T41 LCD inverter problems

[jimmy274]

Hi guys,
Quick question - I have a T41 with XGA screen - no backlight. At first I thought it's the CCFL, but when I tried it with a good CCFL (I have it from a broken T40), it still did not lit. So, next obvious thing is the inverter. This is where the problem is - I can't seem to figure out how an inverter would die out. I did, however, measure the secondary of the inverter transformer (just before the backlight jack) and it's an open connection - I suppose this would have to be a problem - a transformer should never have an open secondary...
Anyone care to comment?

[poshgeordie]

You are correct - the secondary coil should not be open circuit - have a look at the circuit diagram of a typical (not Thinkpad) inverter here.

The link is to a UK company which sells some replacement inverter coils but it may be cheaper to buy a replacement inverter.

Note that the last time I checked with this company they did not have any replacement inverter transformers for T4x series, but this was over a year ago.

I guess you have checked that the Thinkpad works OK on an external display?

[jimmy274]

Yeah, works fine on an external display. Come to think of it - I'm not even sure what a secondary is - T41 inverter transformer has 4 pins (+1 unused) on the input side and 2 (+1 unused) on the output.

p1 |---- ----| p6
p2 |----
p3 |----

p4 |---- ----x u2
u1 x---- ----| p5

p6 is not connected to anything, according to my multimeter. p1, p2 and p3 are forming first coil, p4 and u1 second, and p5 and u2 third coil. u1 and u2 are not connected to the pcb. Just weird...
By any chance, do you know what are those CPx pads on the inverter - it seems that, even though the FRU's the same, CPx pads are wired differently on the same inverters?
Thanks for such a quick reply btw

[poshgeordie]

Having rechecked the link in my first post, the nearest inverter I can find is this, but note that there are two more segments worth of secondary coils than the standard 5 segment secondary on the T4x inverter, which will produce a higher output and could I guess shorten the life of the CCFL.

However referring to this diagram (which matches the pins on the T4x inverter) the primary coil connects between pin 1 and pin 3 (pins 2 and 3 are also connected together on the PCB).

The secondary is connected between the little stubs next to pins 4 and 5 respectively.

I don't quite understand your "diagram", but hope this clarifies. Also do you want an explanation on how the inverter works?

[jimmy274]

Thanks, but I've covered most of the inverters on my master thesis - theoretically that is - I've designed an inverter control pattern. Practically, I don't know anything about them (fresh outta school) - you might fill me in with this
Anyway, just found out the inverter is good - I've tried it with a known good T40 and the whole screen works (CCFL + inverter) - the problem lies somewhere on the mainboard. It would be helpful if you know something about what I need to check. I've already checked fuses F3, F7 and F8 (I saw that from one of your posts) and they're all fine. I'm thinking some of the TPC8xxx power transistors are messed up, but I can't seem to find one - for now, I saw that a working T40 generates 16.4V on the drain of the first TPC8014 on the upside of mainboard, while my broken T41 board produces 15.4V at the same spot - could be a problem (this is measured with the adapter in, battery out). I'll repeat the measurements tomorrow, but I would still like to know more about the power circuitry for the inverter on the mainboard itself, so I could know where to look for a malfunction.

[poshgeordie]

Sorry, I'm not sure about voltage readings and the circuits in that area.

Going back to your first post:

T41 with XGA screen - no backlight.

Is the screen completely blank or can you see the display very faintly but with no backlight on?

If you know the display, inverter and backlight etc are all working on another Thinkpad, then there may be a number of reasons:

A component fault (which you've mentioned).

Fuses (which you've mentioned).

A soldering fault on the motherboard LCD connector - sometimes these soldering joints become open circuit (broken).
I'd suggest not trying to resolder unless you are very experienced. Note that when resoldering the connector do not make solder bridges between the tiny pins - use liquid flux over the joints which will help prevent the solder bridges.

