Hello, folks,

I have trouble with my a21p: Symptoms are:

When booting or shortly afterwards (perhaps up to 10 minutes) the laptop
switches off. More accurately: The displays goes off, harddisk doesn't
spin any more, connections via ethernet are interrupted.


What I noticed:
the processor remains moderately warm (purhaps 30 - 45 degrees Celsius)
while it is working.  After the above declared  switch off it gets hot
(above 50 degrees approximately).


What I did so far:
- checked the ram (2 x 128 MB)

- changed the mains adapter

- removed the mini pci card

- switched off the subcard 

- changed the system board

- replaced the processor, intel pentium III mobile 850MHz, by a used one
from ebay (seller wrote me that it was tested 7 hours with the help of
the program prime95)


What I plan:
Check the logs. As far as I know: they don't contain anything.


What I wonder: 
How likely is it that processors bought at ebay are as good as they are
declared? In other words: is it too unklikely that both processors are
faulty? 
What do you say? 
Which other reason do you see for the malfunctioning?


What I beg to note: 
The switch off is often so fast that functioning or non-functioning of
the fan cannot be the reason in my opinion.  (The fan spins up at boot
time and stops afterward, it did not start up again, so far). 
Additonally: I check the temperature of the processor by hand (feeling
the fan in the inside of the laptop) and it does not get hot before the
switch off. 


Tilo

On Sat, 6 Oct 2007 14:22:40 +0000 (UTC) Tilo Baumann wrote:

> Hello, folks,
> 
> I have trouble with my a21p: Symptoms are:
> 
> When booting or shortly afterwards (perhaps up to 10 minutes) the laptop
> switches off. More accurately: The displays goes off, harddisk doesn't
> spin any more, connections via ethernet are interrupted.
> 
> 
> What I noticed:
> the processor remains moderately warm (purhaps 30 - 45 degrees Celsius)
> while it is working.  After the above declared  switch off it gets hot
> (above 50 degrees approximately).
> 
Suggestion:  Download and install Everest home edition, freeware, from 
<http://www.majorgeeks.com/EVEREST_Home_Edition_d4181.html>.  Monitor the CPU 
temperature and fan speed: Computer > Sensor.

Visit <http://processorfinder.intel.com/list.aspx?ProcFam=55> and verify the 
temperature specifications for your processor.  Most Intel processors will 
shut down, rather abruptly, when they overheat.
> 
<snip>
> 
> What I wonder: 
> How likely is it that processors bought at ebay are as good as they are
> declared? In other words: is it too unklikely that both processors are
> faulty? 

Not very likely imo.  If a CPU works at all it's probably all right.

> What do you say? 
> Which other reason do you see for the malfunctioning?
> 
The cooling fan sensor is malfunctioning.  I had a similar problem on a TP 
R40, the cooling fan wouldn't start.  The computer would run until the 
processor overheated, then it would shut down.  The processor was unharmed but 
the hard drive was damaged and required replacement.
> 
> What I beg to note: 
> The switch off is often so fast that functioning or non-functioning of
> the fan cannot be the reason in my opinion.  (The fan spins up at boot
> time and stops afterward, it did not start up again, so far). 

.... but the CPU continues to generate heat.  The fan should cycle on and off 
to maintain a set temperature.

> Additonally: I check the temperature of the processor by hand (feeling
> the fan in the inside of the laptop) and it does not get hot before the
> switch off. 
> 
The *fan* wouldn't get very hot.  Monitor the CPU temperature as suggested 
above.

A suggestion:  Rest the computer on an ice bag, keep it dry of course and 
don't block the air inlet, and see if the computer operates normally for a 
longer period of time.  My guess: it will do fine as long as it is kept cool.
-- 
Ernie B.

Communication:  The art of moving an idea from one mind to another, hopefully 
without distortion.

On 2007-10-06, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:
> On Sat, 6 Oct 2007 14:22:40 +0000 (UTC) Tilo Baumann wrote:

>> What I noticed:
>> the processor remains moderately warm (purhaps 30 - 45 degrees Celsius)
>> while it is working.  After the above declared  switch off it gets hot
>> (above 50 degrees approximately).
>> 
> Suggestion:  Download and install Everest home edition, freeware, from 
><http://www.majorgeeks.com/EVEREST_Home_Edition_d4181.html>.  Monitor the CPU 
> temperature and fan speed: Computer > Sensor.

There are possibilities under linux to measure cpu temperature.  I did
measure the old processor and the temperature remained moderate until
the switch off.  After that nothing more is measured. 


> Visit <http://processorfinder.intel.com/list.aspx?ProcFam=55> and verify the 
> temperature specifications for your processor.  Most Intel processors will 
> shut down, rather abruptly, when they overheat.

It says: Thermal Design Power: 27.5W and Thermal Specification: 100�C. 


>> Which other reason do you see for the malfunctioning?
>> 
> The cooling fan sensor is malfunctioning.  I had a similar problem on a TP 
> R40, the cooling fan wouldn't start.  The computer would run until the 
> processor overheated, then it would shut down.  The processor was unharmed but 
> the hard drive was damaged and required replacement.

>> What I beg to note: 
>> The switch off is often so fast that functioning or non-functioning of
>> the fan cannot be the reason in my opinion.  (The fan spins up at boot
>> time and stops afterward, it did not start up again, so far). 
>
> ... but the CPU continues to generate heat.  The fan should cycle on and off 
> to maintain a set temperature.

There is a second a21p working on my desk, and I know about the load
necessary in this type of notebook to start the fan. 

The a21p has got a subcard where the fan is plugged into.  Might be that
this card is faulty (too) in that it doesn't deliver the voltage for the
fan appropriately. 

>> Additonally: I check the temperature of the processor by hand (feeling
>> the fan in the inside of the laptop) and it does not get hot before the
>> switch off. 

I measured the temperature of the old processor (as above said) by
program and it remained low.  I will repeat this measurement with the
second processor.  But often booting is not yet finished when the switch
off happens.  So there should be a fault beside the fan potentially not
being switched on appropriately.  I ask for what this fault might be. 


> A suggestion:  Rest the computer on an ice bag, keep it dry of course and 
> don't block the air inlet, and see if the computer operates normally for a 
> longer period of time.  My guess: it will do fine as long as it is kept cool.

Will it help if your processor has an inner shortcut?



Tilo

On Sat, 6 Oct 2007 18:30:57 +0000 (UTC) Tilo Baumann wrote:

> There are possibilities under linux to measure cpu temperature.  I did
> measure the old processor and the temperature remained moderate until
> the switch off.  After that nothing more is measured. 
> 
Okay.  I'm not familiar with linux, I suppose that you achieved an accurate 
temperature measurement.
> 
> It says: Thermal Design Power: 27.5W and Thermal Specification: 100�C. 
> 
.... and the measured temperature was well below 100�C?
> 
>
> > ... but the CPU continues to generate heat.  The fan should cycle on and off 
> > to maintain a set temperature.
> 
> There is a second a21p working on my desk, and I know about the load
> necessary in this type of notebook to start the fan. 
> 
> The a21p has got a subcard where the fan is plugged into.  Might be that
> this card is faulty (too) in that it doesn't deliver the voltage for the
> fan appropriately. 
> 
That was the apparent problem with my R40.  The sub-card didn't start the fan.  
I still have the original assembly and may wire it to bypass the sub-card and 
run the fan continuously.  That will drain the battery a little faster but 
should keep things cool.
> 
> I measured the temperature of the old processor (as above said) by
> program and it remained low.  I will repeat this measurement with the
> second processor.  But often booting is not yet finished when the switch
> off happens.  So there should be a fault beside the fan potentially not
> being switched on appropriately.  I ask for what this fault might be. 
> 
It still seems like a temperature problem to me, however caused.
> 
> > A suggestion:  Rest the computer on an ice bag, keep it dry of course and 
> > don't block the air inlet, and see if the computer operates normally for a 
> > longer period of time.  My guess: it will do fine as long as it is kept cool.
> 
> Will it help if your processor has an inner shortcut?
> 
No, but you've tried two processors with similar results.  That would seem to 
eliminate the processor as the cause.
-- 
Ernie B.

Communication:  The art of moving an idea from one mind to another, hopefully 
without distortion.

