From: Nate Nagel on
Hi all, bought a used Precision M90 for dirt cheap a while back,
upgraded the HDD and memory, and have been using it as my primary
computer. On a recent trip to Japan (of course, couldn't happen at home)
the video card crapped out- garbage characters in ASCII and vertical
lines and lockup with the window manager running (GNOME, Ubuntu 9.04)
when I got home I disassembled the laptop, cleaned the heat sinks, baked
the video card (really!) and it's working again. However neither of the
cooling fans is ever coming on, and I installed the sensors-applet thing
and it says my GPU is getting up over 60C. I just received two "good
used" cooling fans in the mail today and the laptop is sitting on my
desk partially assembled and the fans still are not coming on, still
getting up over 60C with only GNOME running and a movie playing in a
window (and I believe it, the copper heat pipes are hot to the touch.)
Obviously it's only a matter of time until the video card or something
else craps out again with this kind of temp. I know the next step would
be to "run the diagnostics" but I don't have any of the original stuff
the computer came with save for the unit itself. Can I download these
from somewhere? Is this a common problem on this model? If so what is
the usual culprit? (please don't say "motherboard" and "buy a new
computer")

Also, does anyone know the pinout of the fans so that I could verify
that I didn't just somehow freakishly end up with four bad fans?

thanks

nate
--
replace "roosters" with "cox" to reply.
http://members.cox.net/njnagel
From: Paul on
Nate Nagel wrote:
> Hi all, bought a used Precision M90 for dirt cheap a while back,
> upgraded the HDD and memory, and have been using it as my primary
> computer. On a recent trip to Japan (of course, couldn't happen at home)
> the video card crapped out- garbage characters in ASCII and vertical
> lines and lockup with the window manager running (GNOME, Ubuntu 9.04)
> when I got home I disassembled the laptop, cleaned the heat sinks, baked
> the video card (really!) and it's working again. However neither of the
> cooling fans is ever coming on, and I installed the sensors-applet thing
> and it says my GPU is getting up over 60C. I just received two "good
> used" cooling fans in the mail today and the laptop is sitting on my
> desk partially assembled and the fans still are not coming on, still
> getting up over 60C with only GNOME running and a movie playing in a
> window (and I believe it, the copper heat pipes are hot to the touch.)
> Obviously it's only a matter of time until the video card or something
> else craps out again with this kind of temp. I know the next step would
> be to "run the diagnostics" but I don't have any of the original stuff
> the computer came with save for the unit itself. Can I download these
> from somewhere? Is this a common problem on this model? If so what is
> the usual culprit? (please don't say "motherboard" and "buy a new
> computer")
>
> Also, does anyone know the pinout of the fans so that I could verify
> that I didn't just somehow freakishly end up with four bad fans?
>
> thanks
>
> nate

In the picture here, it looks like the pinout is different than
a desktop fan pinout. I would guess the black wire is ground,
red wire is the positive supply wire, the yellow wire is the
RPM signal from the fan to the motherboard. But that is just a
guess. I wonder why it doesn't have the standard pinout ? What
advantage is there ?

http://image.siliconelectronic.net/FAN/Dell/Dell-04.jpg

To control fan speed, the +12V can be switched on and off by a
power transistor. The transistor could have failed, or the 12V
copper track on the motherboard, feeding the transistor, is burned
out. Something needs to control the transistor, and the SuperI/O
chip has fan speed controls. You can use Speedfan from almico, to
set fan speed. If the RPM wire is hooked up, you should also be able
to read the current fan speed. If the fan speed drops below a certain
RPM level, the SuperI/O reads the speed as zero. And that is because
the circuit measures the time period between two pulses on the RPM
signal wire. Once the time is larger than a certain period, the
counter used to measure that time period, overflows. So there is
a "lower limit" to the possible measured RPMs.

http://www.almico.com/speedfan440.exe

Speedfan cannot tell you, whether anything is defective. In good
faith, it can dial the registers inside the SuperI/O and tell the
fans to turn. But if something is broken outside the SuperI/O, there
is no way to tell. You'd need to debug with a multimeter or other
test equipment, with the laptop opened up, to see if it was getting a
signal, check the transistor to see if it is powered and so on.

On desktop motherboards, where the user connected too large an
electrical load to the three pin fan header, sometimes the
copper track gets burned out. For some reason, they don't put
a fuse on that track. So people have managed to damage their motherboard,
while fooling around with the fan headers. But the laptop is typically
a bit more complicated. On a desktop, you might take a multimeter
set to the ohms range, and tell it was broken, by ohming between
the 12V wire on a power connector, and where the red wire goes on
the fan header. With something like a laptop, the power transistor
could be in that path, so the test results may not be valid. You would
need to come up with some other test, to be able to tell what was going on.

