Fan control in the last BIOS for a Clevo d470w "clone"

Hi everybody!
Yes, it seems a Clevo D470W clone. I'm referring to a Hyundai P57V imagequest that has something less than the Clevo model I mentioned (such as a 17" display, a subwoofer and 2 additional fans) but has the same motherboard as the Clevo's one, although Clevo came out with a motherboard code (71-M4000-D04) slightly different from the Hyundai P57V's one (71-M4000-D03). I compared the schematic diagrams of Clevo D470W with the PCB of my Hyundai: there was absolutely no difference. That's why I thought that this section was the most suitable for this thread (even because there is no hyundai section on this website, no longer service assistance and a company to contact). They have an identical chipset.

After these necessary remarks, let me tell you, first of all, my current BIOS revision, 1.05, and kbc/ec revision, 1.00.04. My bios doesn't have any option to control the fans but someone said that in the later BIOS revisions there could be. Is it true?
Secondly, having already the 1.00.04 firmware, would I need to flash it again or could I go straight to flash 1.09 BIOS revision (downloaded from this link: http://repo.palkeo.com/clevo-mirror/D4x0E/ )?
Thirdly, for repairing needs I disassembled my laptop and I disconnected the CMOS battery. All the peripherals are now correctly listed in the POST screen except for the ATAPI CD-ROM and I noticed the line "system BIOS shadowed". But then my system cannot start up because the system file isn't found. Could this have to do with that fact? Or perhaps is it a driver problem (that I should solve by using the drivers cd provided by the manufacturer)? In these cases, is it recommended to flash again the BIOS? Or do I need to format my hard disk (because system file wasn't found)?
Lastly, once unzipped the BIOS files into a formatted floppy disk, I just have to boot the system from the floppy and then type something in DOS to load the BIOS. What do I need to type?

Any help is really appreciated.

 

Yes, according to the change-list:

The 1.09 only is fine.

That message means the bios has been copied to memory (ram), this is normal. You do need to re-insert the battery, otherwise it can't configure (or remember) the IDE/ATAPI drives.

Nothing, the command switched are already pre-set to run on boot:



Nice you've got that still running, btw. Changed the cmos battery for a new one, then?

 

Great! Thanks for your reply!

No, I didn't. I measured its voltage: 2.90 V. Sure, many years ago it was healtier...
So do you think that CMOS battery will be the responsible if Bios keeps on losing some hardware along his way? Should I replace it then?
More, could it be responsible for the fact that the system file in the windows folder wasn't found so that system cannot start up?
Could it cause all these jokes?

Anyway, in a diskette I saved BIOS 1.09 unzipped in a folder with the same name of the original zipped file and I managed to save even the 1.00.04 firmware in the same diskette (again, under a folder with the same name of the original zipped file). Do I have to insert a DOS command to let the system find the right folder?
Or do I have to move the bios files out of their current folder so that no folder contains them and autoexec can be immediately executed?

 

Yes, it could and while 2.9V should be alright, it's also ancient and $1 for a new one, so ...

You have to make a bootable disk first and unzip all bios files in the root of that disk (no folder). Boot from that and it'll flash bios to v1.09.

 

If the system can't remember the settings any more that points straight to the bios battery, but if it was only after removing the battery it forgot it once that is normal since power is required to store the settings.

 

I just created an MS DOS start up floppy disk (in windows XP under format options) and then I sent all the unzipped files to that disk. But a message asked me if I wanted to replace an existing file: autoexec.bat. So I agreed and now there is only 1 autoexec.bat, the one of the BIOS unzipped files. Doing so, I'm afraid that MS dos won't start.
Someone said that once loaded MS DOS with a start up diskette, you have to insert another diskette with the only BIOS files (so all the unzipped files). Is this a proper procedure? But then what should I type in dos to load bios?
I haven't tried to flash BIOS with the diskette I created yet. I firstly want to be sure that I'm going to follow the right steps with the right disk (with just the needed files).

 

That is good, it will boot regardless of an autoexec.bat present. All this does is run automated commands (the bios flash, in this case).

Only if the bios file won't fit to the disk. This it not necessary here, since the bios is small enough, so it's ok to run as is.

Remember that after flashing you go into the bios menu and set ' Reset to defaults' and hit ' Save and reboot'. You only need to do this once, after that just set everything as you like.

