the fan won't notice it should stop, and will run forever..."
What i meant was if the temperature *jumps* from 51 to 49 without stopping at 50°C... Maybe the sensor only report the temp at discret time intervals? -
Considering for one moment this is a transition problem somebody would have noticed the fan would turn off sometimes (another transition happens and then the fan eventually turns off).
As I said personally I don't think fan control is based on transitions.
-
That doesn't change anything. Provided the temperature eventually ends up below the desired value (eg. the temperature at which the fan should stop) well... the fan should stop.
Making it a fan control issue way before being a temperature reading issue in all cases. What I mean is that you don't program the fan when a transition happens, you control it depending on readings at a given time - otherwise you're a bloody stupid programmer.
-
Here's an article about how ACPI fan control works:
http://www.analog.com/library/analogDialogue/archives/34-04/fan/
I'd might try to debug this issue (allthough I'm not that good, and I probably won't) if anyone comes up with a definite way of reproducing this behaviour. I haven't noticed it myself for a while, and quite frankly it never bothered me that much as simply putting the notebook to sleep and bringing it back again takes less than 10 seconds an resets the fan.
One would most likely need access to technical information on the SMBus IC as well as do some logging to figure out what triggers what. After that there are a lot of available source code to draw from to build a fan control program, if possible. Or I guess someone like the SpeedFan author might be nudged into implementing a workaround if the mechanisms controlling the fan were to submitted to him.
Trouble is someone would have to:
1. Be able to consistently reproduce the bug.
2. Have access to the information allowing them to gather data and interpret it.
3. Invest the time neccecary to fix it.
Somehow I doubt that's going to happen. I'm guessing Asus isn't going to spend time on a deprecated product. (At least not by going through the regular support channels. If someone was able to contact one of the responsible engineers directly one might get something.) And while some dedicated user(s)/programmer(s) might provide 2 and 3, that won't do much good without 1 beeing resolved first. -
Don't mess up too much with your notebook. If you do want to open it, write down the full reference of all chipsets being used, at least next to the processor and the fan.
The SMBus seems to not report the presence of a fan control chipset (however it does report the presence of the Maxim temperature sensor), meaning the Asus ACPI driver goes through a proprietary mechanism.Last edited by a moderator: Jan 29, 2015 -
It seems to be quite impossible to find bullet proof way to reproduce this fan issue. Usually I have it when I have several programs running (Word, Winamp, Outlook and web browser are usually running). Sometimes fan stucks without any load on computer. And sometimes it works well with heavy tasks running.
As long as we can't find the way to reproduce fan problem, ASUS says that they don't find anything wrong on their tests (as they told to someone (to escape perharps?))
I'm going solve this thing with quite suspicious solution. I'm going to open my A2D and take a look at that fan and it's wiring. Maybe I put potentiometer on that voltage wire so I can control fan voltage. I took a glimpse and tried to open my notebook few weeks ago, but I couldn't get it fully uncovered and I had no time to figure out how to remove all the covers. There was some sticker which said something about warranty...
Yes, I know that I can kiss goodbye my warranty, but hey, if it breaks I have a good excuse to buy that iBook... []
And never any ASUS products for me, thank you. -
Looks like the M5N guys in this forum got Speedfan working with their notebook fan. See the upper thread about the M5N.
-
<blockquote id='quote'> quote:<hr height='1' noshade id='quote'>Originally posted by escape707
Last edited by a moderator: May 8, 2015 -
Hi All, I have been following all the conversations here about A2D fan.
I have a relatively older Asus L3C p4M-2.2M, I have this problem,too.
The fan never goes off!
Even under 50degrees, it becomes very quite but as soon as you start a program hooop, it kicks.
Fedora Core 3 works, it lowers the temp very good to 48.
I'm wondering if there is a way to stop fan before 50!
Is there anyone else having this problem too?
