T21 Backlight Woes - Possible Bad PCB Connection

[Robbyrobot]

One of my T21s has no LCD backlight and I could use some suggestions as to what to try next. But let me first explain...

I bought this T21 with the expectation that I could fix the backlight fairly easily. The symptoms were that the LCD panel would show an image, but this would be dark because of the lack of a backlight. Moving the screen back and forth often caused the light to go on, but was not a reliable method.

My first thought was a defective LCD cable, so I bought one (thanks, jamiphar) and installed it. But still no backlight. So I took the inverter and put it into another T21 I have that has no backlight problems. The backlight was still OK with the inverter from the defective machine, so both the cable and the interver are OK. That left the LCD panel itself, so I put the LCD panel of the defective machine in my working T21 and... no problems, backlight beautiful, display beautiful.

So the panel, cable and interver are OK and now I really have got a problem. From the fact that the backlight went on when I flexed the screen back and forth, I assume I have some loose connection on the PCB somewhere near where the LCD cable connects to it. Unless, of course, there's some other possibility I haven't yet considered.

As I see things, I'm going to have to disassemble the defective machine enough to get a good look at the PCB around the LCD cable socket and hope I can see a bad connection somewhere.

Does anyone have any other suggestions as to where I could check and for what?

Later:
With the keyboard, bezel, insulator and hinge cover removed I left the LCD attached and hooked up an external monitor. Put the defective T21 on a port replicator to be able to turn it on without the keyboard (thanks to Ray for this tip), then hooked up the keyboard once to set the display to both (LCD and external). The LCD did not light, but had an image, and the external monitor had a normal display.

Removed the keyboard to access the area around the LCD connection, started the computer and tapped and played with the area around the LCD connector. Nothing happened. But when I pulled the right edge of the PCB (next to the Ultrabay) slightly upwards with a fingernail, the image on the external monitor broke up. It normalized again when I pressed down lightly on the same place. So it's no joke, I definitely have a weak connection somewhere, although I wasn't yet able to get the LCD to light by moving the PCB.

[tfflivemb2]

Pay very close attention to the wires that connect the backlight to the Inverter. I had a problem with a T30, where I tested it in another system and it worked, but when I put it back it didn't. It turned out the there was a small slit in the cable that was causing a short, especially when it touched something else.

[rkawakami]

Try playing with the lid switch near the right hinge (and near where you were probably pulling up on the motherboard). If you have specified "no action" in Windows when the lid switch is activated, in other words, no standby or hibernate mode, then if the switch is stuck down the backlight is killed but the O/S keeps running.

[tfflivemb2]

Very good suggestion Ray...I haven't seen this problem since the days of the 600 series...but a VERY high probability.

[SMA]

If playing with the lid switch does not bring up something new, then I would suggest that you try to do another swap between the 2 machine - similar to what you have already done with the inverter and the LCD panel. But this time, try swapping the entire LCD assembly.
It is fairly easy to do and it would halve the number of possible error causes.

[Robbyrobot]

Pay very close attention to the wires that connect the backlight to the Inverter.

Thanks for the suggestion, but the LCD panel from the defective T21 has no problems in the working one, and more to the point, the LCD panel from the working machine also has no backlight when attached to the defective one.

If you have specified "no action" in Windows when the lid switch is activated, in other words, no standby or hibernate mode, then if the switch is stuck down the backlight is killed but the O/S keeps running.

Looks to me like the switch is on the I/O board rather than the mainboard, where the LCD connector is located - I was flexing the right "wing" of the mainboard next to the screen and on the upper left side of the Ultrabay cage when the display of the external monitor became distorted - but I'm sorry for the inexact description, since the I/O board is next to the Ultrabay. The backlight problem doesn't appear to be related to an operating system, however, since the problems and symptoms described are there immediately upon the POST. In any case, I played with the lid switch quite a bit before starting any more involved work since I had previously read your comments about it sticking.