On Sat, 06 Oct 2007 19:53:33 GMT Ernie B. wrote:

> > > A suggestion:  Rest the computer on an ice bag, keep it dry of course and 
> > > don't block the air inlet, and see if the computer operates normally for a 
> > > longer period of time.  My guess: it will do fine as long as it is kept cool.
> > 
> > Will it help if your processor has an inner shortcut?
> > 
> No, but you've tried two processors with similar results.  That would seem to 
> eliminate the processor as the cause.
> 
Something else that you might consider:  There could be a hairline crack in 
one of the printed circuit runs; perhaps near the processor, its connector or 
cable, that is opening when the board/cable gets warm.  Some time with a 
magnifying lens might be well spent.
-- 
Ernie B.

Communication:  The art of moving an idea from one mind to another, hopefully 
without distortion.

On 2007-10-06, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:
> On Sat, 6 Oct 2007 18:30:57 +0000 (UTC) Tilo Baumann wrote:

>> It says: Thermal Design Power: 27.5W and Thermal Specification: 100�C. 
>> 
> ... and the measured temperature was well below 100�C?

I measured the 2nd CPU: starting from 24 to 33 degree Celsius within 25
minutes. At 33 the switch off took place.

When switching off the computer was not dead -- contrary to what I said
before. Instead the harddrive was switched on and off as if the computer
searched for a boot device unsuccessfully. 

Next time I will put a diskette into the floppy drive to see whether
that drive will be  checked, too.

As I assumed in a former posting: the logs don't say anything about the
death. (My sylogd gets really surprised.) 


>> > A suggestion:  Rest the computer on an ice bag, keep it dry of course and 
>> > don't block the air inlet, and see if the computer operates normally for a 
>> > longer period of time.  My guess: it will do fine as long as it is kept cool.
>> 
>> Will it help if your processor has an inner shortcut?
>> 
> No, but you've tried two processors with similar results.  That would seem to 
> eliminate the processor as the cause.

I thought so, too.  But at the moment I don't have any other idea,
though I will look further after what I called 'search for boot device'. 
It would be interesting to learn how often processors fail because of
inner shortcuts. 


Tilo

On 2007-10-06, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:
> On Sat, 06 Oct 2007 19:53:33 GMT Ernie B. wrote:

> Something else that you might consider:  There could be a hairline crack in 
> one of the printed circuit runs; perhaps near the processor, its connector or 
> cable, that is opening when the board/cable gets warm.  Some time with a 
> magnifying lens might be well spent.

Thanks! I installed another mainboard to avoid that case.  Both
mainboards offer the same fault symptoms.  That is against likelyhood. 
(Same as with the processor. For me, however: I trust the 2nd mainboard
more than the 2nd processor.)


Tilo

On Sat, 6 Oct 2007 21:20:08 +0000 (UTC) Tilo Baumann wrote:

> On 2007-10-06, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:
> > On Sat, 06 Oct 2007 19:53:33 GMT Ernie B. wrote:
> 
> > Something else that you might consider:  There could be a hairline crack in 
> > one of the printed circuit runs; perhaps near the processor, its connector or 
> > cable, that is opening when the board/cable gets warm.  Some time with a 
> > magnifying lens might be well spent.
> 
> Thanks! I installed another mainboard to avoid that case.  Both
> mainboards offer the same fault symptoms.  That is against likelyhood. 
> (Same as with the processor. For me, however: I trust the 2nd mainboard
> more than the 2nd processor.)
> 
Did you replace the cables?
-- 
Ernie B.

Communication:  The art of moving an idea from one mind to another, hopefully 
without distortion.

On Sat, 6 Oct 2007 21:12:45 +0000 (UTC) Tilo Baumann wrote:

> I measured the 2nd CPU: starting from 24 to 33 degree Celsius within 25
> minutes. At 33 the switch off took place.
> 
Okay, so the CPU isn't overheating.  The failure of the fan to run when it 
should be running makes me believe that *something* is getting too hot.

> When switching off the computer was not dead -- contrary to what I said
> before. Instead the harddrive was switched on and off as if the computer
> searched for a boot device unsuccessfully. 
> 
That's strange.  Are you sure that the hard drive is all right?  You could 
swap with the other A21p and see if that tells you anything.

> Next time I will put a diskette into the floppy drive to see whether
> that drive will be  checked, too.
> 
Maybe with a boot disk...  

> As I assumed in a former posting: the logs don't say anything about the
> death. (My sylogd gets really surprised.) 
> 
Not unexpected.  
> 
> I thought so, too.  But at the moment I don't have any other idea,
> though I will look further after what I called 'search for boot device'. 
> It would be interesting to learn how often processors fail because of
> inner shortcuts. 
> 
Not very often I think, although they sometimes grow internal "whiskers".  See 
<http://nepp.nasa.gov/whisker/> for some information.  It would be very rare 
for two processors, probably from separate batches with different hours of 
use, to have identical failure modes.
-- 
Ernie B.

Communication:  The art of moving an idea from one mind to another, hopefully 
without distortion.

Well My problem is similar but different.   When I am doing an image
backup, or sometimes when defragging the c: dr takes over 10 minutes
the screen blanks on my lenovo T-32.   I know where must be a way
in XP/PRO to set the length of the "screen blanking time limit"
I dont know where in CTRLPanel or where to go to set a longer period
of time before screen blanking occurs.
                              Joe

On Sat, 6 Oct 2007 13:07:57 -1000 Joseph Fenn wrote:

> Well My problem is similar but different.   When I am doing an image
> backup, or sometimes when defragging the c: dr takes over 10 minutes
> the screen blanks on my lenovo T-32.   I know where must be a way
> in XP/PRO to set the length of the "screen blanking time limit"
> I dont know where in CTRLPanel or where to go to set a longer period
> of time before screen blanking occurs.
>                               Joe
> 
With Win2K Pro, XP should be similar...

Right click the desktop, select Properties.  Select the Screen Saver tab > 
Power button and set the time for the various functions.

BTW, "screen blanking" won't harm whatever the computer is doing, it just 
turns the display off after xx minutes of mouse or keyboard inactivity.
-- 
Ernie B.

Communication:  The art of moving an idea from one mind to another, hopefully 
without distortion.

On 2007-10-06, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:

> Did you replace the cables?

There are no cables,

except very few minor important ones (e.g.  to the disk drive).  The
different parts are mainly connected by plug ins.


Tilo

On 2007-10-06, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:
> On Sat, 6 Oct 2007 21:12:45 +0000 (UTC) Tilo Baumann wrote:

> That's strange.  Are you sure that the hard drive is all right?  You could 
> swap with the other A21p and see if that tells you anything.

It is tested ok.

>> Next time I will put a diskette into the floppy drive to see whether
>> that drive will be  checked, too.
>> 
> Maybe with a boot disk...  

Yes, I will do


>> I thought so, too.  But at the moment I don't have any other idea,
>> though I will look further after what I called 'search for boot device'. 
>> It would be interesting to learn how often processors fail because of
>> inner shortcuts. 
>> 
> Not very often I think, although they sometimes grow internal "whiskers".  See 
><http://nepp.nasa.gov/whisker/> for some information.  It would be very rare 
> for two processors, probably from separate batches with different hours of 
> use, to have identical failure modes.

The link showed 'external' whiskers.

For my first processor I can imagine the following scenario: the subcard
is faulty in that it does not switch on the fan appropriately.  The
processor gets hot => (*) Shortcuts.  I don't know, though, if (*) is a
typical consequence. 

As overheating of a laptop processor is not seldom the second processor
might have been overheated as well. 


Tilo

On Mon, 8 Oct 2007 15:59:04 +0000 (UTC) Tilo Baumann wrote:

> On 2007-10-06, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:
> 
> > Did you replace the cables?
> 
> There are no cables,
> 
> except very few minor important ones (e.g.  to the disk drive).  The
> different parts are mainly connected by plug ins.
> 
Okay.  You could clean the male plug-in connections with a soft pencil eraser, 
e.g. the Faber "Pink Pearl", to eliminate any problems there.
-- 
Ernie B.

Communication:  The art of moving an idea from one mind to another, hopefully 
without distortion.

On Mon, 8 Oct 2007 16:22:43 +0000 (UTC) Tilo Baumann wrote:

> > Not very often I think, although they sometimes grow internal "whiskers".  See 
> ><http://nepp.nasa.gov/whisker/> for some information.  It would be very rare 
> > for two processors, probably from separate batches with different hours of 
> > use, to have identical failure modes.
> 
> The link showed 'external' whiskers.
> 
Yes, to illustrate the problem.  I seem to recall that some early processor 
failures were traced to internal whiskers though.  Probably not very common 
these days.