 

In my case, apart from RAM, floppy disk reader and video, do you think it's necessary to connect even the hard disk or something else to the motherboard for flashing bios?
I suppose they're enough... am I wrong?
Keeping HD and CD/DVD writer disconnected allows me to avoid setting boot priority in the BIOS. Last time I remember I couldn't even change the boot sequence via bios. I solved it by pressing a function key (perhaps F12?) that opened a window with a list of available hardware to choose for booting the system.

 

F12-option is fine, only need to boot from the floppy once. Everything else can stay as is.

That's a clear indication of a failed cmos battery. Either that, or a poor connection; could try cleaning the contacts.

 

Just to point out this particular step...
When BIOS flash is completed, I suppose I'll be notified and asked to reboot my computer. But before rebooting PC, as you suggested, I need to go into the BIOS setup (pressing a function key like F2 or simply selecting a different option when I'll be asked to reboot my PC?) and then set 'reset to defaults' ('load BIOS defaults' or something similar) and finally 'save and reboot'. Before this last step, I suppose I should remove the floppy disk and set boot priority as it was before the update. Is this the correct procedure or may I expect some surprises?
Someone said that after the flashing an error message appeared showing "CMOS Checksum Error - Defaults Loaded”. So I ask you, does this have something to do with CMOS battery? And for a successful flashing process is it necessary to have the CMOS battery at 100%?
I even have some CR 2032 batteries at home, but I'm not sure I have watched the right tutorial.
It seems that the black and red wires must be connected to the negative and positive poles and then attached with an insulating tape.
Finally a plastic material to wrap battery... Is a simple insulating tape or stronger? Because I've seen someone heating the wrapped battery with a hot gun (in order to melt the plastic). Isn't that exagerated?
Isn't it enough to wrap the battery with insulating tape without heating anything? If not, what kind of plastic material should be used to do this job? And finally, would a simple hairdryer be fine for this purpose?

 

Yes, spot on.

No and no.

Don't know, have you looked yet? Since it's a Clevo, chances are it's implemented like this:

Spoiler: good

... instead of this:

Spoiler: bad

... or this:

Spoiler: worst

If it happens to be one the the last two options, then get a battery holder and either plug it in or clip its wires and solder on. Forget about soldering the battery itself; a real pain and not a very solid connection (unless you have proper tools).

 

Sorry, I cannot see the second image, perhaps there is a mistake in the link.
Anyway, my CMOS battery is located in the bluetooth module, easily accessible from the bottom case:

As this service manual's image shows, the battery sits in that module (bluetooth board is in another image). The black and red wires coming from the battery are then connected to the motherboard with a stardard connector.
I don't know how the wires are attached inside the battery because I haven't removed the external tape yet. I've watched a tutorial where wires are laid on their sides and covered with a tape (I suppose an insulating tape but more likely is something different, I forgot its name...). The next step I saw is to wrap the battery with a plastic material (perhaps a kind of insulating tape), then cover all with a shrink sleeve and finally heat it with a hot gun to give the classic shape. I'm referring to this tutorial:
In my particular case, would you recommend such a method? The only thing is that I don't have any hot gun (but just hair dryer).

 

Most are soldered on (with tabs), this is a much better connection than with tape. Only things is that a battery's alloy is tricky stuff to solder; need to sand both sides of the battery and use good flux. The battery holder would be the easiest option by far.

For the simple tape method; splice the wires widely, so they have a nice, flattened surface area. Then wrap it tight with the insulating tape. Don't bother with the shrink-wrap, that's redundant.

 

Great! Precious information!
Now, back to BIOS issue, please. I got an annoying message concerning Phoenix Phlash16 Error: "Cannot flash if Memory Managers (e.g. HIMEM) present".
Perhaps some files of the formatted DOS diskette are disturbing the flash process. OK, but which, considering that there were only the files created with the Format option (MS dos start up floppy) under WinXp?
I even created a simple formatted diskette with the BIOS files. Could I use it after having pressed any key to exit flash page? It seems that after that I'm finally in DOS: A:\>
What should I type then to make the flash process start?

 

Odd ... usually it's the other way around; ' insufficient memory' error, which indicates himem isn't present. If there's a 'config.sys' file on the drive then rename its extension to .bak. That will disable all memory mangers, if present.