If playing with the lid switch does not bring up something new, then I would suggest that you try to do another swap between the 2 machine - similar to what you have already done with the inverter and the LCD panel. But this time, try swapping the entire LCD assembly.

Already did that, and luckily it wasn't as difficult as I had expected... didn't even have to remove the display from the computer to do it (removed the bezel and the screws on the sides of the display, lifted out the LCD panel). Unfortunately - or fortunately, depending on how you see things - the LCD panel of the defective machine appears to be just fine, and works (with either inverter) without problems in the other (working) T21. And as mentioned, the assembly from the working machine has no backlight in the defective one.

I plan to remove the mainboard and inspect the upper and right "wing" portion for breaks or bad solder joints today, but only give myself a 50-50 chance of seeing anything because of my inexperience, bad eyes and the difficulty of inspecting a board with a 30x magnifier (I can't see much at less magnification) while simultaneously trying to hold a strong light on the area. Still want to give it a try. I'll report here if I see anything suspicious.

[Robbyrobot]

Just one last post on this unfortunate situation. In the meantime, I completely dismantled the computer, held the mainboard in my hands, examined it with a 30x magnifying glass, scanned it on both sides at high resolution and examined the scan - and saw nothing remarkable.

Just to be sure, I then re-mounted the mainboard in the base, added the CPU and the fan, attached a good LCD panel and started the computer both with the help of a port replicator, and using a keyboard. All the lights went on, then except for the power light off; the CPU fan was operating. No beeps, as the speakers weren't attached. And there was no video signal either to the LCD panel or to an external monitor I attached (after repeated pressing of Fn-F7 with an attached keyboard). Both displays stayed totally black.

Oh yes, the bottom of the graphics chip and the AGP chip (on the bottom of the mainboard below the CPU) became warm to the touch after the computer had been switched on for a couple of minutes - but there was still no signal to the LCD panel or external monitor.

During all this time, I flexed various parts of the mainboard up and down, tapped it in various places with a fingernail and in general did everything I could think of to detect a bad solder joint or cracked PCB lead. But nothing had even the slightest effect.

So I have one of these fine, mysteriously "defective" mainboards of which there are apparently hundreds if not thousands floating around - if I trust the various reports I've read here and elsewhere. And really only one question to the general readership: are these T2x mainboards really as delicate and easily damaged as it would appear? And if so, is there somewhere a source of cheap mainboards as replacements? The ones I see at Ebay generally fetch nearly as much as a whole "parts computer".

[rkawakami]

Given that you seemed to have eliminated the panel, inverter and ribbon cable and have isolated the fault to the motherboard, then about the only thing left is checking all of the fuses that you can find. This would be my assumption since the LCD panel is not working but I don't know why the external video is dead as well. Oh, I hope you had a working memory module installed as well as the CPU and fan, otherwise a blank screen can result.

These symptoms are not the dreaded "blink of death", which I believe you are referencing. At least the fan is spinning and the power LED remains lit on your board. The life on a BoD board is just a 0.1 second flash of the hard drive LED. As it appears that the T2x (and others) may have been some of the first systems assembled with lead-free solder, coupled with a flexible chassis, this appears to me to be the cause of many failed motherboards. I know of no "cheap" source of boards other than eBay.

[Robbyrobot]

Given that you seemed to have eliminated the panel, inverter and ribbon cable and have isolated the fault to the motherboard, then about the only thing left is checking all of the fuses that you can find. This would be my assumption since the LCD panel is not working but I don't know why the external video is dead as well. Oh, I hope you had a working memory module installed as well as the CPU and fan, otherwise a blank screen can result.

You just said the magic words... naturally I had not installed a memory module, so I just did that and tried again. Now I have the original situation back, i.e. a signal to the LCD display, but no backlight, and the external monitor works fine. As previously mentioned, I suspect the external monitor signal was disrupted because I flexed the part of the mainboard with the plug for the port replicator.

Uh-oh... and just to be sure, I attached another working display and started the computer up - and now suddenly the LCD display lights up as if nothing had ever been wrong! OK, I took it out, replaced it with the original display, started the computer... backlight out. Then again the other working display - and this time it also stayed dark. Needless to say, both displays work fine on another T21.