Racoon City will be completely sanitizated. -
Darn! Somebody from an Asus Repair Centre in Australia (AssyrianKing) has apparently popped in this forum, commenting on the A2D sound card issue:
http://www.notebookreview.com/forums/topic.asp?TOPIC_ID=14370
I am sure we would ALL love to know if he got notebooks back because of fan issues.
-
I havent had any laptops come in with the fan issue you are all having in here....its strange, have you all tried bios updates and pc probe update to the latest?
http://www.asus.com/support/download/item.aspx?ModelName=Tools
Asus Repair Centre Technician -
Hi AssyrianKing,
Yes we have tried different BIOS versions, tried removing Power4Gear and using SpeedswitchXP instead, made sure the latest CPU hotfixes and SP2 were installed. Reinstalled the system several times etc. There is a BIOS issue.
The problem concerns notebooks here in New Zealand but also from Australia, Finland, Turkey, France, The Netherlands. Pretty much from everywhere around the world.
Hyde (Finland) had his notebook motherboard replaced and experienced the same problem a couple of hours after using his computer.
When you use this notebook this problem has to happen sooner or later.
Would you mind investigating it a little bit for us?
Make sure you are using the laptop in AC mode. Surfing the internet for some time should be sufficient to show off the problem.
-
Ok ill have a look and see if i can see this problem and see what i can do. So im just looking for an A2D's cpu fan starting up once it gets to a certain temp then it will never turn off? just sit there working forever? and run the AC power.
Asus Repair Centre Technician -
You have to use the notebook normally. It is more difficult to reproduce this problem if not effectively using the notebook on an everyday basis.
If using Power4Gear, use the Game power management mode. If using SpeedswitchXP, use dynamic switching mode. Don't use modes where performance would be 100%, the fan would stay on all the time to cool down the notebook (which in this case would be perfectly normal). Of course don't play any game, because CPU activity would be 100% as well.
The problem shall happen, just be a little bit patient if it doesn't show up after one or two hours.Last edited by a moderator: May 12, 2015 -
...and let me say that A2D fan issue doesn't happen with Linux OS, if you have powernow daemon and other AMD-related stuff running with kernel.
Although, with linux asus_acpi module reports errors. This seems to happen with latest linux distributions (now I'm using Ubuntu). I read that some asus models have somehow broken bios or something like that... -
There are two new BOIS-files on the german ASUS ftp site, 209A and 210A. No information on what's changed, though.
ftp.asuscom.de/pub/NOTEBOOK/A2x00D-DC/Bios/ -
Has anyone tested the new bios yet? Is there a changelog somewhere?
-
Hi all,
<blockquote id='quote'> quote:<hr height='1' noshade id='quote'>Originally posted by bugmenot
ftp.asuscom.de/pub/NOTEBOOK/A2x00D-DC/Bios/
<hr height='1' noshade id='quote'></font id='quote'></blockquote id='quote'>
Thanks for the info. I was frequently checking ftp.asus.com.tw for updates, however updates did come from the german site
Just check the site and downloaded bioses. BTW, German Asus ftp site includes some new drivers for vga (V8.01RC3-040421a-014736E.zip), smartio, and modem.
I'm thinking of updating the BIOS. So far, primary problem was the FAN issue... I hope they have resolved it.
Currently using v0208a which came installed with the notebook.Last edited by a moderator: May 8, 2015 -
Hi again,
I have updated A2D bios from 208 to 210 successifully using winflash 2.11. After a reboot, I have reconfigured my bios settings (no new options picked my eye).
On Xp, I have opened the device manager if anything new listed. Guess what "AMD Special tools Driver" is listed under system devices. Which I couldn't recall from previous setting. Now, There is a total of 27 items listed under system devices. Please check with your setting.
So far I'm testing fan issue by running powergear in game mode (which I did not ever uninstalled), so far everthing works fine (fan stopping at 50 C and reactivating 55 C degrees), and also resource confilict report under "Microsoft ACPI-Compliant Embedded Controller" still exists (maybe it is not the main issue), though.
When I'm done with powergear, I'll switch back to crystalcpu which runs even cooler with voltage adjustment 800Mhz at 0.925 volts.
