Thinkpad T21 Power Problem

Message

Russell.reed · #1 Post by **Russell.reed** » Wed Dec 13, 2006 9:29 pm

I have a IBM Thinkpad T21 that I am working on for a friend. When I start up the computer on the battery Windows XP starts to load and when it loads windows the laptop shuts off and all of the panel lights by the power switch below the monitor blink off and on a bunch of times until the battery dies. When I plug in the adapter the battery charges but when the battery is connected and I try to start windows, it will not load and the lights all blink on and off a bunch of times.

My friend wants me to fix this laptop but this is the first time I have ever had a machine do this. Any help would be appreciated.

Thanks!

Website · #2 Post by **tfflivemb2** » Wed Dec 13, 2006 11:03 pm

Welcome to the forum, wish it was under better circumstances...

Does the T21 boot and run normally with the battery removed?

It sounds like their might be a problem with the battery itself, as in either not enough life in the battery, or the Thinkpad can't properly read the battery.

Does it have Battery Maximizer installed on it? (if you aren't familiar with it, it should be a green battery next to the system tray)

nikki605 · #3 Post by **nikki605** » Thu Dec 14, 2006 9:03 am

I'll be following this thread with interest. I also have a T21 with nearly the same problem.

Russell.reed · #4 Post by **Russell.reed** » Thu Dec 14, 2006 9:10 am

Thank you for the welcome.

The T21 does not boot with the power adapter attached. I do have a green battery on the system tray on the laptop, but it does blink and change to orange after a while.

nikki605 · #5 Post by **nikki605** » Sun Feb 04, 2007 2:40 pm

From reading this and other topics on the forum, it seems like power problems, with varying symptoms, are fairly common. So, I took a different approach. While searching around Ebay for T21 systemboards (Most were under $100, but I just didn't feel comfortable with a board described as "removed from working system") I came across a repair company that replaces components on the systemboard - LaptopOnCall.com out of Cincinnati. So I sent it off to them after shredding all of my personal data files from the hard disk.

I've had it back now for several days and seems to be running fine. It hasn't powered off, yet. They replaced 2 power chips and 3 capacitors for $158 + shipping. Maybe I would have been OK with a board from Ebay, but I didn't feel like risking it.

#6 Post by **RealBlackStuff** » Wed Feb 07, 2007 7:47 am

Is there any chance of you taking pictures with the replaced parts marked on them, and attach the links to them (rather than publish them inside the thread)?
This might help a lot of others who seem to have similar/the same problems.

I realize you'd have to disassemble your laptop for that, but in the name of T2x-humanity, this is a small sacrifice!

Msmax · #7 Post by **Msmax** » Wed Feb 07, 2007 2:42 pm

Or...
please post the item numbers/tags which are replaced.
That will be very helpfull too.

Thyanks, Msmax

nikki605 · #8 Post by **nikki605** » Wed Feb 07, 2007 3:22 pm

I popped out the keyboard the day I got it back looking for the same thing, but I couldn't see any components that were obviously re-soldered. Whoever did the work was very neat.

I'll take another look (I may have to totally remove the systemboard to look at the bottom) and see if I can figure out which components they replaced. If I can, I'll post what I find.

nikki605 · #9 Post by **nikki605** » Wed Feb 07, 2007 11:59 pm

I removed the systemboard this afternoon so I could examine both sides. Try as I might, even with a Maglite flashlight and a magifying glass, I could not spot the components that were replaced! I looked for anything "re-soldered", anything that looked "new" or "different" and I couldn't see anything out of the ordinary. Whoever did the work was very, very good at un-soldering old parts and soldering in new parts. Must have been a real small tip on his iron!

Sorry

If anyone has any ideas how to locate replaced parts, I'm open to suggestions. I don't think the shop that repaired it will tell me specifically which chips & caps they replaced. That's how they make their living.

I've been using it for over a week now and it hasn't lost power yet! So, whatever they replaced seems to have solved the problem. Before I had it repaired, it wouldn't stay powered up for more than a few hours.