Unsoldered Graphics chip - it's possible to have video output on the external VGA connector and none on the (internal) display since there are separate video outputs on the chip.
Maybe try pressing hard on the chip and powering up the Thinkpad and see if you get display on the screen.

[jimmy274]

The screen is working (I see the IBM logo in the dark), but without backlight - typical symptoms of a bad CCFL or a bad inverter. However, I've established that both the CCFL and inverter are fully functional by trying them out on another Thinkpad T40 which incidentally has the loose GPU problem, but fires up display just fine. So - fuses or connectors are the problem. So far, I've checked fuses F3, F7, F8, F13, F15 and all of them are fully functional.
I've ruled out loose GPU problem, because the laptop works flawlessly on external monitor - no freezing, no problems whatsoever, even if I tilt it, flex it - nothing.
So - it must be a fuse or a connector. Now, my question to you is - Do you know which pins of the T41 LCD connector are dedicated to inverter power, that is - which pins are dedicated for power supply of the inverter? This would be very handy for me to know, because I could just try and see if there's sufficient voltage on them. If not, back-trace to the faulty component...
Another thing I did notice is that there is a large coil beneath the LCD connector on the underside of the mainboard, driven by TPC8203 dual power MOSFET. It seems that it's connected to F8 fuse, and thus could be a DC/DC converter supplying power to the inverter, but I'm not sure...
Any help would be greatly appreciated... Oh, and thanks for your other posts, you really seem to know this stuff better than anyone on the forum and you really did help me. If I do find out something in the meantime, I'll post.
Thanks in advance...

Ivan

[SMA]

There are 3 supply voltages going to the LCD assembly. They are all going through a fuse. That is F3, F7 and F8.
Since you do see the IBM logo the F3 fuse (3.3V) will have to be fine.
Also, if the power led lights up, then the F8 fuse (3.3V) will have to be fine.
This leaves only the F7 fuse, and yes that one does carry a voltage that is dedicated to the inverter.

There is at least 2 ways of checking fuses.
The usual one, where the resistance through the fuse is measured
and the one, where the DC voltage on both sides of the fuse is measured. The voltage on both sides of the fuse should be the same.
Note that, besides from telling whether or not the fuse is god, the second method mentioned also provides information about whether or not there
is voltage coming to the fuse.

When the ThinkPad is running from the AC Adapter, then there should be 16V on the F7 fuse.
Since the 16V on the F7 fuse is the same 16V (the System DC) that is used to power the rest of the ThinkPad, then only a broken trace
can prevent that voltage from being present on the F7 fuse.
The 16V out of the F7 fuse, goes through a mosfet to the LCD connector, pins 58, 60, 62.
Most likely, the mosfet will be a TPC6003 - top mark S2D.

Are you sure that you have checked the F7 fuse? Could it be that you have mistaken a resistor as being the fuse?
It should be easy to see/follow the traces from F7 to the mosfet and further on to the LCD connector.

The large coil you are referring to is used for battery charging and has nothing to do with the LCD / backlight.

[jimmy274]

That's what I'm talking about!
Ok, so, I'll fire up the T41 and check F7 for 16V. I sincerely doubt that the fuse itself is a problem, since I've already checked it twice - couldn't be more conductive I'm sure it's the fuse (black and has that little 'N' on it).
Afterwards, I'll try to back-trace to the underlying cause, if any. I'll post tomorrow when I'm done with it.

[poshgeordie]

Hi Ivan

I've ruled out loose GPU problem, because the laptop works flawlessly on external monitor - no freezing, no problems whatsoever, even if I tilt it, flex it - nothing.

I wouldn't rule this out since it's still possible - I'll need to check though to see if it also controls the inverter voltages.
I don't think it's Southbridge though - the inverter voltage is one of the few components controlled by it!

I'm somewhat mystified by this fault and don't know what to suggest next.

you really seem to know this stuff better than anyone on the forum

Absolutely not! SMA for instance is the technical guru here and has taught me. See his post above compared with mine. I could mention many others too.
Thanks for the comments though...