Second option; can it boot from usb stick as well? If so, that's much easier; create bootable stick using Rufus and copy the flash file to its root, replacing the autoexec.bak again. Rufus would need these parameters:

Spoiler: Rufus - bootable MS-DOS

 

Renamed Config.SYS file to Config.bak, still got that message. I tried to launch autoexec.bat via a formatted floppy disk (in WinXp) with the only BIOS files, but once again that message appeared.
Are there Memory Managers within the six BIOS files? Is it due to the fact that I created both diskettes via WinXp?
Someone said that Win98 doesn't have these problems.
Perhaps, I should create an MS DOS 4.01 (seen at this link: http://www.allbootdisks.com/disk_contents/dos.html ) bootable diskette just because it doesn't include Config.SYS and autoexec.bat (see the post #4 in this thread: http://www.computing.net/answers/windows-xp/phoenix-bios-upgrade-himem-message/192069.html ). Maybe those files cannot be loaded, otherwise the error message will appear. In fact, I overwrote autoexec.bat (coming from the MS DOS booting disk under winxp) but even this file disturbs the flashing process, besides Config.SYS that you suggested to rename. I think the mistake could be traced in the MS DOS version to be loaded at the beginning. Booting disks created under winxp cause that error. So I'd try with an earlier MS dos version, like the 4.01, Dos4.01.exe at this link: http://www.allbootdisks.com/download/dos.html
Otherwise, could it be an alternative solution to rename even autoexec.bat? Said that because I took a look at this thread: http://forum.notebookreview.com/threads/clevo-bios-thread.416916/
It seems that no one of the .bat files mentioned there, is the same as autoexec.bat. Is this because that's an interfering file name?
Please, give me some suggestions. I wish to run the floppy disk route before the bootable USB pen drive. Thanks!

 

Try DOS 4.01 first and, failing that, the usb second.

 

Finally, I managed to update the BIOS (diskette with 4.01 DOS) but fan control only provides 2 options: "continuos" and "automatics". I didn't even notice the difference. None of them works for my aim: I wish my CPU fan immediately started at full speed, so at full voltage (because otherwise it wouldn't start spinning).
In the BIOS there is no possibility to set voltage or Pwm. Could I set these parameters in a software like speedfan?
If I set "continuos", will I have the possibility via speedfan or similar software to manage the fan in order to let it start at full voltage?
I asked you these questions because I wish to avoid useless trials, as my motherboard is now disconnected and I should lose my time to reassemble hardware and case. And then? Realizing that speedfan or similar is unable to better manage fan control?

 

Not sure if this function was present then, but try Fn+1 key combination. That's ' 1', mind, not F1. Failing that, you may want to clean the fan and re-grease its bearing. Should expect that's long overdue and it won't run very well when dry. Not to mention noisy, at that.

Never had much luck with SpeedFan as a fan-control method, think it works mainly for desktops (nice program, though). May try HWiNFO instead. Problem here is that fans are controlled via the EC and Clevos were rather late to be supported. The D470W is probably not in its lookup-table yet, so you'd have to ask the author to add it (i.e, if possible).

 

Can I try that key combination when I'm in the BIOS? I mean, I don't need to load the operating system, do I?

 

Sure, if implemented then it works at all times.

 

Last update... Sadly, while I was in the BIOS, that key combination didn't seem to work. I even tried other combinations, as reported in this thread: http://forum.notebookreview.com/threads/clevo-m570u-bios-update-fan-speed-question.354890/
And that thread was opened in 2009. There, the combination was FN+F2 and Gophn, referring to a previous generation of Clevo notebooks, mentioned some Clevo models but not the D470W (a laptop with latest firmware and bios released in mid 2004).

Then, as said in this thread, http://forum.notebookreview.com/threads/np5792-5793-fan-speed-control.208249/ , the FN+1 hotkey was introduced in 2008 (January).
The question to ask is this: can d470w be included in the list of previous generation laptop? Or perhaps is too old to have any hotkey for full speed?
Do you think it's worth to ask this in a new thread?

 

No, it won't be updated. That needs a new firmware version and Clevo will not spend time on discontinued models. It is possible to force it full speed all the time, but not sure you'd want that (noisy). Make a photo of the fan, its cable and the fan cable's motherboard connector, in that case.