#10 Post by **rkawakami** » Thu Feb 08, 2007 2:47 am

If they have done the repair correctly, you probably would not be able to tell by looking at the soldered connections. A hot air reflow station is the preferred method to detach/attach surface mount components, not a soldering iron. About the only thing I could suggest that would help identify which two power chips have been replaced is to look for a "date code" marked on the top of the package. Normally a four-digit number is used to indicate the year and week of when the silicon die was packaged. For example, "0528" would mean 2005 and the 28th week. Seeing a part with a date code which is newer than the date when the laptop was assembled is a good clue

. The capacitors are probably too small to have any kind of identifying marks on them. Unless they are the larger electrolytic types (big "can" or "block" looking parts with a "Cnn" next to them).

nikki605 · #11 Post by **nikki605** » Thu Feb 08, 2007 8:50 am

Thanks for the info, Ray. The next time I have reason to take the laptop apart, I'll try reading the date codes. I agree with your comment about the caps. Depending on the type, it may be impossible to determine which ones were replaced.

#12 Post by **RealBlackStuff** » Thu Feb 08, 2007 12:09 pm

Not to say anything detrimental about that shop (or others like them), but it would be quick(er) and cheap(er) for such shops to buy motherboards in bulk, and then just replace the whole motherboard, while charging the customer a few 'fantasy' replaced parts, whose sum exceeds that of the mobo.
T2x mobos are quite cheap.

If someone can perhaps point me to a picture of what these 'power chips' look like, I could try and locate/replace them before I replace the whole mobo.

The reason why I am interested is, that my T30 mobo is fuxored (no LCD but external monitor is OK, already replace cable, inverter and LCD), where a replacement used mobo will run into at least $150.-
I am looking for the assembly-2 mobo with security chip (P/N: 91P7212 or 26P8423 91P7698).

#13 Post by **rkawakami** » Thu Feb 08, 2007 2:59 pm

realblackstuff wrote:<snip>... but it would be quick(er) and cheap(er) for such shops to buy motherboards in bulk, and then just replace the whole motherboard, <snip>

Yes, that thought entered my mind too. Current eBay Buy-it-Now prices start out at about $70 so with a total of $150 that was charged to nikki605, that would be about right. However, if the shop did indeed find something or know that this is a fairly common problem, the cost of replacement components is probably less than $10. Removal and re-installation of the motherboard is still required so that cost is identical. Should only take a couple of minutes to use the hot-air reflow station to remove the 5 components and install the replacements. That adds some $$$ in labor costs over a mobo swap. If I was the owner of the shop and knew which components to replace, I'd go with the re-flow operation as that would maximize my profits and I wouldn't have to keep $50 boards in stock for these kinds of repairs (assuming that's the wholesale cost for refurb'ed motherboards).

I just checked out their website. They list some typical power on problems and say what they replace to fix those issues. The "MOSF power chip" is actually a MOSFET power transistor, similar to the one I blew up recently. So given that, I'd go with the opinion that they perform some component replacement.

One way to determine if the motherboard was swapped would be to boot into BIOS and check the system unit serial number, the system board serial number and the UUID. If the shop didn't take the time to reset those fields then it should be obvious what happened, if you had made note of those items prior to repairing the system.

Seeing as I'm probably the only one who documents these things in the 28 or so systems that have been in my possession, that's probably not going to happen in this case. So, check the system unit serial number in BIOS against what is printed on the bottom of the laptop. If they are different, then a mobo swap occured. If they are the same, then you can't really say anything definite about a swap unless you locate the system board serial number on the motherboard and compare THAT with what's in BIOS.

nikki605 · #14 Post by **nikki605** » Thu Feb 08, 2007 5:27 pm

For what it's worth, I compared the unit model/serial number in BIOS to the label on the bottom and to my original paperwork when I bought it new. They all agree! The correct model/serial number is displayed in the BIOS. The BIOS also displays the systemboard serial number, however, I didn't have that number recorded anywhere, so I have nothing to compare it to. Same with the UUID - I didn't have it recorded anywhere.

A related question - the label on the bottom appears to include a date code. Mine says 00/11. If that means it was manufactured in the year 2000 and either week 11 or month 11, then that is about right since I bought it in July, 2001.

#15 Post by **rkawakami** » Thu Feb 08, 2007 6:02 pm

nikki605 wrote:The BIOS also displays the systemboard serial number, however, I didn't have that number recorded anywhere, so I have nothing to compare it to.