[jimmy274]

Ok, still hadn't checked for anything since I'm not home yet, but I will post later today. I agree with you poshgeordie about the loose GPU - it still could be that this is causing it. However, I think that ATi is not directly driving inverter - there should be some sort of power management unit in the path from GPU to the inverter, as there's no dedicated pins in the DVI standard for backlight - tft lcd monitors do all of their power saving with seperate circuitry (for instance, I have an LG that dims when not used for more than 5 minutes - not controlled by th monitor or the GPU driver, but by the monitor itself). More about DVI:
http://en.wikipedia.org/wiki/Digital_Visual_Interface
As you can see pin 14 is for the whole monitor standby, which means it shuts off the entire screen+backlight. The connector on the ThinkPads is probably a customized DVI, and GPU is probably only having DVI connections - anything inverter related is controlled only by power management units on the mainboard.
Does this make any sense SMA?

[jimmy274]

Ok, I've just measured the voltages. First my setup - I just plugged in the mainboard, CPU, memory, battery, speakers and keyboard. I did NOT plug in the screen.
Voltages on F3 and F8 are the expected 3.3V, running through an additional 6-pin ICs and on to the pins of the LCD connector - everything checks out.
Voltage on F7 is some 15.6V and goes all the way to the 6-pin IC, but this is where it stops. I suppose that 6-pin ICs are the additional driving MOSFETs that SMA was referring to. Now - I guess that either the driver is dead, or the gate drive voltage is dead.
I would like to emphasize that I haven't plugged in the screen and that pins 58, 60 and 62 are at 0V. So, no power coming to the LCD connector. It's either dead driver or dead drive voltage on the driver (I don't really know what the drive voltage should be, but I'll measure it and post later).
I'm posting in an hour or so, I just need to measure the rest of the voltages.

[jimmy274]

Ok, Murphy's Law strikes again - it's the gate driving voltage. Should be some 23.4V, as measured on a working T40, in my case it's 0V. Now, I can't seem to back-trace the gate driving line to anything (my guess is that the power management IC for this particular driver is somewhere on the underside of the motherboard).
So... SMA, do you know what and where is the IC driving the 6-pin driver mosfet for pins 58, 60 and 62? This is the key - it's either a broken trace, or an IC malfunction.
Regards,

Ivan

[poshgeordie]

Does the IC / MOSFET have an ident marked on the board?

If not could you post a photo of where it is on the board and I'll see if I can get some more info on what the gate's s connected to, or could you desribe which one it is (eg "it's the IC connected to F3").

Also with the power off, is the gate shorted to ground?

V RISKY idea - if it's not shorted what happens if you apply 24V from elsewhere on the board to the gate?

Again V RISKY - Also on the working T40, is the MOSFET acting just as a switch in that Vin = Vout when gate voltage is applies? If so what happens when you short Vin to Vout pins?

Again I state the last two ideas are V risky so do them only if you're prepared for further component damage.

[jimmy274]

Hi Nick,
Ok, I'll post in an hour or two with the correct markings and pics.
About your very risky ideas:
No 1: Have thought about it, but I don't know where to look for another 24V on the board.
No 2: Mosfets are used as switches, but I think they might be PWM driven, in which case, something would probably burn if I short their sources and drains.
Btw, 24V could be provided by a DC/DC boost converter or a simple bootstrap circuit. Either way, I can't find any
Did find some schematics by SMA, though:
http://dbhome.dk/sma/
but none contains any inverter related power electronics...

[jimmy274]

Ok, I've managed to trace gate drive voltage to the infamous toshiba tb62501f ic - it's pin 33. I've also noticed that battery charging doesn't seem to work (I didn't install drivers until now).
The big bad thing after all this is that I somehow managed to fry ADP3806, or even something more - I accidentally left multimeter measuring resistance (ohm range) and shorted ground to pin 24 of ADP3806. Looking at the schematics and datasheet, I really don't see what could have been fried, but I don't have 5V ALWAYS on MAX1631 anymore, which means, I don't have 3.3 and 5V switched power supplies either - in other words, the laptop is now dead
I'll try my best to trace what's been fried in the process tomorrow...

