T41 repair story

[SMA]

I bought a defect T41 locally. It came with no hard drive, but otherwise all parts were there, including a port replicator.

It was said to be completely dead. I was hoping for a quick repair - like a blown F2 fuse.
But that hope disappeared as soon as I plugged it in, because the battery started charging.

The problem became obvious when I lifted the keyboard. Some liquid had been spilt on it.
Seemed like a completely lost board, but never the less I went on with it.

Here are some links to a couple of pictures of what I found

Picture 1
Picture 2
Picture 3
Picture 4

First I took the board out and cleaned it using dematerialized water and after some days of drying it was time to measure voltages.

The test setup was just the bare board. No cpu and no ram. Sometimes, and sometimes not, attached to the port replicator.
The reason for choosing this setup was that I knew the CPU would not turn on, because, when I earlier had tried the machine,
I had it hooked up to a laboratory power supply with a built-in ampere meter and the current consumption was very low.
And I would need the port replicator to enable me to press the power button.

I was a little in doubt about whether I would have to measure voltages before or after pressing the power button, but I
concluded that I would have to do both.

Since I was already familiar with the MAX1631 chip and it was located in the area of the board where the spill had been, I decided to let that
chip be the first one to measure upon. Here is a link to the voltages I measured on it.

MAX1631 Measures

It is a table with one row for each pin on the chip and 4 columns.
The first 2 columns are the T41 in the 'off' and 'on' state and the next 2 columns are the same, but measured on another
machine, a T40. This T40 has the loose GPU error, but it has no influence here. The purpose of including the measures from the T40 is
that it is easier to evaluate the measured voltages, when one knows what they are supposed to be.

I found that there were three MAX1845 chips on the board. They became my next targets.
Two of them did behave like one would expect, but on the third one (U51) I found something strange. Here is a link to the measure sheet.

MAX1845 Measures

Can you see the problem for this chip?

If you are interested in trying out your own bug-hunting skills, then now is the time for you to look carefully on the measure sheets and make your
own evaluation - before reading on. If you do not know these chips, you may want to download their datasheets.

So far, I had measured voltages on only 4 chips and I had already found 2 problem indications.
The MAX1631 ON1/ON5 wasn't there in the 'off' state and the U51 MAX1845 did not get its VCC. It should have 5V on pin 21.
The two other MAX1845 chips did have 5V on their VCC pins in the 'on' state.

At this point I did not know how these chips were interconnected, so I had to follow some traces on the board and I ended with this.

Diagram

I quickly found that the MAX1630 ON3/ON5 was connected to the TB62501F chip, but since I was unable to find a datasheet for that chip it seemed to be a dead end.
I had to follow the second error indication - the missing 5V on the U51 MAX1845 chip.
I discovered that two of the three MAX1845 chips are getting their VCC from the MAX1631 OUT5, but the last one (U51) was not connected to the same 5V.
The missing 5V did not come from the MAX1631 '5V ALWAYS' and I could not find another chip making 5V.

For a long time that remained as a mystery. I could not figure out where that 5V was supposed to come from, but finally I found the 3B transistor and
its connections and I concluded that pin 25 on the TB62501F would have to be a sort of VCC pin for the TB62501F.
That kind of closed the ring.

The U51 MAX1845 side 1 is supposed to turn on before the MAX1630 since the U51 MAX1845 side 1 output enables the 3b transistor to pass the System DC on
to the TB62501F so that it can make the ON3/ON5 signal for the MAX1630.

Since the U51 MAX1845 was to turn on before the MAX1631, the missing 5V could not come from the MAX1631 OUT5 and, as already said, there was no other
chips making 5V. The missing 5V would have to come from the MAX1631 '5V ALWAYS'. But again - there was no connection there.
I decided to make a control measure on the T40 and found that that machine actually had such a connection. So my T41 had a broken trace.
From then on, things started to move rather quickly.

I rolled out some 10 cm of a thin orange wire between pin 21 of the MAX1631 and pin 21 of the U51 MAX1845.
Then, attached the power and found that the board's current consumption in the 'off' state was now normal.
I installed the board in the chassis. With cpu and ram and up came the logo.

But since there had been some liquid spill on this board, I knew that there could still be more errors lurking around, so I had to test everything.
I had to replace the CMOS battery as it was completely dead and using PC Doctor I found that the 'M' key had stopped working, so the keyboard was replaced too,
but besides from this I did not find any other errors.
The XP installation went well and after having installed all needed drivers, there were no yellow exclamation marks left in Device Manager.

The last thing to do was to inspect the main battery information. I was expecting a rather old and dead battery, but it wasn't so. It had a cycle count of 1.

Note that I did not replace any of the components shown on the pictures at the beginning of this tread. I just cleaned them.

The model number of this T41 is 2373-4GG and it now has 1.5 GB of memory and a 160 GB hard drive.

I know that this story has been rather technical, but I hope that some of you will get something out of reading it.

[sjthinkpader]

Great for you, excellent work.

[sparta.rising]

Great story, I wish I had the patients for that kind of detective work.

[Temetka]

Excellent work you did.

[Beaver]

Still wondering why there was a problem with all these spill proof T series keyboards and drain holes

[rkawakami]

@SMA: Excellent debugging! It might give me some insight into a couple of "dead" T41 systems that I still have.