I believe that there's a label on the motherboard near the MiniPCI slot which has the system board serial number on it. I can't remember if you can see it if you remove the MPCI card or whether you have to remove the entire motherboard. I'm at work at the moment and don't have access to my free T23 motherboard to know for sure. I would hope that the T21 mobo would have the same type of label/markings.

You analysis of the date code seems correct. I would lean toward the last two digits as being the month. I'd have to check all my systems at home to see if any of them exceed 12 to know for sure. I probably should update the documentation I keep on my systems to include this data.

#16 Post by **RealBlackStuff** » Thu Feb 08, 2007 6:45 pm

You may be right that they (the shops) know which parts to check/replace, if not the whole mobo.
Then there is always the Hardware Maintenance Diskette, with which you can re-enter the serial no. and UUID, if you DO change the mobo.
Can I ask again, would you also have a picture of a fuse, or better, is there any guide available that describes the individual parts, such as resistor, capacitor, diode etc. that are not in their standard (round) format 'of old'?

#17 Post by **rkawakami** » Thu Feb 08, 2007 7:19 pm

Yes, you are correct about the HMD. However, if I'm correct about the serial number label on the motherboard, then they would also have to remove/replace that label to match what's been programmed into the BIOS.

My post entered at 11:59am PST (two back from here) has a link to a thread were I have posted a picture of a portion of a T23 motherboard. The part that blew up is a MOSFET. Here's a link to the large picture:

http://www.rkawakami.net/ibm_t2x/let.jpg

There are MANY such components that look similar. Probably too many to replace all at once unless you can physically see something wrong with it (ref: my picture). You also have the problem of selecting the proper replacement part. Not all MOSFETs are the same. The picture has six components which look identical to the one that now has a hole in it. Without looking closely at the part numbers on them, they could all be different from each other. You need the part number off of the chip (assuming you can still read it) and then you have to locate a supplier that's willing to sell you ONE of them. Digikey.com and Mouser.com are such companies. THEN you have to be able to unsolder miniature small outline ICs (SOIC) that have four or more leads without destroying the circuit board and solder on the replacement.

As far as identifying components on the motherboard, generally it's easiest if you refer to the legends next to the parts. "Cnn" are capacitors, "Rxx" are resistors, "Qnn" are normally transistors, "Fnn" are fuses, "Lxx" are coils/inductors/chokes, "Dxx" are diodes. That tells you what type of component it is but NOT the "values" associated with them.

re: fuses

If you look at the picture I've linked to above, there's a fuse next to the black terminal on the right side of the picture. That's the connector to the Ultrabay. The fuse is the creme-colored block that has silver endcaps on it. It's a 10A fuse made by Littelfuse and is labelled F11 on the board. There's a much smaller fuse to the right of it, marked F10. If you look at the top of the picture you will see a small connector just barely visible to the right of the mounting hole in the motherboard. That connector is what the DC power input jack plugs into. Next to the connector is the label "F2". That signifies a fuse just below the connector. It's the upside-down "FU" component. I believe that is a 7A fuse, but can't remember who makes it. I had identified who made it and order a couple from Mouser.com. There are also what I believe to be thermal fuses in the pictures. Those are the green squares with "7" and "15" written on them. In short, fuses can be in various packages and ratings.

#18 Post by **RealBlackStuff** » Fri Feb 09, 2007 3:32 am

Thanks Ray,
that was a very informative explanation.
Armed with this knowledge, I'll re-inspect my mobo, and measure all the fuses to start with. Maybe I get lucky.

#19 Post by **rkawakami** » Fri Feb 09, 2007 4:49 am

To futher expound the details about changing motherboards...

I just ripped open another newly acquired T23 tonight to fix a problem. It had a slightly different looking bar code label that has been placed over the "normal" label. It does not have the apparent date code at the end (i.e., 02/02). It says it's a type 2647-7NU and serial number 78-LHA0G. In the BIOS it says it's a type 2647-1U1 and the serial number matches. The system board serial number in BIOS matches the paper label on the motherboard underneath the MiniPCI modem/802.11b combo card. Going online and checking the serial number, IBM says it's supposed to be a 2647-1U1, which matches the BIOS but not the "new" label. Confusing?