[gabrield]

Hi,
I was following your discussion full of hope, thinking you are my salvation.

I have this Thinkpad R50, which won't start anymore.
The system board is similar with T40, I think, anyway, I was following everything you said and matches very good.

Right now, I noticed F3 and F8 are grounded, so, I ask you kindly, where the 3.3 V is made, so I can trace the problem.

In few words I don't have 3.3 V neither on F3, or F8 and this fuses are grounded.

jimmy or poshgeordie, if you can help me with this...

[jimmy274]

Both 3.3V and 5V come from MAX1631 ic. It's near the system battery - pretty easy to find. When a voltage is applied to MAX1631 (Vdd), it first uses it's linear 5V regulator to supply logic with 5V voltage. After this voltage exceeds 4.5V or so, MAX1631 automatically bootstraps and switches on the 5V SMPS to supply voltage, along with 3.3V SMPS. Less dissipation with and more efficiency of the SMPS being the reason for that.
I have couple of boards with totally dead 5V ALWAYS, and I can't seem to find out the reason for it... I have one board which bumps the 5V ALWAYS to some 4-ish V when I put the AC in, but that soon falls down to zero. Almost as if there's a problem with the SMPS section... This is as far as I came, I'm hoping that you can do better. If you need schematics for T-series or anything else, PM me with your e-mail - I'll gladly pass it on. I think that ADP3806 might also have something to do with all of this, but I haven't been able to trace what.
BR,

Ivan

[gabrield]

Thank you for your quick reply, I already sent a pm to you.
I measured on MAX1631, 5V exists, but 3.3 no, and it is grounded, 0 ohm to ground.
What do you suspect ?

Later :
Based on schematics, could be D11, or D27, or C214 and C216.

If I can remove them, tomorrow, I will write here the results.

[gabrield]

Bad news, for me...

I remove the above, but still VCC3M is grounded.
In this case, I think I must trace down all the board where VCC3M is present, and remove every chip or whatever will be.

Any more ideas, please ?

[jimmy274]

What do you mean by "grounded"? VCC3M cannot be grounded, per say - only if a certain component had a massive burnout, then VCC3M could be grounded, in which case, you should be throwing away the board, since anything that actually burns through a resistor, diode, or cap had probably burned everything else. Another way is that the actual VCC3M trace got broken and somehow got short-circuited to GND, which is, let's face it, a long shot and in this scenario you should be looking at a couple of dead fuses, and there aren't any, as far as I could tell from your posts.
No, as I said - VCC3M cannot be grounded per say - you can read 0V from GND to VCC3M, but this probably means that a switching power mosfet has either a bad drive, or it has no drive at all, or simply - it's dead and not driving VCC3M as it should. First, these TPC8xxx or Fairchild Semi 4xxx are very ESD sensitive and could be fried by only touching them. Second - look at the left side of the application circuit in the datasheet for MAX1631 - it has 2 mosfets driving 3.3V SMPS - one from the positive rail (Vdd) to load - high side drive, and one from GND to load - low side drive. If one of them is dead, there's your problem. Check gate drive voltages, and then check the pin 26. You'll need an oscilloscope for any SMPS, since all SMPS are usually switched at 50-500 kHz. Since these babies are N-channel mosfets, you should see gate drive voltage pulsed at 3.3+V (maybe even bootstrapped to 5-6V), in regular/irregular periods at let's say 100-200 kHz. Pin 26 (LX3) should also be pulsed, accordingly with the gate drive - if it's not, either Q18 is dead or Q100 is dead. Multimeter is, of course, out of the question - this is a completely new level, requiring a low-end scope for everything.
Ok, even I hadn't got this far, but I would have if I had some free time. Since I've returned the dead boards... I'm not in the game anymore Power electronics is somewhat my specialty and it's notorious for being one of the most complex electronics there is - combining digital and analog, being super sensitive to component tolerances etc. Anything else you need - fire away, but don't count on quick responses, since I'm in a sort of transition period right now...
BR,

Ivan

[gabrield]

When I say grounded, I mean shorted to ground.
I did a lot of removes, I took out F3 and F8, but still shorted to ground.
I checked up all components receiving VCC3M, excluding big chips.
It seems to be all right, none of them is burned.
Intriging is that, measuring with multimeter, is around 5 ohm, on VCC3M and ground.