 

Fan cable and motherboard connector:

Fan photo from above

CPU fan electronic components

Someone suggested to short collector and emitter of transistor Q21 (the one controlling the fan). In this case, the full voltage arrives to the fan, making it spin at full speed.
Actually, I verified by shorting those pins with a screw driver and it works. To make it permanent I should apply a little solder between them. It's quite tight but doable with my soldering iron.
But I was wondering another thing... If I did something like that, the tachimeter wire (the yellow one) should become useless. What if I cut the yellow wire? No more signals to CPU... What's the reaction? Full speed all the time because the absence of signal is interpreted as a non-spinning fan? Or, as someone hypotized, a system block because CPU thinks that no CPU fan is connected and then blocks any possible activity in order to prevent damage?
Or perhaps none of the previous hypothesis?
Do you think it could be an alternative solution? What do I risk? That the fan could stop spinning even shorting collector and emitter of the transistor (in case the cut solution didn't work for my purpose)? If it's so, I'd prefer to leave the yellow wire where it is.
Any suggestions?

 

On my Clevo it automatically shuts down when it can't reach minimum desired rpm. Just leave it as is and use the short-pin option.

Alternatively, you can also use the JFAN2 solder blobs (if these are powered). Solder the red/black wires to pins 1 and 2, while leaving the tachy on pin #3 of JFAN1. That way you have full speed all the time, while leaving tachy/rpm report intact. Could also use a splice voltage and ground to both JFAN's; double the amperes. JFAN1 wouldn't do much then (since temperatures will be down), but it does provide twice the headroom (200% fan oc, or however much it can handle).

Cleaned/greased the fan, yet?

 

I normally knew something like this: http://m.newegg.com/Product/index?itemnumber=N82E16812200455
But here you're referring to the opposite situation: not a Y cable (2 fans on 1 header) but 1 fan connected to 2 headers.
By the way, the other fan of the laptop is a DDR fan, very small but working well. I was thinking of buying a splitter cable to connect the 2 fans to the header of the DDR fan (which is on the opposite side of the motherboard than JFAN1, JFAN2 and JFAN 4, all these very close instead). But then how will CPU react? And then, the 2 fans have 2 different connectors, so it would be impossible to find a splitter with different connectors. Finally, the electronic components of DDR fan circuit have different characteristics than the ones of the main fan. How would this affect the correct operation of the 2 fans? These doubts made me give up considering this solution.

So, the red and black wires coming from the fan should be soldered to pins 1 and 2 of JFAN2, while the yellow wire to pin 3 of JFAN1 (this perhaps to prevent automatic shutdown but are you sure that leaving out of connection the other 2 pins won't have negative consequences such as an automatic shutdown as well?). Next step would be to join the pins #1 of JFAN1 and JFAN2 with a small wire. Then I'd need to join the pins #2 of JFAN 1 and JFAN2 with another wire. Is this the right procedure of splicing voltage and ground? Would you use specific products to make something like that?

Yes, it was the very first thing.

 

Either short Q21, which will give you 100%, or use JFAN1 and JFAN2 and get ~100-200%:

 

Nice! Simple and effective explanation on the image!
I think I firstly opt for the fastest solution: shorting Q21. I only need to apply a very small amount of solder between pins. By the way, I heard someone talking about flux-resin core to put before solder. In my case, do you think it's necessary or can I just clean between the pins with alcol or something else?
Another doubt... I noticed that the smaller fan stops spinning when the main fan is at full speed (due to the short). And when I turn my motherboard on, shortpinning from the beginning (CPU fan's transistor), the smaller fan doesn't even start spinning. Is this because cpu doesn't send any command anymore to both fans, being CPU temperature always well below the fan activation's level?
In this case, I'd need to short transistor's pins of the smaller fan if I want even that fan spinning as the other one (and at full speed, no choice due to the shortpin unless I apply a resistor to decrease voltage and, consequently, speed). Do you agree?
While it wouldn't be a wise choice to apply a resistor to the CPU fan transistor (to decrease voltage-speed-noise) because in the last period I've heard that CPU quite often needed its fan at full speed (so I shouldn't impede to satisfy this CPU request), on the other hand it could make sense to apply a resistor to the other fan (DDR fan), since, as I explained before, it doesn't even start spinning when the main fan is at full speed (due to its short), so I shouldn't need the smaller fan spinning at full speed all the time. I could decrease its voltage-speed-noise with a simple resistor.
Since the DDR fan is a DC 5V 0.12A fan, how many ohms should the resistor need to guarantee an acceptable result?
And should I apply it between emitter and collector of the ddr fan transistor where I'm going to add some solder to short the 2 pins?