When I disassembled the laptop, I noticed that it used 2 wrong screws at the front edge of the laptop (the ones that hold down the keyboard bezel). Inside, it was missing the long screw underneath the trackpoint buttons. So obviously someone was fooling around with the system before I got it.

So, which type number is correct? The 7NU (the one visible on the outside of the laptop) is supposed to be a non built-in wireless model with a Windows XP COA. What I have is a wireless system (with the combo card and antennas) and a Windows 2000 Pro COA, which matches the 1U1 type.

Here's the eBay auction:

http://cgi.ebay.com/ws/eBayISAPI.dll?Vi ... 0075037179

I took a leap of faith that the seller was describing the system well enough that it was something I wanted (the built-in wireless; I'm a sucker for these models). I asked him to double-check the type and serial numbers he gave out in the auction because when I checked the serial number, it came up as 2647-1U1. I had purchased several other T23s from him in the past so I knew that he was describing everything accurately, from his standpoint anyway.

nikki605 · #20 Post by **nikki605** » Fri Feb 09, 2007 2:01 pm

Thanks Ray for the great information! After reading your latest posts, I did some more looking at my own T21. I found the systemboard labels under the mini-PCI card, as you described. Here is a link to a photo of these labels:

Label photo

The uppermost barcode label does include the systemboard serial number being reported in the BIOS. They are the same - J1FKN1269X2. But then I got to thinking - if they did replace the systemboard the paper label serial number on that replacement would most likely agree with the serial number in that replacement board's BIOS. So does this really prove anything? Probably not.

All they would have to do is change the unit model & serial numbers in the new systemboard's BIOS. I know nothing about the UUID number, so I'm not sure of its significance.

So, I popped the keyboard one more time to look for date codes on some of the ICs. I could see 7 ICs with obvious date codes:

1 - 0035
1 - 0036
4 - 0039
1 - 0040

Reading the date of the newest IC, that puts it at the week of September 24, 2000 (if my week number calculation is correct). If we are reading the unit date code correctly (00/11) and say the laptop was manufactured in November, 2000, then all this date information falls in line and seems to be correct.

Two other notes of interest:

1 - I mentioned that I took delivery of the new laptop originally in the latter half of July, 2001. What I forgot until I looked back at my paperwork, was that the laptop had a floppy drive problem right out of the box. Since it was under IBM warranty, I got an RMA and sent it to their repair facility in Memphis. THEY REPLACED THE SYSTEMBOARD! That replacement was done August 10, 2001, which was within the first month after I received it.

2 - When I sent the T21 to laptoponcall.com, I asked about WinXP (and CPU-Z) displaying that the CPU was only running at 795MHz. (The 100MHz FSB was only running at 99.5MHz). When they contacted me for authorization to repair the power problem, they said the crystal(s) oscillators were old and they could replace it/them for an additional $100. I said "no thanks" since I already had a 1GHz CPU sitting here ready to install when I got the unit back. When I got the unit back, WinXP & CPU-Z still said the CPU was running at 795MHz. My point is this - if they did replace my system board with a referbed one, what are the chances that the CPU would be running at the same exact slower speed of 795MHz? A referbed board would have probably fixed the slower speed. If not completely, the reported speed may have been slightly different.

I don't think laptoponcall.com replaced the systemboard, based on what I've been able to see. I believe they really did replace bad components.

#21 Post by **rkawakami** » Fri Feb 09, 2007 4:06 pm

nikki605 wrote:But then I got to thinking - if they did replace the systemboard the paper label serial number on that replacement would most likely agree with the serial number in that replacement board's BIOS. So does this really prove anything? Probably not.

Sorry if my earlier rambling was a little confusing on this point. What I was basically saying is that if you already had documented the system board serial number before sending it in for repair, they would have had to transfer the label and fix the BIOS so that it matched what you had originally, IF they changed the motherboard.

nikki605 wrote:So, I popped the keyboard one more time to look for date codes on some of the ICs. I could see 7 ICs with obvious date codes:

Your analysis of the date codes appears to be correct.

nikki605 wrote:I asked about WinXP (and CPU-Z) displaying that the CPU was only running at 795MHz. (The 100MHz FSB was only running at 99.5MHz).