> For my first processor I can imagine the following scenario: the subcard
> is faulty in that it does not switch on the fan appropriately.  The
> processor gets hot => (*) Shortcuts.  I don't know, though, if (*) is a
> typical consequence. 
> 
Does the fan work properly now?

An Intel processor will usually shut down but will survive.  See 
<http://www.youtube.com/watch?v=XgOmMAasqto>.

> As overheating of a laptop processor is not seldom the second processor
> might have been overheated as well. 
> 
I doubt that this is the problem since you seem to have identical failures 
with either processor.  If the processor is bad it usually won't boot at all.

Did you operate the laptop while taking steps to keep it cool?  What was the 
result?

Have you looked at the power supply circuitry?  Something happening there 
could cause loss of one or more of the voltages.

I've about run out of ideas at this point. :-(
-- 
Ernie B.

Communication:  The art of moving an idea from one mind to another, hopefully 
without distortion.

On 2007-10-08, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:
> On Mon, 8 Oct 2007 16:22:43 +0000 (UTC) Tilo Baumann wrote:

>> For my first processor I can imagine the following scenario: the subcard
>> is faulty in that it does not switch on the fan appropriately.  The
>> processor gets hot => (*) Shortcuts.  I don't know, though, if (*) is a
>> typical consequence. 
>> 
> Does the fan work properly now?

I can't know. My 'switch-off' takes place at 33 degree C.


>> As overheating of a laptop processor is not seldom the second processor
>> might have been overheated as well. 
>> 
> I doubt that this is the problem since you seem to have identical failures 
> with either processor.  If the processor is bad it usually won't boot at all.

Where do you know that from?

> Did you operate the laptop while taking steps to keep it cool?  What was the 
> result?

33 degree Celsius _is_ cool. 

> Have you looked at the power supply circuitry?  Something happening there 
> could cause loss of one or more of the voltages.

I don't know anything about the circuitry on the board. The (external)
adapter is ok.

 
> I've about run out of ideas at this point. :-(

Let us cry for fresh helpers!


Tilo

On Mon, 8 Oct 2007 21:52:02 +0000 (UTC) Tilo Baumann wrote:

> > Does the fan work properly now?
> 
> I can't know. My 'switch-off' takes place at 33 degree C.
> 
Okay, 33�C measured at the CPU?  The fan runs to begin with?  If so, this 
seems odd since the fan shouldn't run until the CPU reaches a warm 
temperature.
> >> 
> > I doubt that this is the problem since you seem to have identical failures 
> > with either processor.  If the processor is bad it usually won't boot at all.
> 
> Where do you know that from?
> 
Mostly from experience in trying to install a faster processor in an old 
Hyundai laptop.  The processor overheated, got entirely too hot to touch, 
unless we blew cool air directly onto it.  It would run fine if kept cool even 
after being overheated to the point of shutdown several times but wouldn't 
boot unless allowed to cool off.  Also, my Thinkpad R40 is running fine on the 
original processor even after the fan sensor failure and replacement.  The 
hard drive was damaged, made a clicking noise and was unreliable, so it was 
replaced.

> > Did you operate the laptop while taking steps to keep it cool?  What was the 
> > result?
> 
> 33 degree Celsius _is_ cool. 
> 
33�C = 91.4�F.  The CPU is cool enough I suppose since the CPU in my desktop 
is presently running at 38�C, 100.4�F.  Are there any noticeably hot areas to 
be felt anywhere on the computer?  Have you tried running it while placed on 
an ice pack?

> > Have you looked at the power supply circuitry?  Something happening there 
> > could cause loss of one or more of the voltages.
> 
> I don't know anything about the circuitry on the board. The (external)
> adapter is ok.
> 
The external adapter supplies a DC voltage to the computer.  The circuitry I 
was referring to modifies this voltage and supplies power to the various 
components inside the computer.  For instance: five volts to the fan, 1.4 
volts to the CPU core, 3.3 volts to the I/O system, etc.

You do have a service manual?  I would suggest that 
<http://tinyurl.com/37exy8>, especially the first item, looks promising.
>  
> > I've about run out of ideas at this point. :-(
> 
> Let us cry for fresh helpers!
> 
Absolutely!  If anyone following this thread spots something we've overlooked, 
please jump in.
-- 
Ernie B.

Communication:  The art of moving an idea from one mind to another, hopefully 
without distortion.

On 2007-10-09, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:
> On Mon, 8 Oct 2007 21:52:02 +0000 (UTC) Tilo Baumann wrote:
>
>> > Does the fan work properly now?
>> 
>> I can't know. My 'switch-off' takes place at 33 degree C.
>> 
> Okay, 33�C measured at the CPU?  The fan runs to begin with?  If so, this 
> seems odd since the fan shouldn't run until the CPU reaches a warm 
> temperature.

Measured at the CPU, yes.
The fan runs while booting (like all thinkpads afaik) and then stops. It
never starts up again. Starting again is not necessary until the faulty
switch off because temperature is normal until that time.
Behavior of the fan after the faulty switch off is ominous, though. 


>> > Did you operate the laptop while taking steps to keep it cool?  What was the 
>> > result?
>> 
>> 33 degree Celsius _is_ cool. 
>> 
> 33�C = 91.4�F.  The CPU is cool enough I suppose since the CPU in my desktop 
> is presently running at 38�C, 100.4�F.  Are there any noticeably hot areas to 
> be felt anywhere on the computer?  Have you tried running it while placed on 
> an ice pack?

The situation here is as follows: the keyboard bezel is removed.  The
keyboard itself is attached by being plugged into the appropriate slot. 
I seem to need the keyboard to switch on the laptop.  

So the part between hinges and keyboard along the whole width of the
laptop is open.  That's the place for the upper part of the fan and for
the subcard.  This part of the fan is approximately as warm as is
measured. 

After the faulty switch off it gets warm and hot until I switch off the
laptop manually.  That takes about 2 minutes.  During the time between
the faulty switch off at 33�C and the manual switch off at estimated 
> 70�C the system tries to start the harddisk all the time.  (I didn't yet
check if the floppy drive is started, as well.)

I did neither try using an ice pack for the following reason: it seems
that the faulty switch off is not caused by too hot the processor.


>> > Have you looked at the power supply circuitry?  Something happening there 
>> > could cause loss of one or more of the voltages.
>> 
>> I don't know anything about the circuitry on the board. The (external)
>> adapter is ok.
>> 
> The external adapter supplies a DC voltage to the computer.  The circuitry I 
> was referring to modifies this voltage and supplies power to the various 
> components inside the computer.  For instance: five volts to the fan, 1.4 
> volts to the CPU core, 3.3 volts to the I/O system, etc.

How can I check if they are functioning? 


> You do have a service manual?  I would suggest that 
><http://tinyurl.com/37exy8>, especially the first item, looks promising.

I have got the hmm printed for nearly all of my thinkpads. I have got it
for the a21p. 

>> > I've about run out of ideas at this point. :-(
>> 
>> Let us cry for fresh helpers!
>> 
> Absolutely!  If anyone following this thread spots something we've overlooked, 
> please jump in.

Thanks for yelling.


I see two aspects to explore beside others:

1. If it is true that the system tries to boot, I wonder how it gets to it.
Booting seems setting the instruction pointer within the processor to 
instruction zero. That seems to be a reset. Which events do reset the processor?

2. What about the speedstep property of the pentium 3?
In my intact a21p when booting linux detects different speeds:

Apr  1 03:01:51 localhost kernel: Detected 697.798 MHz processor.
Sep 21 14:03:48 localhost kernel: Detected 190.253 MHz processor.
Sep 22 08:29:33 localhost kernel: Detected 190.777 MHz processor.
Oct  6 11:06:47 localhost kernel: Detected 488.061 MHz processor.

with different evaluated BogoMIPS.

As far as I remember there are no speeds of 190MHz or 488MHz in a 850MHz
Pentium 3 mobile. 

Linux changes the speed of the cpu by so called governors. It seems
useful for me to control the governors, and I appreciate hearing proposals. 

On the other side I acted on the faulty laptop with two harddisks:
- one with a new linux installation
- one with the linux installation of the working a21p



Tilo

On 2007-10-09, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:
> On Mon, 8 Oct 2007 21:52:02 +0000 (UTC) Tilo Baumann wrote:
>
>> > Does the fan work properly now?
>> 
>> I can't know. My 'switch-off' takes place at 33 degree C.
>> 
> Okay, 33�C measured at the CPU?  The fan runs to begin with?  If so, this 
> seems odd since the fan shouldn't run until the CPU reaches a warm 
> temperature.