Did Asus resolved the fan issue at last? We'll wait and see... []
Cheers. -
> "AMD Special tools Driver" is listed under system devices. Which
> I couldn't recall from previous setting.
Not BIOS related. Comes from AMD PowerNow! Dashbord or other software from AMD.
> and also resource confilict report under "Microsoft ACPI-Compliant
> Embedded Controller" still exists (maybe it is not the main issue),
> though.
It's not related to the fan. I have no resource conflicts.
> When I'm done with powergear, I'll switch back to crystalcpu which
> runs even cooler with voltage adjustment 800Mhz at 0.925 volts.
Tried RMClock (cpu.rightmark.org/products/rmclock.shtml)? I switched to regain dynamic speed adjustment while still undervolting. Running 0.925-1.45V, 3X-16X (and a sliding scale on any multiplyer inbetween). -
I really hope the new BIOSs fix the fan problem. Fortunately I have not sent the laptop to technical support yet for them to fix the sound card (which is defective). I'm going to test the new BIOS first otherwise I will talk about the fan problem too.
BTW don't forget that to test the new BIOS you should not attempt to reclock the CPU with any software. I would forget RMClock and CrystalCPU for a while until the situation really looks perfectly normal for a while with the new BIOS.
-
now I'am using bios a210. it's total crap......
worst than a208. fan stays on almost every time.... -
The purpose here in the first place is not to see whether the fan stays on too often but rather if it stays on permanently. There is a big difference between the two incase you haven't noticed after 9 months discussing in this forum the fan issue!
Last edited by a moderator: Jan 29, 2015 -
Hello,
I'am fallowing this thread almost from the start and I'am aware of its nature.
What are you trying to tell me? didn't I wrote the the fan stays almost every time on - I mean that almost after every reboot the fan constantly stays on 3000 rpm.... -
Alright. Is your diagnostic exactly the same as the one you had with previous BIOSes I.e.:
- notebook seems to work fine. Fan on at 55C, off at 50C.
- sometimes however, the fan gets stuck, runs forever. Meaning your notebook temperature can go well below 50C but with the fan still on.
- only cure for fan-always-on issue is standby, reboot or fiddling with Speedfan MAX6657 advanced settings.
Please elaborate.
-
Hello,
My notebook seems to work fine at the beginning with a210 BIOS. Fan starts at 55C, off at 50C. On Previous BIOS - a0208 (I bought a2d with it) it was the same, however now it stucks on almost every time with Windows Power Management, Power Gear and even with Crystal CPU. Meaning my notebook temperature go on 33 deg. but with the fan still on.
Yesterday I installed RightMark CPU (min 3x max 14x) clock utility and as for now this fann problem is gone with bios a210.....
Unfortunately it was the same at the beggining with a208 and Crystal CPU -
hi I have a M6ne and my fan kicks in at start up for 5 sek, stops and then start as windows starts here it starts at 4000 rmp runs for some time and settels at 3600 rmp. the cpu temp is then 75-80 C is this normal? if I pull the plug and set power4 at e-mail-office the temp at once drops to 58 C and the fan runs 3000 rmp if i then reconnect
the power plug the temp goes up to 75-80 again!!!!
Very odd.
PM 1.8, 512 mb ram, 80 gb hdd, 15.4" -
Hi.
This helpe on my pc!! Bat time emproved and the fan stops running once in a wile. Bios version i 0208.
I hope that it is okay to copy the texst!!
Hi all
A bloke nicknamed Irion (from Germany I think) has posted this on the VIP Asus forum and I thought it could do a lot of good on this one, so I hope he doesn't mind but here it is
(A Small modification: this all works for sonoma as well as I found out now. I have updated the tutorial.)
Hi Folks!
I want to share something great with you
all of the following tricks are from the german unofficial ASUS M6N-board http://www.encke.net/encke/modules.php?op=modload&name=phpBB_14&file=index. I will give you a step-by-step instuction. This all works perfectly on my ASUS M6710NEUP, but basically should work on any centrino-based computer (banias and dothan, and as I found out now for Sonoma as well!). For example it works well on a Acer Aspire 2020.