 

Clean is fine. No need for good flux either; you're just adding some lead-based solder, so doesn't require high temperatures. Most solder wire has a rosin flux core anyway.

A 5V/0.12A≈40Ω resistor rated fo (5V/40Ω)*5V≈0.65W (minimum).

Only, that may not be necessary; fan circuit (or the fan itself) should be limited already. Check how the small fan is doing when shorted first. Apply the resistor only when it's running way too fast.

In-between the voltage or ground wire of the ddr fan; cut wire, solder resistor and wrap with tape or, better, heatshrink.

 

Actually, it's too fast and noisy... Perhaps because DDR fan sits on the bottom case of the laptop.
Being at full contact with plastic, at full speed it seems to have a Boeing engine...
So, having modeled the fan as a resistor with about 40ohms, what if I try a 10ohm resistor to start? Let's say a quarter of its theoretical resistance.
How much power will be dissipated?

 

Four times as much, if there's that much available and the fan could handle that. The 40Ω would limit the fan to 105%, a 60Ω to 70% and an 80Ω to 52% of its maximum speed. Either will also diminish the resistor's required watts (the 0.65W @ 105%).

If you're unsure about what fan speed will be acceptable, then use a trimpot instead. That's a 200Ω rated for max. 0.75W. Since it sounds like there may be more than that on the connector you'll have to set it to a high value first; standard fan connector can do ~2.5W at most and that would kill this trimpot. Turn all the way to the right and connect it using the farthest two pins. The fan would run at ~20% when started. Then simply turn counter-clockwise until you've found a satisfactory speed.

Actually, it would be better to use a 300Ω or 400Ω version (since you could turn the fan completely off), but these are sold per 100 or 2,000, so ...

 

I'm just trying to figure out the resistance issue and last days I came across some interesting web pages.
Let me report these calculations:
5/0.12≈42ohms (let's say 40)
Current draw: 5/40=0.125A
Total power dissipated: 5×0.125=0.625W, with the resistor portion being 0.625×(10/40)=0.15625W (with a 10ohm resistor).

Now, it's time to verify with a couple of formulas (found on the internet) whether a 10ohm resistor could be fine for my purpose:

R=(cV-dV)/A
where
R=resistance
cV=current Voltage
dV=desired Voltage
A=fan ampere

So, hypothesizing a 10ohm resistor, we only have dV unknown:
10=(5-dV)/0.12 and then
dV=3.8V

With a 10ohm resistor the desired voltage should be 3.8V. It sounds correct.
Now, another formula for the power (always found on the internet):

P=R×A²
Then,
P=10×0.12²=10×0.0144=0.144W

0.144W is about a quarter than 0.625W. So, just correct me if I'm wrong, I should get the DDR fan spinning at about 3/4 of its maximum speed.

With those formulas I cannot verify your figures. For example, putting 100ohms in the first formula would give 7V of desired Voltage, more than its current theoretical voltage (5V).
Perhaps, those formulas are not correct. If you have other formulas more reliable than those, please post them. I didn't have enough time to find something else on the internet.

 

The A is also unknown, unless you've measured circuit with a multimeter. The fan can claim 0.12A, but that is not to say there won't be more on the circuit. You could supply 5V at 10A to the fan and it does 'whoosh'. It's also a fan, not a led. Maybe there'll be a little voltage drop, but nothing like a diode.

But sure, try a 10Ω, perhaps the result will be satisfactory. The trimpot is guaranteed success, though.

 

Get a variable resistor to test and then a fixed value when you are happy.

 

No, I haven't measured the current on the PCB. It would be necessary to break circuit, perhaps desoldering components (thing that I won't do, having already run the risk to burn both the fan transistors with a not so thin tip of my soldering iron and a not so thin solder wire). I-prober would be the ideal but it's very expensive. After all, I'd only need to measure voltage before and after applying a resistor to know how much current there is.

So I'm gonna test with a few resistors. Applying them in-between the voltage or ground wire of the fan won't be so difficult.

Anyway, just let me thank you for the precious support you've given me and the very useful tips since the beginning of this thread. Starting with the BIOS issue, ending with a "manual adjusted" fan control... now my system works fine thanks to the very effective contribution of NotebookReview community!