Measured at the CPU, yes.
The fan runs while booting (like all thinkpads afaik) and then stops. It
never starts up again. Starting again is not necessary until the faulty
switch off because temperature is normal until that time.
Behavior of the fan after the faulty switch off is ominous, though. 


>> > Did you operate the laptop while taking steps to keep it cool?  What was the 
>> > result?
>> 
>> 33 degree Celsius _is_ cool. 
>> 
> 33�C = 91.4�F.  The CPU is cool enough I suppose since the CPU in my desktop 
> is presently running at 38�C, 100.4�F.  Are there any noticeably hot areas to 
> be felt anywhere on the computer?  Have you tried running it while placed on 
> an ice pack?

The situation here is as follows: the keyboard bezel is removed.  The
keyboard itself is attached by being plugged into the appropriate slot. 
I seem to need the keyboard to switch on the laptop.  

So the part between hinges and keyboard along the whole width of the
laptop is open.  That's the place for the upper part of the fan and for
the subcard.  This part of the fan is approximately as warm as is
measured. 

After the faulty switch off it gets warm and hot until I switch off the
laptop manually.  That takes about 2 minutes.  During the time between
the faulty switch off at 33�C and the manual switch off at estimated 
> 70�C the system tries to start the harddisk all the time.  (I didn't yet
check if the floppy drive is started, as well.)

I did neither try using an ice pack for the following reason: it seems
that the faulty switch off is not caused by too hot the processor.


>> > Have you looked at the power supply circuitry?  Something happening there 
>> > could cause loss of one or more of the voltages.
>> 
>> I don't know anything about the circuitry on the board. The (external)
>> adapter is ok.
>> 
> The external adapter supplies a DC voltage to the computer.  The circuitry I 
> was referring to modifies this voltage and supplies power to the various 
> components inside the computer.  For instance: five volts to the fan, 1.4 
> volts to the CPU core, 3.3 volts to the I/O system, etc.

How can I check if they are functioning? 


> You do have a service manual?  I would suggest that 
><http://tinyurl.com/37exy8>, especially the first item, looks promising.

I have got the hmm printed for nearly all of my thinkpads. I have got it
for the a21p. 

>> > I've about run out of ideas at this point. :-(
>> 
>> Let us cry for fresh helpers!
>> 
> Absolutely!  If anyone following this thread spots something we've overlooked, 
> please jump in.

Thanks for yelling.


I see two aspects to explore beside others:

1. If it is true that the system tries to boot, I wonder how it gets to it.
Booting seems setting the instruction pointer within the processor to 
instruction zero. That seems to be a reset. Which events do reset the processor?

2. What about the speedstep property of the pentium 3?
In my intact a21p when booting linux detects different speeds:

Apr  1 03:01:51 localhost kernel: Detected 697.798 MHz processor.
Sep 21 14:03:48 localhost kernel: Detected 190.253 MHz processor.
Sep 22 08:29:33 localhost kernel: Detected 190.777 MHz processor.
Oct  6 11:06:47 localhost kernel: Detected 488.061 MHz processor.

with different evaluated BogoMIPS.

As far as I remember there are no speeds of 190MHz or 488MHz in a 850MHz
Pentium 3 mobile. 

Linux changes the speed of the cpu by so called governors.  It seems
useful for me to check for the governors, and I appreciate hearing
proposals. 

On the other side I acted on the faulty laptop with two harddisks:
- one with a new linux installation
- one with the linux installation of the working a21p



Tilo

On Wed, 10 Oct 2007 22:52:16 +0000 (UTC) Tilo Baumann wrote:

> > Okay, 33�C measured at the CPU?  The fan runs to begin with?  If so, this 
> > seems odd since the fan shouldn't run until the CPU reaches a warm 
> > temperature.
> 
> Measured at the CPU, yes.
> The fan runs while booting (like all thinkpads afaik) and then stops.

Well, you caused me to look. ;-)  On my R40 the fan _does_not_ run while 
booting.  Placed on a desk, it runs for about ten seconds after the CPU warms 
up, then off for a time then cycles on and off as required to maintain a CPU 
temperature of 42-44�C.  When the R40 is used on a laptop cooler, 
<http://meritline.stores.yahoo.net/laptop-cooler-3-fans-4-usb-port.html>, the 
laptop fan doesn't run at all.

> It
> never starts up again. Starting again is not necessary until the faulty
> switch off because temperature is normal until that time.
> Behavior of the fan after the faulty switch off is ominous, though. 
> 
How is it ominous?  If I understand correctly the fan doesn't run after the 
shut-down.
> 
> The situation here is as follows: the keyboard bezel is removed.  The
> keyboard itself is attached by being plugged into the appropriate slot. 
> I seem to need the keyboard to switch on the laptop.  
> 
Understandable, most computers require a keyboard in order to boot.  Speaking 
of errors, do you hear any "beeps" when booting?  I believe a single beep 
indicates normal operation and there is a code for multiple beeps that can 
give a clue as to an error.

> So the part between hinges and keyboard along the whole width of the
> laptop is open.  That's the place for the upper part of the fan and for
> the subcard.  This part of the fan is approximately as warm as is
> measured. 
> 
> After the faulty switch off it gets warm and hot until I switch off the
> laptop manually.  That takes about 2 minutes. 

"It" being the CPU?  Do you use the power switch to manually shut the computer 
down or do you have to remove power?

> During the time between
> the faulty switch off at 33�C and the manual switch off at estimated 
> > 70�C the system tries to start the harddisk all the time.  (I didn't yet
> check if the floppy drive is started, as well.)
> 
That makes me wonder about the BIOS but I supposed that a different BIOS chip 
came with the replacement system board.

> I did neither try using an ice pack for the following reason: it seems
> that the faulty switch off is not caused by too hot the processor.
> 
Well, *something* is causing it to switch off.  It seems that the usual 
suspects are temperature or a bad connection.
> >> 
> > The external adapter supplies a DC voltage to the computer.  The circuitry I 
> > was referring to modifies this voltage and supplies power to the various 
> > components inside the computer.  For instance: five volts to the fan, 1.4 
> > volts to the CPU core, 3.3 volts to the I/O system, etc.
> 
> How can I check if they are functioning? 
> 
I would use an oscilloscope or a digital voltmeter.  You'll want to use 
something with a high input impedance to avoid loading the circuit you're 
trying to examine.  You'll also want some information on what voltage/signal 
can be expected on a given pin.  That might be hard to find as it may be 
proprietary.
> 
> I see two aspects to explore beside others:
> 
> 1. If it is true that the system tries to boot, I wonder how it gets to it.
> Booting seems setting the instruction pointer within the processor to 
> instruction zero. That seems to be a reset. Which events do reset the processor?
> 
I believe that boot-up is controlled by the BIOS but I don't know the details.

> 2. What about the speedstep property of the pentium 3?
> In my intact a21p when booting linux detects different speeds:
> 
> Apr  1 03:01:51 localhost kernel: Detected 697.798 MHz processor.
> Sep 21 14:03:48 localhost kernel: Detected 190.253 MHz processor.
> Sep 22 08:29:33 localhost kernel: Detected 190.777 MHz processor.
> Oct  6 11:06:47 localhost kernel: Detected 488.061 MHz processor.
> 
> with different evaluated BogoMIPS.
> 
> As far as I remember there are no speeds of 190MHz or 488MHz in a 850MHz
> Pentium 3 mobile. 
> 
> Linux changes the speed of the cpu by so called governors.  It seems
> useful for me to check for the governors, and I appreciate hearing
> proposals. 
> 
I don't know enough about Linux to offer a useful comment here.  I could 
invite a Linux guru into the conversation if you wish.  I doubt that he knows 
much about Thinkpads but he seems familiar with most of the Linux 
distributions.

> On the other side I acted on the faulty laptop with two harddisks:
> - one with a new linux installation
> - one with the linux installation of the working a21p
> 
I gather from this that you did some drive swapping.  Did you try the drive 
from the faulty computer in the computer that's working?  That would verify 
that particular drive to be healthy.
-- 
Ernie B.

Communication:  The art of moving an idea from one mind to another, hopefully 
without distortion.