First, what it is all about: We will use the tool "RM Clock" to reduce the core voltage of the CPU. This dramatically reduces power consuption, therefore the battery lasts longer and the cpu temperature is much lower. There is absolutely no risk of damaging the cpu, if you follow my instructions. As well, the CPU performance is not modifyed, you still have the same power for games. On my machine, i can play Half-Life 2 or World of Warcraft with the CPU staying so cool, that the fan keeps off half of the time (BIOS 0205).
Here is what these tools can achieve:
Without these tricks, my computer uses around 33 Watts in prime torture test (see below) at 1.5GHz, and the temperature rises over 70°C.
With RM Clock, it only needs 22 Watts, the temperature doesn't go over 56°C but it still has the same performance!
The Watt numbers were measured via Mobile Meter 0.3.0.0 in Battery mode.
I can not guarantee that this will work, and i don't take any responsabilities, i can say this perfectly works for me, and for all my friends on whose computers i tried it. I am also not sure if you lose your warranty or something. basically this should not happen, because we only decrease the cpu voltage, and it is only done if the program is active in windows. if you reboot or just deactivate the program discribed below, all changes are gone, there is nothing saved in the hardware.
What we need:
The tool "RM Clock" from http://cpu.rightmark.org/download/rmclock_15_bin.exe
The tool "CPUMSR" from http://www.cpuheat.wz.cz/BetaVersions/CPUMSR_0_88.zip
The tool "prime" from http://mersenne.org/gimps/p95v238.exe for stress testing
Centrino Hardware Control from http://www.pbus-167.com/chc.htm
1 - Uninstall the Power4Gear Software, it sux ;-)
2 - Download the above tools and extract/install them. it is not important, where prime and CPUMSR are extracted, RM Clock should be extracted to some folder, where it can stay, like Crogram FilesRM Clock3 - Now, start CPUMSR. Click "Configuration", and uncheck the "enable user protection" box. now, click the "Pentium M" button in the list on the left side (be sure not to click any of the other buttons!!)
4 - Choose "Frequency & Voltage Control"
5 - Here you can see the current multiplier and core voltage, write them down somewhere, or keep them in mind.
6 - Run prime (should be in the start menu under "Prime 95") and click "Just Stress Testing"
7 - In prime, choose Options -> Torture Test... and choose the first box (Small FFTs).
8 - click "OK" to start the "torturing" of the CPU, you will note text lines beginning with "Test 1, 1000 Lucas-Lemer iterations...", they keep appearing (=torturing) until you choose "Test -> stop"
9 - switch back to CPUMSR and move the windows so that you can see the text lines appearing while you change the settings in CPUMSR.
10 - in the "Change Multiplier to"-field you choose The maximum available multiplier for your system. In order do find it out, open the CHC-Helpflile from the START-menu, it has some great tables, including each CPU's multipliers and the correspondent default Core Voltages. it has no effect if you choose a multiplier above your maximum, you can't overclock your CPU.
11 - in the "at a Voltage of:"-field you choose the correspondant Voltage from the CHC-Helpfile-table, but 1.308V should work for everybody.
12 - click "Set"
13 - now, if your computer hasn't crashed, which should not yet happen ;-), you slowly decrease the voltage, but keep your maximum multiplier, and click "Set" each time.
14 - while you do this, keep watching the lines in prime, as long as they appear regularly, everything is ok
15 - At a certain voltage, one of these things will happen: either prime reports some kind of error, or your computer crashes. This is NOT dangerous, just restart (press the power button 5 secs, wait 5 secs and press it again to turn the computer on). It is important, that you keep in mind the last voltage that reported no error! it varies from system to system, on my system it is 0.924V but this will definately vary on each system!
16 - write down this voltage that was the lowest one to work and the correspondend multiplier on some piece of paper!