@Beaver: Well, the best advice that I can give concerning T4x systems and spills is this:

- don't tilt the system around in an attempt to shake the liquid out. Let the drainage canals and hole do their job while the laptop is kept level.
- obviously, remove all power from the system as soon as possible. This means unplugging the AC adapter AND removing the battery.
- put the system on top of a cloth so that it soaks up any exiting liquid.
- if the liquid is particularly sticky (soda, highly sugared coffee, etc.), resign yourself to the fact that you'll probably end up having to replace the keyboard. I understand that people have had success in bathing the keyboard in demineralized (i.e., reverse osmosis) water and/or alcohol but you may not be so lucky.
- after a day or so of letting the liquid drain out, carefully remove the keyboard (keeping the system level if possible) and inspect the motherboard for signs of liquid intrusion.

[poshgeordie]

Interested to know if anyone has tried washing their keyboard in the dishwasher.

There are plenty of posts around about people who have done it, and in theory it should work even if it does take a while to dry out.

Since the key tops tend to come off quite easily, putting it in one of those mesh bags, used for small items of clothes in the washing machine, might help to keep it in one piece.

I've got one with a couple of sticky keys which I might try and report back.

And SMA - Many thanks for the hard work - very useful

[thomaslp]

I had a mobo that could not charge the battery so I decided to take a shot and change the adp3806 chip. I understand this chip could be the problem. The laptop powered on and it seemed that the battery started charging....but during booting into windows the laptop shut down....and has been dead since. I checked my soldering.....and learned always check the soldering carefully


I hoped that I only had fried the adp3806 chip so today I soldered a new one onto the board....and checked it carefylly - but the mobo is still totally dead. I checked the fuses but they all seem to be fine so now I will give the 1631 and 1845 a shot.

I am not very familiar with a voltmeter and this kind of debugging. Should I always place the minus (black) on leg one and plus (red) on leg X on the chip?

Also you write you check before and after power on. Does this mean that you plug in the psu and then makes the measurement. Then you simply push the power button once and make the measurement again?

[SMA]

There is always a risk when replacing chips on boards like these.
They are multilayer boards and the distance from the top layer and down to the next layer is less than 1 mm, I believe.
It is rather easy to apply a little too much heat and by doing that, burning a hole in the board, which may cause some traces in different layers to short.
Also, too much heat may cause one or more traces to come loose.

Remember to orientate the new chip correctly.


Place the black probe on ground when measuring voltages. There are many places on these boards where one can find the ground plane to hook on to.
Ground can be found on the cobber islands surrounding the screw holes like this, on the flaps holding the memory modules in place like this,
on the battery connecter like this and a lot of other places.
Making the ground connection a solid one will save you a hand - having only two hands, it seems like there is always a lack of hands available
when working with electronics.

The probes that come with the cheap multimeters are to fat for measuring on the tiny legs of the SMD chips.
Try not to measure directly on a chip leg if there is an alternative. Sometimes one can easily see where the traces are going and measuring
on some other component, like a resistor or a capacitor, is much safer. Here is the equipment I use when measuring.


And, yes, I did check before and after power on. You have got that right. I plugged in the PSU and measured. Then I pressed the power button and measured again.
But that does not mean that this will always be the right thing to do. The strategy to choose should depend on the error at hand.

With a board like the one thomas has I would start by measuring the System DC.
If the System DC is there, I would move on and try the max1631 chip. If the system DC is not there - and I do not think it is - then
I would take a look into the power selection circuitry.

The System DC can be found on the bottom side of these 2 capacitors, which is located somewhere between the MAX1631 chip and the south bridge on
top side of the board. The System DC can also be found on pin 22 of the MAX1631 chip.

[AdaSch]

Bravo! great work!

p.s

do You have on sold adp3806 12.5 ?

my is broken and charge don't work

this chip is obsolete and i can't buy in regular analog/on semi distributors

cheers

[norm360X]

Interested to know if anyone has tried washing their keyboard in the dishwasher.

There are plenty of posts around about people who have done it, and in theory it should work even if it does take a while to dry out.

I placed a fairly dusty keyboard from a T41 in a basin of 50c water with some Fairy Liquid, later rinsed it off. I left it on a radiator to dry for about 8 hours. After which it worked fine.

PS. Very interesting thread. Nice to know even the deadest of dead system boards can be resurrected some times.

[Norway Pad]

Amazing!

I didn't even know that such knowledge and skills were out there among us "DIY's". (Or maybe this is your occupation?) However, as we said back in Norway: "Problems are challenges, and the impossible takes a bit more time"

[Tasurinchi]

@SMA: Great work you did there! I wish I had time to learn more about repairing computers and having the detective skills/patience you had.

Interested to know if anyone has tried washing their keyboard in the dishwasher

Not in the dishwasher. But I also cleaned a dusty, with (I guess coke) spilled keyboard using running hot water in my kitchen's sink. The water tap in my kitchen has quite strong pressure so I put the keyboard under the running hot water for a while, (before I've removed some keyboard caps that were very sticky), let it dry for a couple of hours (plus a short hairdresser session at the end), and installed it again. Worked like a champ!