Next step, I will follow your extremely precious advice, and try to test with oscilloscope.

Anyway, I did took out entirely Q18, and still shorted the circuit on L4, and everywhere else on VCC3M.

About Q100, couldn't find it.

[jimmy274]

Ok, first, you're sure it's shorted? You've checked with a regular ohm-meter which gives you 0 ohms, and you've checked with a pulsed mode ohm-meter and that one also gives you 0 ohms? And of course, I hope you didn't use an ohm-meter on a live board. If this is true - then you could be facing the fact that a trace is broken and shorted in any of the 5-6 layers of the board (ok, not any, there's GND plane, that one's off). Second, if you have a component that was burned and thus making a short... well, let's just say that I wouldn't play with that - for any massive burnout there's an underlying cause, and it's usually it's chained to something else. It's like shooting in the dark...
Why would you remove F3 and F8? They're not line connected to the VCC3M - if I can recall, you have several driver mosfets before them...

[jimmy274]

Just one more thing - 5 Ohms is NOT 0 Ohms, therefore it's not shorted. It simply isn't there.

[gabrield]

Well,
First I must assure you that I did not use ohmeter on live board, and I used a regular one.

Second, on both F3 and F8 I have 0 ohm, regarded to ground.

According with schematics R50_U40 at page 19, F3 and F8 are direct connected with VCC3M, wich I noticed does not exists on my board, I only have VCC5M at Max chip.

Therefore, I assumed, is a short somewhere on VCC3M line.

[SMA]

Given the high number of components sitting between the 3.3V and ground, it can be very difficult to find a short there.
The only way of doing - that I know of - is to unsolder all possible components one by one.
It can be a chip that has shorted, but it doesn't have to be so. There is also a lot a decoupling capacitors to consider.

But, before starting to solder (or giving up) there is a few things to inspect visually.
Check all chips for cracks.
Look for signs of a liquid spill.
And if there is a chance that someone else might have been in there soldering, then look for shorts between the legs of any chip that looks
as though it has been soldered upon.

[gabrield]

Thank you for advises.
I will check all, but, there is a way to interrupt somehow the VCC3M circuit ?
I mean, remove a capacitor, or diode, or anything else, to isolate the short

As I noticed it appears to go in one of the layers inside...

[jimmy274]

Buy a new board!
I'll have a second glance at the schematics tomorrow, I'll let you know if I find some more info on what could be wrong. The problem is that we don't know the layer layout and this could be crucial - if one of the layers is bad, you can just throw the board away...

[gabrield]

Well,
All I know is that my thinkpad was running on the desk, I went out 5 min, after that, all was black, no power, no led.

Can the layer be broken in that situation ?

I wait for you to reply tomorrow, and I will consider buying a new one, some guy already offers me a T40 system board, should I check for a R50, or the T40 is better ?

Finally, I really want to fix my R50 board, it was a good one.

[jimmy274]

Yeah, similar thing has happened to a R51 that I had too. Plugged in the USB printer, and all gone... I've had the board for like 2-3 years now, and gave up on it.
T40/T41/T42 and R50/R51 all have the notorious GPU issue. You can read more about it on the forums, but I would stay away from any of those. R52 and T43 have this issue too, but it's fixed to some extent and with these you'll have a problem of fitting them since the boards are a bit different than the rest. In your case, I would save the money and buy a used R60/T60 - they're perfect and have no problems whatsoever.