17 - now, you repeat the same (start prime and CPUMSR, if it is not yet running) and set the multiplier to 6x, which is the lowest possible setting, with the lowest possible Voltage, 0.700V and check out if prime reports an error (really should not happen).
From now on, we don't need CPUMSR any more. you can uninstall/remove them if you want, we only needed it to get the limit-voltages. You can now decide, if you want to use CHC or RM Clock. Personnaly, I prefer RM Clock, so I am explainig it in detail. CHC works very similar, you will find out! CHC offers the possiblity to set the temperatures at which the fan starts and stops, this might be interesting to V6V-Users. I strongly recommend M6Ne-Users not to activate fan control in CHC, it makes the fan never stop again ;-).
1 - Now, start RM Clock by double-clicking the .exe-file.
2 - Click the "Management"-Tab and activate the "Use P-State Transitions (PST)" checkbox
3 - At "Minimal FID" you choose 6x
4 - At "Minimal VID" you choose 0.700 V
5 - At "Maximal FID" you choose your maximum multiplier, 15x in my case
6 - At "Maximal VID" you choose the voltage you wrote down, that was the lowest one to work, it is recommendet to choose 2 or 3 levels above, as a safety backup for voltage fluctuations.
7 - At "AC Profile", choose "Automatic Management"
8 - At "Battery Profile", choose what you want, "Minimal" for 600MHz and super-low power consumption ;-), or "Automatic Management" for high performance, but still lower power consumption than 600MHz without RM Clock.
9 - Click "Apply" and close the RM-Clock window.
10 - Now start prime again and do the "Small FFTs"-Torture Test again to check if everything works alright.
You will note the small icon in the system-tray-bar, where you see a circle, that turns red if the cpu is at a high frequency and stays gray, if the frequency is low. The frequency adapts automatically to the cpu usage, as long as you have "Automatic Management" activated under "PST Performance Profile" in the right-click-menu of the system-tray-icon!
11 - if everything works alright without a crash or an error, right-click on the icon in the tray-bar
12 - and activate the "Start Minimized to Tray" and "Run Automatically at Startup" options.
congratulations, you're done
if you have problems, feel free to post them here
Greetings, Luiz.
Thanks Luiz []
-
Thanks for the information, you talk about something for Centrino notebooks but this should work for A2D notebooks with AMD Athlon CPUs, although I am not sure for the P-state transitions, which are I believe unsupported on the XP-M Barton Model A of the A2D...
Still, this is hacking the voltage and this is risky in terms of stability.
To get back to the A2D fan issue: if you uninstall the ATK0100 ACPI driver, the fan issue disappears. CPU heat monitoring and active cooling is still fully handled by the BIOS and the MAX6657 does its job - there is no worry. Speedfan reveals in this case the threshold is 55 for the remote temperature (which is probably MAX6657's default value), whereas ATK0100 seems to set this to 70. It is definitely not a long term solution, but it is worth noting...
For the general heat produced by the A2D, it is interesting to understand WHY the A2D produces so much heat. Let me explain.
Intel Centrino CPUs enter a low power mode during OS idle time, a mechanism which goes far beyond a simple decrease in CPU frequency. Athlon CPUs can do this as well, provided two conditions are met. First, the STPGNT (Stop Grant, or Bus Disconnect) state has to be enabled on the northbridge (in our case a SiS 746FX northbridge with a SiS 963 south bridge chipset). Second, the OS idle time core loop needs to read continuously the PLVL_2 register on the southbridge, not just HLT the processor. Intel processors automatically disconnect their front side bus when the HLT instruction is issued, this is not the case for the Athlons.
The problem with our SiS 746FX is the second thing to do. It is not done by the OS because SiS does not publish a way to read the PLVL_2 register... Plenty of northbridges can do it, AMD 76x, nearly all VIAs... etc.