As far as I can remember, none of my systems run exactly at 100Mhz FSB, but I normally don't pay any attention to the speed readings when executing memtest86+ either. A couple of 0.1Mhz doesn't seem to be that out of line. When it gets to be HALF of what it should be (running Windows 2000 on battery, for example), then I notice!

Msmax · #22 Post by **Msmax** » Sat Feb 10, 2007 2:17 pm

My guess is the ADP3421 is causing these powerup problems.
It is located just beneath the keyboard. Please check if you can see if there is some residue of the soldering paste used for these kind of repairs. Should be visible on the board, even if soldered with a hot gun.

As for the capacitors, I think the SMD caps in the output lines.

As soon I will have some more spare time I will try to fix a T21 board which I could get hold off. The ADP3421 chips are already delivered... 9.90 euro p/p. Bought 2 working MB's from Ebay for only $29 few months ago.

Msmax

nikki605 · #23 Post by **nikki605** » Sun Feb 11, 2007 1:20 pm

I found the ADP3421. I also read the datasheet on the device, a DC-DC Converter Controller, and found that it directly interfaces with the ADP3410 Dual MOSFET Driver. This sounds like the same "MOSF" device mentioned on the laptoponcall.com web site when it describes failed power chips.

I took a photo of the area of the systemboard where these 2 chips are located. Sorry its not of better quality, but my digital camera is old and not up to the task of taking good, detailed, close-up photos.

Click here to see the photo.

I looked at these 2 chips closely, but could not see anything that would make me believe they had been replaced. No rosin residue, no visual indication that anything had been over-heated during re-soldering. I just can't be sure if these were the chips or not. Neither chip had what I could determine to be a date code, either. I certainly can't tell which 3 caps they replaced. They are too small.

Sorry I couldn't be more help.

vlyne · #24 Post by **vlyne** » Wed Jun 27, 2007 2:18 am

Dredging up an old thread.... I looked at the image of the motherboard and two things I noticed are: 1) it looks like the guide for the screw mount appears to be missing (south east of the ADP chip). Not sure what implications it has for the rigidity/flexibility of the motherboard and connectors?, and 2) there appears to be some thin black wire coming from the end of the bank of 8 capacitors (top right of the last one). Is this an image artifact? If not where does it go to?

As for the 3 caps that were replaced, there are 5 large capacitors, one in the top left (out of the picture) that seems to be related to the docking station connectors, two visible in the image and another two to the right outside the image. The two in the image must be suspects but I'm not sure about the other.
Cheers

nikki605 · #25 Post by **nikki605** » Wed Jun 27, 2007 7:39 am

I should have updated this thread, but I just plain forgot.

Whatever component replacements laptoponcall.com made to my mobo, they didn't hold up. Sometime in late February, it started failing again with the same symptoms. I sent the T21 back to them under warranty in March. It didn't fail for them and they could find no problem. I got it back (after paying shipping both ways) and within a few days it started failing again. I saw no point in sending it back a 2nd time as shipping was getting expensive and my TP was still unuseable.

I finally broke down and bought a used mobo off eBay ($70) and installed it early April. The T21 has not failed (knock on wood) since.

My point being - for those of you looking to replace mobo components to fix a problem like this, even a company in the repair business didn't get it right.

Unless you're just the curious type, you may be wasting your time.

vlyne · #26 Post by **vlyne** » Thu Jun 28, 2007 4:47 am

That’s really disappointing that a repair shop didn’t get it right –twice- and charged you enough to buy another laptop. I’m at the point where I’m not as tempted to buy another T20/21/22 even a working one at a good price. Why bother when there are more reliable, or easier to fix, models around? Well…not unless some curious and determined soul finds that elusive fix for the power-on problem. Speaking of which, if were in CA I’d organise a working bee in the TP dungeon of a moderator who has been a Ray of hope on this forum

C’mon! it’s summer there, and you could have a barbeque as well, and a toxic bonfire of dead motherboards if all fails.

Thinkpad T21 Power Problem

Thinkpad T21 Power Problem

Thinkpad T21 Problem

Who is online