On 2007-10-11, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:
> On Wed, 10 Oct 2007 22:52:16 +0000 (UTC) Tilo Baumann wrote:


>> The fan runs while booting (like all thinkpads afaik) and then stops.
>
> Well, you caused me to look. ;-)  On my R40 the fan _does_not_ run while 
> booting.  Placed on a desk, it runs for about ten seconds after the CPU warms 
> up, then off for a time then cycles on and off as required to maintain a CPU 
> temperature of 42-44�C.  When the R40 is used on a laptop cooler, 

That's what I meant to say.


>> It
>> never starts up again. Starting again is not necessary until the faulty
>> switch off because temperature is normal until that time.
>> Behavior of the fan after the faulty switch off is ominous, though. 
>> 
> How is it ominous?  If I understand correctly the fan doesn't run after the 
> shut-down.

The fan doesn't run after what I call(ed) a faulty switch off. I meant with it

- display off
- ethernet off
- no reaction to keyboard
 
Being in this state 

- the processor gets warmer
- the system tries to 'boot', i.e. permanent shut on and off the
  harddisk. (The floppy drive is not affected.)

I called the behavior of the fan ominous, because the temperature of the
processor gets so hight that the fan should run if the whole system was
working correctly.  


>> The situation here is as follows: the keyboard bezel is removed.  The
>> keyboard itself is attached by being plugged into the appropriate slot. 
>> I seem to need the keyboard to switch on the laptop.  
>> 
> Understandable, most computers require a keyboard in order to boot.  Speaking 

PCs do not need it.

> of errors, do you hear any "beeps" when booting?  I believe a single beep 
> indicates normal operation and there is a code for multiple beeps that can 
> give a clue as to an error.

No beeps

>> So the part between hinges and keyboard along the whole width of the
>> laptop is open.  That's the place for the upper part of the fan and for
>> the subcard.  This part of the fan is approximately as warm as is
>> measured. 
>> 
>> After the faulty switch off it gets warm and hot until I switch off the
>> laptop manually.  That takes about 2 minutes. 
>
> "It" being the CPU?  Do you use the power switch to manually shut the computer 
> down or do you have to remove power?

'It' is the upper part of the fan.
I use the power switch.


>> > The external adapter supplies a DC voltage to the computer.  The circuitry I 
>> > was referring to modifies this voltage and supplies power to the various 
>> > components inside the computer.  For instance: five volts to the fan, 1.4 
>> > volts to the CPU core, 3.3 volts to the I/O system, etc.
>> 
>> How can I check if they are functioning? 
>> 
> I would use an oscilloscope or a digital voltmeter.  You'll want to use 
> something with a high input impedance to avoid loading the circuit you're 
> trying to examine.  You'll also want some information on what voltage/signal 
> can be expected on a given pin.  That might be hard to find as it may be 
> proprietary.

I am not able to manage those devices.


>> [ some linux things ]
 
> I don't know enough about Linux to offer a useful comment here.  I could 
> invite a Linux guru into the conversation if you wish.  I doubt that he knows 
> much about Thinkpads but he seems familiar with most of the Linux 
> distributions.

See below. I first need someone to know the a21p better than me.


>> On the other side I acted on the faulty laptop with two harddisks:
>> - one with a new linux installation
>> - one with the linux installation of the working a21p
>> 
> I gather from this that you did some drive swapping.  Did you try the drive 

Yes

> from the faulty computer in the computer that's working?  That would verify 
> that particular drive to be healthy.

Yes


BTW: I made new measurements of the development of temperature. 
time it took about 6 minutes to get from 28�C to 51�C when it switched
off.  (The loop to check the temperature was run in shorter distance.)
I.e.  those numbers are not at all constant. 

When booting into DOS the system switched off today after 22 minutes.
The fan got hot but did not run.

When getting into the BIOS Setup Utility (F1) the system switched off as
well. 

Questions from the experiences today:

- is the subcard responsible for running the fan?
  I think the subcard must get the signal for running the fan from the
  processor, whose temperature is controled. If the fan is only running
  at boot time (for ten seconds) it must not be the subcard, must it not?
  
  (What about your r40?)
  Note: the fan itself is intact as it runs when booting (for some seconds).
 
- is it true that the fan in the a21p is running all the time with low speed?
  (In my other a21p the temperature now is 38�C and the fan is running, perhaps
  all the time. (I will pay attention to it.)
  I know that it can run in different speeds.
  What about your r40 in this regard?

What do you think about my buying another subcard?


Tilo

On 2007-10-11, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:
> On Wed, 10 Oct 2007 22:52:16 +0000 (UTC) Tilo Baumann wrote:


>> The fan runs while booting (like all thinkpads afaik) and then stops.
>
> Well, you caused me to look. ;-)  On my R40 the fan _does_not_ run while 
> booting.  Placed on a desk, it runs for about ten seconds after the CPU warms 
> up, then off for a time then cycles on and off as required to maintain a CPU 
> temperature of 42-44�C.  When the R40 is used on a laptop cooler, 

That's what I meant to say.


>> It
>> never starts up again. Starting again is not necessary until the faulty
>> switch off because temperature is normal until that time.
>> Behavior of the fan after the faulty switch off is ominous, though. 
>> 
> How is it ominous?  If I understand correctly the fan doesn't run after the 
> shut-down.

The fan doesn't run after what I call(ed) a faulty switch off. I meant with it

- display off
- ethernet off
- no reaction to keyboard
 
Being in this state 

- the processor gets warmer
- the system tries to 'boot', i.e. permanent shut on and off the
  harddisk. (The floppy drive is not affected.)

I called the behavior of the fan ominous, because the temperature of the
processor gets so hight that the fan should run if the whole system was
working correctly.  


>> The situation here is as follows: the keyboard bezel is removed.  The
>> keyboard itself is attached by being plugged into the appropriate slot. 
>> I seem to need the keyboard to switch on the laptop.  
>> 
> Understandable, most computers require a keyboard in order to boot.  Speaking 

PCs do not need it.

> of errors, do you hear any "beeps" when booting?  I believe a single beep 
> indicates normal operation and there is a code for multiple beeps that can 
> give a clue as to an error.

No beeps

>> So the part between hinges and keyboard along the whole width of the
>> laptop is open.  That's the place for the upper part of the fan and for
>> the subcard.  This part of the fan is approximately as warm as is
>> measured. 
>> 
>> After the faulty switch off it gets warm and hot until I switch off the
>> laptop manually.  That takes about 2 minutes. 
>
> "It" being the CPU?  Do you use the power switch to manually shut the computer 
> down or do you have to remove power?

'It' is the upper part of the fan.
I use the power switch.


>> > The external adapter supplies a DC voltage to the computer.  The circuitry I 
>> > was referring to modifies this voltage and supplies power to the various 
>> > components inside the computer.  For instance: five volts to the fan, 1.4 
>> > volts to the CPU core, 3.3 volts to the I/O system, etc.
>> 
>> How can I check if they are functioning? 
>> 
> I would use an oscilloscope or a digital voltmeter.  You'll want to use 
> something with a high input impedance to avoid loading the circuit you're 
> trying to examine.  You'll also want some information on what voltage/signal 
> can be expected on a given pin.  That might be hard to find as it may be 
> proprietary.

I am not able to manage those devices.


>> [ some linux things ]
 
> I don't know enough about Linux to offer a useful comment here.  I could 
> invite a Linux guru into the conversation if you wish.  I doubt that he knows 
> much about Thinkpads but he seems familiar with most of the Linux 
> distributions.

See below. I first need someone to know the a21p better than me.


>> On the other side I acted on the faulty laptop with two harddisks:
>> - one with a new linux installation
>> - one with the linux installation of the working a21p
>> 
> I gather from this that you did some drive swapping.  Did you try the drive 

Yes

> from the faulty computer in the computer that's working?  That would verify 
> that particular drive to be healthy.

Yes


BTW: I made new measurements of the development of temperature.  This
time it took about 6 minutes to get from 28�C to 51�C when it switched
off.  I.e.  those temperatures are not at all constant. 

When booting into DOS the system switched off today after 22 minutes.
The fan got hot but did not run.

When getting into the BIOS Setup Utility (F1) the system switched off as
well. 

Questions from the experiences today:

- is the subcard responsible for running the fan? I think the subcard
must get the signal for running the fan from the processor, whose
temperature is controled.  If it is the case that the fan is no more
running after boot time the reason must not be the subcard, must it not?
(What about your r40?) 

- What do you think about my buying another subcard?