As a result of our ACPI driver's inability to disconnect the bus, the "C2" state is reported to be not supported by Windows. This is why CPU Idle Pro complains, if you ever tried this program. CPU Idle Pro indeed does not known either how to read the PLVL_2 register. -
Just my 2 cents:
On my M6Ne with latest Bios 0208Ae (I hope I'm not off topic) the fan starts at 55°C and stops at 49°C - Well, more exactly it switches from 3000rpm (very quiet) to 3600rpm (somewhat noisy), and back.
However sometimes the fan "starts" between 49° and 55°C, and in this case it never returns to its quiet state, even when the temp drops below 49°C. I have to restart the notebook to stop the fan.
I found a workaround, though: I observed that if I stress the CPU to make the temp rise above 55°C, then stop and let the temp drop below 50°C, then the fan "stops". It's seems that the fan control code gets confused because the fan speed and the temp range don't match, and by making the temp rise to the "normal", the fan code gets sort of resynced.
I even wrote a 3 line Python script to do the job (stress the CPU for 30sec then stops. Dumb, but worksCould be smarter if I knew how to read the temp) and added an icon on my desktop to keep ready !
Maybe something similar might work on other Asus notebooks...
BTW, I tried to remove the ACPI driver (escape707's suggestion). Got worse so I restored it (with some pain).
HTH
Richard
In case you don't already know the tool, have a look at Centrino Hardware Control, it's free, far better than Asus Probe, does monitoring (temp, voltage, clock, battery..), voltage tweaking, fan control, etc... -
Hello,
Ok, I have bought my a2d (amd2800 ati 9600) second hand as a reall bargain and had that **** fan issues (overheat/jam/low cpu temp, hot case temp) Though I dislike AMD (Nothing about the performance, I simply dont like the name of it, intel inside sounds cool) I love the performance of this babe (especially the 9600) and didnt want to change it.
I am from a country where all guarantees are solely on papers and does not apply in real life, service personel is always chosen from the ones paid the lowest fee and as a result having the tiniest brains blah blah blah... not excluding asus.
So I went out for a solution by myself. Read all threads in this site and many other sites including asus technicians posts without finding a solution.
Now what I have acieved so far...
1. I have totally dismantled the notebook to the tiniest bit, hacked ino the fan. I dismantled it too , it was a ac center rotor surrounding stator engine so It was easy to set it apart.
2. I cleaned it with various gentle chemicals (no alcohol glass cleaners first, then mild soap (ph7), rinsed few times with water then used several times degreaser contact cleaner (not the oily one) . I looked up for a good degreaser but as the prices were sky high, so went on with a cheap chinese one which worked like a miracle (Before I bought this one I was using Philips now I will never buy Philips they sell canned water instead of greaser when compared to this chinese crap)
3. Finally I rinsed every bit of the fan with pure water then
washed with pure water and immediately dried it with a hairdryer. I applied this procedure from step 1 to 3 to all of the fan casing.
4. I used some silicone grease on the stator and assembled the fan. (This grease is mostly used for harsh conditions theoretically it never dries or bonds , I knew it from my scuba gear we depend our lifes on that lubricant but thats another story)
5. now the coppersinks : There are two copper heatsinks. One with L heatpipe for cpu and one under the top cover for the graphics card attached to the cover of it) When I opened them they were nearly clogged with dust. I took them directly to the bath washed then applied the same procedure in 1-3.
7. The original thermal paste crap (was gray color on mine) was dried out on both sinks and was hard as a mules tooth. I worked about half an hour to clean it out without scratching the contact point. For the first 25 minutes I used every chemical I had (mild acis, detergents alcohol, acetone etc etc) but could not remove it totally. Then I used that infamous chinese degreaser on that one and worked out like a miracle again. When finished both heatsinks looked like brand new and fresh out of the factory.
8. This time I didnt went on cheap and bought the white thermaltake heat paste crap in a case which looks and probably was produced hipodemic injector needle (Those eastern people have a great imagiation for using one thing for another purpose) I applied it by a plastic card to make a smooth surface. voila ! it's finished.