In any case, put it together, and give Speedfan a try. Dial the fan to
100% and see if it spins. If you're lucky, the power transistor used
to switch the power, is located right next to the fan header. But there
is nothing in electronics design, that makes that an absolute necessity.
If they wanted, the power transistor could be place a foot away from the
header. So while proximity is something you might come to expect in a
layout, there is no guarantee the transistor and associated circuits are
right next to the header.

Speedfan, while probing the hardware, should tell you what SuperI/O chip
is used. This is the chip used on my motherboard. There are details
in there, as to the fan interface, supported control methods and so on.
If you want a little bit of background, with luck, you can get some info.

http://www.nuvoton.com/NR/rdonlyres/7E862DA6-DFBD-4FAC-A2E0-D549FF86663C/0/W83627DHGP_W83627DHGPT.pdf

If you look at PDF page 37 on this one, it shows a circuit for converting
from SuperI/O PWM fan control output, into a voltage the fan can use.
Figure 2.5 shows how a PNP power transistor, connected to +12V, "pumps" up
a capacitor tied across the fan. If that transistor was damaged, there
would be no conversion from some PWM (pulse width modulation) setting,
into a voltage the fan could use. The fan would then fail to spin.
With luck, that transistor is next to the fan header.

http://web.archive.org/web/20061118133939/http://www.winbond.com/PDF/sheet/w83627hf.pdf

Paul
From: N8N on
On Apr 27, 12:17 am, Paul <nos...(a)needed.com> wrote:
> Nate Nagel wrote:
> > Hi all, bought a used Precision M90 for dirt cheap a while back,
> > upgraded the HDD and memory, and have been using it as my primary
> > computer. On a recent trip to Japan (of course, couldn't happen at home)
> > the video card crapped out- garbage characters in ASCII and vertical
> > lines and lockup with the window manager running (GNOME, Ubuntu 9.04)
> > when I got home I disassembled the laptop, cleaned the heat sinks, baked
> > the video card (really!) and it's working again. However neither of the
> > cooling fans is ever coming on, and I installed the sensors-applet thing
> > and it says my GPU is getting up over 60C. I just received two "good
> > used" cooling fans in the mail today and the laptop is sitting on my
> > desk partially assembled and the fans still are not coming on, still
> > getting up over 60C with only GNOME running and a movie playing in a
> > window (and I believe it, the copper heat pipes are hot to the touch.)
> > Obviously it's only a matter of time until the video card or something
> > else craps out again with this kind of temp. I know the next step would
> > be to "run the diagnostics" but I don't have any of the original stuff
> > the computer came with save for the unit itself. Can I download these
> > from somewhere? Is this a common problem on this model? If so what is
> > the usual culprit? (please don't say "motherboard" and "buy a new
> > computer")
>
> > Also, does anyone know the pinout of the fans so that I could verify
> > that I didn't just somehow freakishly end up with four bad fans?
>
> > thanks
>
> > nate
>
> In the picture here, it looks like the pinout is different than
> a desktop fan pinout. I would guess the black wire is ground,
> red wire is the positive supply wire, the yellow wire is the
> RPM signal from the fan to the motherboard. But that is just a
> guess. I wonder why it doesn't have the standard pinout ? What
> advantage is there ?
>
> http://image.siliconelectronic.net/FAN/Dell/Dell-04.jpg
>
> To control fan speed, the +12V can be switched on and off by a
> power transistor. The transistor could have failed, or the 12V
> copper track on the motherboard, feeding the transistor, is burned
> out. Something needs to control the transistor, and the SuperI/O
> chip has fan speed controls. You can use Speedfan from almico, to
> set fan speed. If the RPM wire is hooked up, you should also be able
> to read the current fan speed. If the fan speed drops below a certain
> RPM level, the SuperI/O reads the speed as zero. And that is because
> the circuit measures the time period between two pulses on the RPM
> signal wire. Once the time is larger than a certain period, the
> counter used to measure that time period, overflows. So there is
> a "lower limit" to the possible measured RPMs.

So the fan is typically 12VDC then? That's good news as then I could
just use a car battery or the 12V power supply I have kicking around
to bench-test automotive type stuff.

(rest of excellent reply snipped)

If I can verify that the fans are good and they're still not working,
I may just download the utility you recommended and see if I can force
them on; obviously I am getting temp readings as sensors-applet is
reporting same. If I can't force them on I guess it's new mobo time
unless I want to get into component level troubleshooting. Not sure
how motivated I really am vs. just spending $100ish for a "good used"
mobo...

thanks

nate
 | 
Pages: 1
Prev: McAfee MVT
Next: T410 vs T510