Tilo

On Thu, 11 Oct 2007 20:39:22 +0000 (UTC) Tilo Baumann wrote:

> >> The fan runs while booting (like all thinkpads afaik) and then stops.
> >
> > Well, you caused me to look. ;-)  On my R40 the fan _does_not_ run while 
> > booting.  Placed on a desk, it runs for about ten seconds after the CPU warms 
> > up, then off for a time then cycles on and off as required to maintain a CPU 
> > temperature of 42-44�C.  When the R40 is used on a laptop cooler, 
> 
> That's what I meant to say.
> 
Now I'm confused.  Does the fan run when you first apply power to the computer 
or does it not run?
> >> 
> > How is it ominous?  If I understand correctly the fan doesn't run after the 
> > shut-down.
> 
> The fan doesn't run after what I call(ed) a faulty switch off. I meant with it
> 
> - display off
> - ethernet off
> - no reaction to keyboard
>  
> Being in this state 
> 
> - the processor gets warmer

.... Which means that the system isn't shutting down completely, i.e.: power is 
still being supplied to the CPU.

> - the system tries to 'boot', i.e. permanent shut on and off the
>   harddisk. (The floppy drive is not affected.)
> 
The floppy drive not trying to boot could be a product of the BIOS boot order.  
Most BIOS can be adjusted so that the computer tries to boot from the floppy 
first and then from the hard drive.

> I called the behavior of the fan ominous, because the temperature of the
> processor gets so hight that the fan should run if the whole system was
> working correctly.  
> 
Okay.
> >> 
> > Understandable, most computers require a keyboard in order to boot.  Speaking 
> 
> PCs do not need it.
> 
Some do some don't, it depends on the BIOS.  I've seen PCs that beep like mad 
if you try to start them with the keyboard disconnected.

> > of errors, do you hear any "beeps" when booting?  I believe a single beep 
> > indicates normal operation and there is a code for multiple beeps that can 
> > give a clue as to an error.
> 
> No beeps
> 
Grrrrr, not so good.  See "IBM BIOS beep codes"
<http://www.computerhope.com/beep.htm>, about halfway down the page.  
Additional information is at <http://www.pchell.com/hardware/beepcodes.shtml>.

Do you still have the mini pci card removed?
> >
> > "It" being the CPU?  Do you use the power switch to manually shut the computer 
> > down or do you have to remove power?
> 
> 'It' is the upper part of the fan.
> I use the power switch.
> 
Okay.
> >> 
> > I would use an oscilloscope or a digital voltmeter.  You'll want to use 
> > something with a high input impedance to avoid loading the circuit you're 
> > trying to examine.  You'll also want some information on what voltage/signal 
> > can be expected on a given pin.  That might be hard to find as it may be 
> > proprietary.
> 
> I am not able to manage those devices.
> 
Too bad.  Good test equipment can save a lot of headaches.
> 
> See below. I first need someone to know the a21p better than me.
> 
Okay.  I can drop him a note whenever you get ready.
> 
> > from the faulty computer in the computer that's working?  That would verify 
> > that particular drive to be healthy.
> 
> Yes
> 
Okay.  I wasn't sure.
> 
> BTW: I made new measurements of the development of temperature. 
> time it took about 6 minutes to get from 28�C to 51�C when it switched
> off.  (The loop to check the temperature was run in shorter distance.)
> I.e.  those numbers are not at all constant. 
> 
I can see that.  I still believe that something is reacting badly to 
temperature though, the trick is to identify it/them.

> When booting into DOS the system switched off today after 22 minutes.
> The fan got hot but did not run.
> 
> When getting into the BIOS Setup Utility (F1) the system switched off as
> well. 
> 
Were you able to verify that the BIOS settings are correct?

> Questions from the experiences today:
> 
> - is the subcard responsible for running the fan?
>   I think the subcard must get the signal for running the fan from the
>   processor, whose temperature is controled. If the fan is only running
>   at boot time (for ten seconds) it must not be the subcard, must it not?
>   
I'm still confused about exactly when the fan runs.

>   (What about your r40?)
>   Note: the fan itself is intact as it runs when booting (for some seconds).
>  
As I said, the fan on the R40 does not run on boot-up.

> - is it true that the fan in the a21p is running all the time with low speed?
>   (In my other a21p the temperature now is 38�C and the fan is running, perhaps
>   all the time. (I will pay attention to it.)
>   I know that it can run in different speeds.
>   What about your r40 in this regard?
> 
The fan assembly for the R40 looks like <http://tinyurl.com/2z6auf>.  The 
cable contains three wires which plug into a sub-board.  I can *speculate* 
that the part that senses the CPU temperature, on the right in the photo, 
*may* contain a thermistor which could control the fan speed but I haven't 
seen a schematic of the circuit.

> What do you think about my buying another subcard?
> 
That would be for you to decide.

<sigh>  I wish that you could borrow or rent some test equipment.  At the 
moment you are exchanging FRUs, saying "No, that didn't fix it" and trying 
something else.  You may finish with enough parts to build another computer, 
you'll certainly never lack for spares.
-- 
Ernie B.

Communication:  The art of moving an idea from one mind to another, hopefully 
without distortion.

On 2007-10-11, Ernie B <ebaresch_REMOVE_@cox._THIS_net> wrote:
> On Thu, 11 Oct 2007 20:39:22 +0000 (UTC) Tilo Baumann wrote:

>> > Well, you caused me to look. ;-)  On my R40 the fan _does_not_ run while 
>> > booting.  Placed on a desk, it runs for about ten seconds after the CPU warms 
>> > up, then off for a time then cycles on and off as required to maintain a CPU 
>> > temperature of 42-44�C.  When the R40 is used on a laptop cooler, 
>> 
>> That's what I meant to say.
>> 
> Now I'm confused.  Does the fan run when you first apply power to the computer 
> or does it not run?

BTW: do you notice a difference when placing the r40 on the floor e.g.?

Let me try to describe what mine does: 

Switching on the laptop the fan starts to run.  It runs for about ten
seconds.  (That is it runs about ten seconds during the laptop is
booting, but not during the whole time of booting.  The fan never
runs exept above said.
 
(To be more precisely: if you remove the harddisk, there is a timeout
for the BIOS to wait for it, which is more than 10 seconds. During this time
the fan is running.)


>> > Understandable, most computers require a keyboard in order to boot.  Speaking 
>> 
>> PCs do not need it.
>> 
> Some do some don't, it depends on the BIOS.  I've seen PCs that beep like mad 
> if you try to start them with the keyboard disconnected.

I know BIOS that can be instructed to not bother if the keyboard is missing.


>> > of errors, do you hear any "beeps" when booting?  I believe a single beep 
>> > indicates normal operation and there is a code for multiple beeps that can 
>> > give a clue as to an error.
>> 
>> No beeps
>> 
> Grrrrr, not so good.  See "IBM BIOS beep codes"
><http://www.computerhope.com/beep.htm>, about halfway down the page.  
> Additional information is at <http://www.pchell.com/hardware/beepcodes.shtml>.

I will do.

> Do you still have the mini pci card removed?

No, it is installed. I tested it before, it seems to be okay.

>> > I would use an oscilloscope or a digital voltmeter.  You'll want to use 
>> > something with a high input impedance to avoid loading the circuit you're 
>> > trying to examine.  You'll also want some information on what voltage/signal 
>> > can be expected on a given pin.  That might be hard to find as it may be 
>> > proprietary.
>> 
>> I am not able to manage those devices.
>> 
> Too bad.  Good test equipment can save a lot of headaches.

I envy you for your equipment and knowlidge.


> Were you able to verify that the BIOS settings are correct?

They are more or less standard. I would have had time to check them all.

>> Questions from the experiences today:
>> 
>> - is the subcard responsible for running the fan?
>>   I think the subcard must get the signal for running the fan from the
>>   processor, whose temperature is controled. If the fan is only running
>>   at boot time (for ten seconds) it must not be the subcard, must it not?
>>   
> I'm still confused about exactly when the fan runs.

Was I able to remedy that?


>> - is it true that the fan in the a21p is running all the time with low speed?

In the meantime I doubt the continuos running of the fan. It seems that I
confused it with the harddisk. 


>> What do you think about my buying another subcard?
>> 
> That would be for you to decide.

My question did not mean to deliver my decision to you.
A subcard is bought.

><sigh>  I wish that you could borrow or rent some test equipment.  At the 
> moment you are exchanging FRUs, saying "No, that didn't fix it" and trying 
> something else.  You may finish with enough parts to build another computer, 
> you'll certainly never lack for spares.

I bought a mainboard, a cpu and a subcard.  The mainboard was necessary
anyway as the BIOS checksum was no more correct (due to my testing the
mainboard). 