9. I reassambled the notebook. (How did I do that ?? of course with reverse of the method I used to disassemble it)
Now the results :
1. Fan is much much silent than the original, both in decibel and the turning on intervals are much less. Working wih it for a month without a problem.
2. cooling problem is solved much cooler it is ! Noise problem is partially solved, it is much more silent and turns on much less (dont have a tool to mesaure decibel but roughly half the noise) but still not as good as I would liked.
3. About jamming problem : I tried half diassambled run tests to find out if the heat receptor(thermocoupler) has the problem , or it is a general heating problem, or a software one. Voila ! the cpu heat and heatsink is approximately as the monitor programs say, and as the posts in this forum. But.. the pretty ati 9600 gets extremely hot. With using speedstep and running on 400 mhz I have found out that %95 the stacking of the fan is not a real stuck but caused by the over heating of the graphics card main processor.
4 . why the graphics card gets hot? 2 reasons,
a) the heat pipe which carries the heat to the heatsink is wrongly designed in the factory. It is round at the edges, half of it sits on the graphic card's main processor following with a gap (about 2 mm) which sits right on the graphics processor. The thermal paste helps a bit but still it is a bad thermal contact in this way.
b) the graphics card heatsink is too small for cooling . I didnt calculate or mesaure anything but it gets hell of a heat when even notebook is running idle.
5. What can be done from now on : The heatsink of the graphics card must be enlarged for a total solution. 3 step idea for this :
a) A direct heat pipe between cpu and graphics card heatsinks which can be dissasambled easily must be constructed so that excess heat will pass onto cool and much larger cpu heatsink which eventually equalise both of their heats, eventually lessen the fan turn on times and probably put an end to the fake fan stuck problem.
b) Luckily there is a great hollow in this notebook hust behind the cpu heatsinks head which sits on cpu. Probably it was for the pentium version of the notebook as the cases are the same. You can not miss it. directly under the laptop wih a square grid. It is a 2,5 cm heigt clynderic hollow with a radius more than the diagonal of the square metal grid. With a radial copper heatsink which can fit there and a heatpipe the heat consumption for the present heatsinks can drastically be increased. Even a 5v low rpm low noise cental fan (like the ones used on graphics cards) can be attached there which will take power directly from an usb port internally. With such a modification the main fan most probably will be fully functional and will also never need to work again except emergency issues (overclok etc)
c) for the main fan I did the best treatment and still it is a bit noisy. For a solution it can be changed with a same diameter same voltage low noise same rpm (sky high pricy) fan. For improving the aerodynamics of the casing som modifications can be made (inside of te fan casing is quite rough and this pretty much can cause the fan noise)
Finally, Jamming problem is still there but much much less it happens. and now I know why it happens and have an idea which I will apply and inform you if I succeed
Warnings !
1. This worked for me, might not wor for you .. no guarantees !
2. Bewarned thee !! If u dont know how to open a notebook dont do this at home ! Besides cooking your notebook, you can seriously hurt yourself with the chemicals or tools.
note : sorry for the sloppy english, It is not a native language for me, I learned from holywood movies
Thats it, have a nice and a silent day !
Bluemax -
Wow, you certainly went through a lot of trouble just to make your notebook cooler and quieter!
I'm glad it worked for you though... -
I had read about for a week every thread for this issue and couldnt stop myself posting this soluton though it's a bit long -
Hi,
I found this solution on the web regarding a M6Va ( http://forum.rightmark.org/topic.cgi?id=6:336), could it work for the A2D fan problem?
You will need Centrino Hardware Control( http://www.pbus-167.com/chc.htm#anchor_download), SpeedFan 4.27beta ( http://www.almico.com/speedfan.php) and eventuraly Everest ( http://www.lavalys.com/).
1)Centrino Hardware Control is almost the same as RM CLock, but in a way simplyer. Just install it (and deactivate any other power saving software just like RM Clock or Asus PowerGear). All you have to do here is enable dynamic switching and undervolt CPU, which is very efective (you can undervolt in when on 800Mhz vrom original 0.9V to 0.7V and its still stable), if you like to undervolt it in other modes, you can do so, but you have to run some test to find out if your setting is stable enough (for example I have my CPU undervolted even when on 2Ghz, and its stable on 1.180V). And thats all on CHC... to explayn it in every detail would be too long, so I count on your abily to work with such a tools ...