Playing around with the display plug - my first idea, and one that would have agreed with the original observation that the backlight sometimes went on when I moved the display - had no apparent effect. I've already checked the fuses I can find, in particular the ones at the top of the mainboard near the display, but they all have contact and are apparently OK.

And now? Didn't you write one time that the CCFL needed an initial "burst" of high-energy power to ionize the gas in it, and then less power was required to keep it running? I'm wondering if a specific pin on the mainboard socket controls this "firing" impulse to the inverter and is loose/corroded or whatever (although nothing is visible even under magnification).

I know of no "cheap" source of boards other than eBay.

Such a pity... at this point, I'd probably buy a dozen just to be prepared for all eventualities

[rkawakami]

You just said the magic words... naturally I had not installed a memory module, ...<snip>

I thought it might be a possibility when you also mentioned that the speakers were not connected. You couldn't hear the 1-3-3-1 beep sequence .

Didn't you write one time that the CCFL needed an initial "burst" of high-energy power to ionize the gas in it, and then less power was required to keep it running?

Yes, that's my understanding on how fluorescent lamps work. Don't know if there's a specific signal which tells the inverter to "fire up". I would have thought that the design of the inverter itself would provide for that, given it is properly supplied with the power it needs.

[frankie66]

How did you conclude the replacement LCD cable is ok? If you bought it used is it possible that it's also damaged?

Try what SMA suggested. Connect the working screen (with the lid and everything, not just the LCD part) to the suspect thinkpad. If this works, the problem is not with the base.

Does anyone know how to examine the LCD cable for breaks?

[Robbyrobot]

How did you conclude the replacement LCD cable is ok? If you bought it used is it possible that it's also damaged?

Actually, it looked brand new when it arrived, but may have been used.

Try what SMA suggested. Connect the working screen (with the lid and everything, not just the LCD part) to the suspect thinkpad. If this works, the problem is not with the base.

That's exactly what I did after I found that the replacement cable didn't fix the backlight problem. I had bought another T21 in the meantime, with a definitely working display, and rather than keep swapping things piecemeal, I just attached the working display to the T21 with the backlight problem. And found that the working display also had no backlight on the problem T21. I then took the display from the T21 with the backlight problem and attached it to the other T21 with no such problems - and found the display worked just fine.

No, obviously this would have been the first thing to do before taking the whole computer apart. And frankly I was hoping against hope that there would be some defect in the display, since that would have been fairly easy to fix. But unfortunately that was not the case

[dickchailey]

Are you sure you don't have a defective lid switch or connection to the subpanel. As posted above it will give this problem and a cheaper fix than a new board.

[Robbyrobot]

Are you sure you don't have a defective lid switch or connection to the subpanel.

At the moment, I'm not really sure of anything except that the display itself is OK

I've played with the lid switch again and again, and it doesn't stick. However, I've never actually removed the subpanel, gone underneath and checked the function with a multimeter. That would be something I'll have to do now.

The connection to the subpanel is a plastic plug just above the CPU on the underside of the mainboard, and to the extent I can check it appears to be OK. Since the screws holding the fan to the CPU also fasten the mainboard tightly down against the socket on the subpanel, I'd be surprised if this were a problem - particularly in view of the original symptoms, that the backlight sometimes lit when I moved the display back and forth. And one time just yesterday, when I was switching a couple of displays (both good), the backlight went on just one time again. But nothing I did afterwards - pushing the connection down, rocking it - brought the backlight back, so I'm still looking for a loose or bad connection if that is, in fact, the cause of this.

I may be bidding on an Ebay lot of spare parts tomorrow, including three T21 mainboards "without function" (whatever that means), and so may have some more stuff to play around with in a little while.

Edit: Got the three mainboards, so now I can play around with them and see if I can get at least one to work... I'm still amazed that the most expensive component of the T2x series is also the one most likely to fail.