Tilo

On Fri, 12 Oct 2007 21:13:58 +0000 (UTC) Tilo Baumann wrote:

> BTW: do you notice a difference when placing the r40 on the floor e.g.?
> 
I'm more comfortable using the TP at a desk or table.  Floors tend to be 
dusty.

> Let me try to describe what mine does: 
> 
> Switching on the laptop the fan starts to run.  It runs for about ten
> seconds.  (That is it runs about ten seconds during the laptop is
> booting, but not during the whole time of booting.  The fan never
> runs exept above said.
>  
> (To be more precisely: if you remove the harddisk, there is a timeout
> for the BIOS to wait for it, which is more than 10 seconds. During this time
> the fan is running.)
> 
Okay.
> >> 
> > Grrrrr, not so good.  See "IBM BIOS beep codes"
> ><http://www.computerhope.com/beep.htm>, about halfway down the page.  
> > Additional information is at <http://www.pchell.com/hardware/beepcodes.shtml>.
> 
> I will do.
> 
Okay.

> > Do you still have the mini pci card removed?
> 
> No, it is installed. I tested it before, it seems to be okay.
> 
Okay.
> >> 
> > Too bad.  Good test equipment can save a lot of headaches.
> 
> I envy you for your equipment and knowlidge.
> 
I don't have much knowledge, am always trying to learn.  The test equipment 
was acquired used, from a pawn shop.
> 
> > Were you able to verify that the BIOS settings are correct?
> 
> They are more or less standard. I would have had time to check them all.
> 
Okay.  Is it possible for you to select the default BIOS settings?

> >> Questions from the experiences today:
> >> 
> >> - is the subcard responsible for running the fan?
> >>   I think the subcard must get the signal for running the fan from the
> >>   processor, whose temperature is controled. If the fan is only running
> >>   at boot time (for ten seconds) it must not be the subcard, must it not?
> >>   
> > I'm still confused about exactly when the fan runs.
> 
> Was I able to remedy that?
> 
Yes.
> 
> My question did not mean to deliver my decision to you.
> A subcard is bought.
> 
> I bought a mainboard, a cpu and a subcard.  The mainboard was necessary
> anyway as the BIOS checksum was no more correct (due to my testing the
> mainboard). 
> 
Okay.

I've run out of suggestions, good luck in fixing your computer.
-- 
Ernie B.

Communication:  The art of moving an idea from one mind to another, hopefully 
without distortion.

Tilo Baumann wrote:
> 
>
> 
> See below. I first need someone to know the a21p better than me.
> 
> >> On the other side I acted on the faulty laptop with two harddisks:
> >> - one with a new linux installation
> >> - one with the linux installation of the working a21p
> >>
> > I gather from this that you did some drive swapping.  Did you try the drive
> 
> Yes
> 
> > from the faulty computer in the computer that's working?  That would verify
> > that particular drive to be healthy.
> 
> Yes
> 
> BTW: I made new measurements of the development of temperature.
> time it took about 6 minutes to get from 28�C to 51�C when it switched
> off.  (The loop to check the temperature was run in shorter distance.)
> I.e.  those numbers are not at all constant.
> 
> When booting into DOS the system switched off today after 22 minutes.
> The fan got hot but did not run.
> 
> When getting into the BIOS Setup Utility (F1) the system switched off as
> well.
> 
> Questions from the experiences today:
> 
> - is the subcard responsible for running the fan?
>   I think the subcard must get the signal for running the fan from the
>   processor, whose temperature is controled. If the fan is only running
>   at boot time (for ten seconds) it must not be the subcard, must it not?
> 
>   (What about your r40?)
>   Note: the fan itself is intact as it runs when booting (for some seconds).
> 
> - is it true that the fan in the a21p is running all the time with low speed?
>   (In my other a21p the temperature now is 38�C and the fan is running, perhaps
>   all the time. (I will pay attention to it.)
>   I know that it can run in different speeds.
>   What about your r40 in this regard?
> 
> What do you think about my buying another subcard?
> 
> Tilo

Pointless. Now that you have mentioned that the system even locks up
when you are in the BIOS it's not an operating system, fan or sub card
problem - you have a defective main board.

I just went through this with an A21m 2628-PRU system. The random
hardware lockup "feature" happens to a lot of A2x and T2x series main
boards. If/when you can get into the BIOS setup without it locking up
try disabling all power management features in the BIOS. See if that
improves or changes the behavior. (It won't fix it, but it will confirm
it.) Just google "dead IBM Thinkpad" and see what you come up with.

You should be able to find a reconditioned main board with a non-DOA
warranty for about $150.00 through many dealers. You don't have to take
a chance buying a used main board on eBay. Get the *correct* part number
from the HMM - the A21p has an UXGA display and you need the correct
main board with the correct video system to drive that display.

Rick

On 2007-10-13, Rick <rickajho@rcn.com> wrote:
> Tilo Baumann wrote:

>> What do you think about my buying another subcard?

> Pointless. Now that you have mentioned that the system even locks up
> when you are in the BIOS it's not an operating system, fan or sub card
> problem - you have a defective main board.

Why?
The processor warms up showing the BIOS Setup, too. Therefore cooling in
this state ist needed, too.


> I just went through this with an A21m 2628-PRU system. The random
> hardware lockup "feature" happens to a lot of A2x and T2x series main
> boards. If/when you can get into the BIOS setup without it locking up
> try disabling all power management features in the BIOS. See if that
> improves or changes the behavior. (It won't fix it, but it will confirm
> it.) Just google "dead IBM Thinkpad" and see what you come up with.
>
> You should be able to find a reconditioned main board with a non-DOA
> warranty for about $150.00 through many dealers. You don't have to take
> a chance buying a used main board on eBay. Get the *correct* part number
> from the HMM - the A21p has an UXGA display and you need the correct
> main board with the correct video system to drive that display.

I will check that later.


Tilo

Tilo Baumann wrote:
> 
> On 2007-10-13, Rick <rickajho@rcn.com> wrote:
> > Tilo Baumann wrote:
> 
> >> What do you think about my buying another subcard?
> 
> > Pointless. Now that you have mentioned that the system even locks up
> > when you are in the BIOS it's not an operating system, fan or sub card
> > problem - you have a defective main board.
> 
> Why?
> The processor warms up showing the BIOS Setup, too. Therefore cooling in
> this state ist needed, too.

It has nothing to do with it. You can power on one of these defective
main board systems and have it hang in under two seconds - even before
you can press F1 and get into the BIOS screen. That's not enough time to
overheat a processor. We can use any premise to speculate about the
problem but IBM/Lenovo has never issued any statement about the nature
of the failure(s). If you talk to IBM tech support they will readily
admit there are known main board failure problems on certain model lines
and if you describe your problem they will jump right on "defective main
board" as the cause. But IBM won't give you any information about what
*they* know about the nature of the failure.

The non-IBM speculation about the cause of the failures has come down to
either 1.) Hairline stress cracks/failed solder points on the main
boards caused by picking up the lap tops from the corners. (That
actually was a known warranty repair problem for IBM on certain model
lines.) 2.) Component failure in the power management circuitry itself.

Regardless the only thing that fixes it is a main board replacement once
you have done all the obvious: swapped out the mini PCI card, swapped
memory, pulled the hard drive and anything in the UltraBay. (I had all
of that done at an IBM repair facility.) If it still hangs while running
diagnostics it's the main board. I've had the defective A21m main board
run the DOS booted PC Doctor diagnostics for over 36 hours before it
locked up. I've also had it hang in under two minutes running the same
tests. I've also had it hang before it could complete booting the
diagnostic tests. There is no obvious or repairable cause/effect
component you can fix here.

You will probably find that when you replace the board it will have a
"revision number" on it that isn't on your original. In my case it has a
"Rev. C" sticker on the replacement board after the IBM part number.
That alone would tell me these main boards have been modified at least
twice for some reason(s).