2)RUn Everest, go to Computer > Sensors here you can see 4 thermal sensors, Main board, CPU, AUX and HDD ... AUX is just some secret name of you video card, after you run Everest (keep it runing), run SpeedFan, it should find 5 sensors, CPU, Local, Remote and Temp1. CPU and Remote is important for us. Remote = AUX= GPU. But remember to first run Everest then SpeedFan, if you would run only speedfan, it wouldnt find the sensors. I had a small email converstaion with creator of speedfan and he helped me out a bid and repaired speedfan code so that it can find the sensors, the patched version of speedfan.exe is here, just overwrite the old file eith this one. In case it wont find the sensors, shut it down and run it again, sometimes you have to do so.
3)Setting SpeedFan ... run SpeedFna ond go press configure then Advanced amd select chip ADT746 there you can see some usefull settins ... Now, change the PWM1 and PWM2 mode to Manually Controled. Now you can manually change FAN speeds in the speedfan main window. You can also set the FAN2 div from 1 to 3. Now go to Temperature tab, uncheck Local and uped all of the checked ones. Now uncheck subsettins (speed1 and speed2) on every checkers, the only two of them to be chcecked are Speed1 under CPU and Speed2 under Remote 2. Now click on CPU and set the desired and warning temperatures (my setting is 70 for desired and 95 for warning, which looks too high, but you will se why in next few moments). Do the same thing with Remote2 (my setting for this one is 85 desired and 97 warning). Now go to tab Fans and uncheck Fan 3 and Fan4. Now go to Speeds tab and uncheck speed3. For Speed1 set nimimum value 35% and for speed2 set minimum value 30%. Dont forget to check "Automaticalyy varited" on both Speed1 and Speed2. The maximum values depends only on how good is your hearing ... I mean if you are willing to listen do 4000rpm, then set 100% for both, but my setting is 65% for speed1 and 55% for speed2.
4)The final touch ... close configuration and go to the main speedfan window, if your fnas are now going on 100% and are absolutely loud, dont worry! The final touch is checking "Automatic fan speed" option in main window above te temperature readings. Fan speeds of your GPU and CPU fan should decrease after 2 or 3 secounds.
Explenation: Now we have that cool and quiet notebook that we all wanted to have. What have changed? Now after you start windows you can hear the fans running really fast, but after the Speedfan pops up it will decrease. Fans are now running on about 1000rpm even on idle, 1000rpm cant be heard by human ears, when you start some game or any other aplication that needs full power CHC will clock CPU back to 2Ghz and you can put up ATI x700 on full speed through ATI powerplay in ati control panel or control center, there will be more heat and speedfan running on background will detect it and it will start fans depending on your desired temperature and on your maximum fan speeds. BUT thanks to those 1000rpm ond idle, it will take more time to produce such a heat. And to finish things up, after you stop your game or whatever, the temperature will decrease really fast, because CPU will be underclocked and undervolted and as a result, speedfan will put back those unhearable 1000rpm. And now I am getting to those high desired temperatures of mine. It is so high because those are operational temperatures of CPU and GPU, its the edge where you need active cooling, and why I have set it this way? Because when the temperature dropsdown from these values it means it wont be going up, so those 1000rpms are enough to cool it down when you are just browsing or writing text after you stopped playing game or whatever. For example my idle temperatures are 40 to 45C onc CPU and 55 to 60 on GPU. Ofcourse, after you stop playing the fans will be running go few secound loud, but just for a few secounds and when tehy get to low rpm, both CPU a GPU will be hotter than that temperatures od mine, but wait few minutes and you will see same temperatures....
cheers,
ANDNow
A2D cooling fan non-stop running
Discussion in 'Asus' started by Hyde, Sep 29, 2004.