> > I just went through this with an A21m 2628-PRU system. The random
> > hardware lockup "feature" happens to a lot of A2x and T2x series main
> > boards. If/when you can get into the BIOS setup without it locking up
> > try disabling all power management features in the BIOS. See if that
> > improves or changes the behavior. (It won't fix it, but it will confirm
> > it.) Just google "dead IBM Thinkpad" and see what you come up with.
> >
> > You should be able to find a reconditioned main board with a non-DOA
> > warranty for about $150.00 through many dealers. You don't have to take
> > a chance buying a used main board on eBay. Get the *correct* part number
> > from the HMM - the A21p has an UXGA display and you need the correct
> > main board with the correct video system to drive that display.
> 
> I will check that later.
> 
> Tilo

I'm just saying that I went through months of screwing around with a
"some times it works for days, some times it works for minutes" A21m. It
doesn't help that you cannot reproduce the set a circumstances that will
cause a system with these main board problems to fail. Like I said: it
might work for days, hours, or minutes before it will lock up. But it
will happen. I was running Win 2000 so I got a different scenario when
the system locked if the OS was loaded. But if yours is locking up when
it's booting and/or when you are just in the BIOS settings screens -
it's the same hardware problem.

Rick

On 2007-10-14, Rick <rickajho@rcn.com> wrote:
> Tilo Baumann wrote:
>> 
Rick <rickajho@rcn.com> wrote:
>> > Tilo Baumann wrote:
>> 
>> >> What do you think about my buying another subcard?
>> 
>> > Pointless. Now that you have mentioned that the system even locks up
>> > when you are in the BIOS it's not an operating system, fan or sub card
>> > problem - you have a defective main board.
>> 
>> Why?
>> The processor warms up showing the BIOS Setup, too. Therefore cooling in
>> this state ist needed, too.
>
> It has nothing to do with it. You can power on one of these defective
> main board systems and have it hang in under two seconds - even before
> you can press F1 and get into the BIOS screen. That's not enough time to
> overheat a processor. We can use any premise to speculate about the
> problem but IBM/Lenovo has never issued any statement about the nature
> of the failure(s). If you talk to IBM tech support they will readily
> admit there are known main board failure problems on certain model lines
> and if you describe your problem they will jump right on "defective main
> board" as the cause. But IBM won't give you any information about what
> *they* know about the nature of the failure.
>
> The non-IBM speculation about the cause of the failures has come down to
> either 1.) Hairline stress cracks/failed solder points on the main
> boards caused by picking up the lap tops from the corners. (That
> actually was a known warranty repair problem for IBM on certain model
> lines.) 2.) Component failure in the power management circuitry itself.
>
> Regardless the only thing that fixes it is a main board replacement once
> you have done all the obvious: swapped out the mini PCI card, swapped
> memory, pulled the hard drive and anything in the UltraBay. (I had all
> of that done at an IBM repair facility.) If it still hangs while running
> diagnostics it's the main board. I've had the defective A21m main board
> run the DOS booted PC Doctor diagnostics for over 36 hours before it
> locked up. I've also had it hang in under two minutes running the same
> tests. I've also had it hang before it could complete booting the
> diagnostic tests. There is no obvious or repairable cause/effect
> component you can fix here.

Did you notice that I changed the mainboard (by a used one, though)?
The seller told me that it was taken from a full functioning laptop.
 

> You will probably find that when you replace the board it will have a
> "revision number" on it that isn't on your original. In my case it has a
> "Rev. C" sticker on the replacement board after the IBM part number.
> That alone would tell me these main boards have been modified at least
> twice for some reason(s).

My old board has no inication of a revision. Whith the 'new' one I don't
know because it is built in and I didn't pay attention when I got it.
 
The rest of your posting will be answered later.


Tilo 

Tilo Baumann wrote:
> 
> On 2007-10-14, Rick <rickajho@rcn.com> wrote:
> > Tilo Baumann wrote:
> >>
> Rick <rickajho@rcn.com> wrote:
> >> > Tilo Baumann wrote:
> >>
> >> >> What do you think about my buying another subcard?
> >>
> >> > Pointless. Now that you have mentioned that the system even locks up
> >> > when you are in the BIOS it's not an operating system, fan or sub card
> >> > problem - you have a defective main board.
> >>
> >> Why?
> >> The processor warms up showing the BIOS Setup, too. Therefore cooling in
> >> this state ist needed, too.
> >
> > It has nothing to do with it. You can power on one of these defective
> > main board systems and have it hang in under two seconds - even before
> > you can press F1 and get into the BIOS screen. That's not enough time to
> > overheat a processor. We can use any premise to speculate about the
> > problem but IBM/Lenovo has never issued any statement about the nature
> > of the failure(s). If you talk to IBM tech support they will readily
> > admit there are known main board failure problems on certain model lines
> > and if you describe your problem they will jump right on "defective main
> > board" as the cause. But IBM won't give you any information about what
> > *they* know about the nature of the failure.
> >
> > The non-IBM speculation about the cause of the failures has come down to
> > either 1.) Hairline stress cracks/failed solder points on the main
> > boards caused by picking up the lap tops from the corners. (That
> > actually was a known warranty repair problem for IBM on certain model
> > lines.) 2.) Component failure in the power management circuitry itself.
> >
> > Regardless the only thing that fixes it is a main board replacement once
> > you have done all the obvious: swapped out the mini PCI card, swapped
> > memory, pulled the hard drive and anything in the UltraBay. (I had all
> > of that done at an IBM repair facility.) If it still hangs while running
> > diagnostics it's the main board. I've had the defective A21m main board
> > run the DOS booted PC Doctor diagnostics for over 36 hours before it
> > locked up. I've also had it hang in under two minutes running the same
> > tests. I've also had it hang before it could complete booting the
> > diagnostic tests. There is no obvious or repairable cause/effect
> > component you can fix here.
> 
> Did you notice that I changed the mainboard (by a used one, though)?
> The seller told me that it was taken from a full functioning laptop.

And they took it out of a fully functioning lap top just for shits and
giggles? Good luck...

At least if you go through a known IBM parts reseller (I.e.: NOT "I sell
crap 'as is' on eBay but tell you whatever you want to hear.") you can
get a non DOA guarantee at minimum, a 90 defect warranty on a
refurbished part at best.

> > You will probably find that when you replace the board it will have a
> > "revision number" on it that isn't on your original. In my case it has a
> > "Rev. C" sticker on the replacement board after the IBM part number.
> > That alone would tell me these main boards have been modified at least
> > twice for some reason(s).
> 
> My old board has no inication of a revision. Whith the 'new' one I don't
> know because it is built in and I didn't pay attention when I got it.
> 
> The rest of your posting will be answered later.
> 
> Tilo

On 2007-10-21, Rick <rickajho@rcn.com> wrote:
> Tilo Baumann wrote:

>> Did you notice that I changed the mainboard (by a used one, though)?
>> The seller told me that it was taken from a full functioning laptop.
>
> And they took it out of a fully functioning lap top just for shits and
> giggles? Good luck...

It was a bit more expensive than that. (But I understand what you wanted
to say.)


Tilo

On 2007-10-14, Rick <rickajho@rcn.com> wrote:

>> > I just went through this with an A21m 2628-PRU system. The random
>> > hardware lockup "feature" happens to a lot of A2x and T2x series main
>> > boards. If/when you can get into the BIOS setup without it locking up
>> > try disabling all power management features in the BIOS. See if that
>> > improves or changes the behavior. (It won't fix it, but it will confirm
>> > it.) Just google "dead IBM Thinkpad" and see what you come up with.

Going into BIOS/Config/Power I find some items which seem to be related
to power management. 

I switched 

	Power Mode for AC		to Max Performance
	Power Mode for Battery		to Max Performance 
	Hibernate by timer		to Disabled
	Low Battery Action 		to Suspend
	the others			to Disabled

I did not test that for a long time and could not find a difference so far.

I started the laptop without any RAM: it took 15 minutes for the laptop to
switch off. Display was not on during this time, the fan was running all
the time. 

>> > You should be able to find a reconditioned main board with a non-DOA
>> > warranty for about $150.00 through many dealers. You don't have to take
>> > a chance buying a used main board on eBay. Get the *correct* part number
>> > from the HMM - the A21p has an UXGA display and you need the correct
>> > main board with the correct video system to drive that display.

That will not be worthwhile, I suppose.

The subcard is/was bought at ebay before you condemned it, the
mainboard as well. If I had known earlier ....

I will test the subcard and am likely to give up, if that doesn't help.


Google offered 800.000 links to 'dead IBM thinkpad'. I am reading a bit
there, as well as in 2 thinkpad groups/forums.

Thanks so far


Tilo  

On 2007-10-21, Tilo Baumann <2007@eene.mine.nu> wrote:
> On 2007-10-14, Rick <rickajho@rcn.com> wrote:

> I will test the subcard and am likely to give up, if that doesn't help.

I got the subcard. There was nothing new with it, same failure. Please give me
further recommendations!


Tilo