T42p inverter/backlight/fuse/? problem

[ajkula66]

Noise usually equals inverter...

[rkawakami]

That certainly sounds like something is causing the inverter to go into a self-protection mode. I would think that brighter CCFL = increased current draw through the inverter.

[sbjoe]

Just managed to replicate, roughly, the result by turning brightness up all the way. Panel went dark fairly soon after and came back after I turned brightness back down to 3 bars and turned off the monitors for a few seconds. The monitor doesn't seem at all overlit as compared to the external lcd that's sitting right behind it, but the popping noise recurred similarly and might well be coming from the vicinity of the inverter. I'll call Lenovo tomorrow and see if I can confirm that the 39T0359 is indeed the specified replacement for the original 93P4661. At the moment the 4661 is holding its own at 3 bars. The garage, where I've got it set up for the duration of this issue, is cooler than it was earlier when it went dark at the same brightness setting so ambient might be part of this equation as well. Haven't actually charted ambient but I have the impression that the panel does better at cooler temps.

[ajkula66]

Proper inverter card for a 15" SXGA+/UXGA is 93P4161 according to HMM, when it comes to T42p.

The newer inverter that you were referring to is RoHS compliant version (which doesn't exist for a T42/p) that was specified for T43 in its 15" XGA form. Wrong one for your purposes.

The correct RoHS compliant inverter for 15" SXGA+/UXGA T43/p is 39T0361. That should work with no issues, if the inverter was your problem to begin with. I also happen to have one in a still sealed box if you decide that you need it...but read this post to the end first.

Now, I know for a fact that LCDs between these (both ID Tech, but with different FRUs) are interchangeable since I've swapped them. I see no reason why an inverter for T43p would not work properly with a T42p...

My assumption is that the backlight (which is on its way out) is draining the inverter too much and causing all the havoc...

[sbjoe]

Thx George. Seems that the friendly sales guy at Lenovo led me down the garden path, but hopefully no further harm done than 50bux wasted. I don't count the time spent entirely a waste. From what I've been able to glean it seems that panel replacement would be in order...? It seems an outfit called Moniserve exists in my general vicinity & they specialize in backlight replacement. Would I do well to dismiss that alternative out of hand - that or ordering up the light and having a go myself absent a clean room? Ifpanel replacement turns out to be the preferred cure I suppose new would also be advisable unless I'm looking for a very short term stopgap...? Ebay as good a source as any?

Ray, I don't recall that you've said as much flat out and I do understand there are no guarantees here, but I have the impression that you concur in the faulty backlight diagnosis even absent the red screen. That more or less correct?

[ajkula66]

I have a 15" T42p with a sickly motherboard but supposedly good LCD coming today...can PM you the details if that indeed proves to be the case...

Having done a backlight replacement myself once a few years back, I stay from it nowadays. Moniserv has received mixed reviews on their work. Personally, I send my laptops to James (jamiphar) when I need CCFLs replaced.

The choices are obviously yours, I'm just sharing my experiences...

Good luck.

[sbjoe]

Sure, George. I've not eliminated any options as yet. Think I'll wait for those neon bulbs before deciding anything just on the off chance they might provide *some* illumination. I could do without the machine for some period although direct pnp would of course be preferable.

[rkawakami]

If you are referring to my use of a neon bulb to determine if the inverter is working, then what you've done so far in the way of diagnosing the problem, may not add any more information than what you already have. The neon bulb will light up (if placed close enough to the inverter) as long as the CCFL is also operating. If you either have the correct inverter and the CCFL is drawing too much current or you are using the wrong inverter for the CCFL, then the neon bulb will shut off when the inverter's self-protection circuit kicks in.

[sbjoe]

Just got home to find a package of NE-2's in the mail. One of them held up next to the bezel at the output end of the inverter does indeed light up, seeming to increase in brightness as I increase screen brightness. Goes out as expected when I shut the monitor down with the FnF3 combination. Also shut the monitor down by leaving it at max brightness for a couple of minutes at which time the NE-2. also expectedly, went out. Does this or anything else I might reasonably do point clearly to the source of the problem? Was anything gleaned from this exercise other than to demonstrate that the inverter is indeed capable of exciting the NE-2? Should I try yet another inverter or break down and spring for a replacement panel? Anything else to consider?

[sbjoe]

Is measuring current draw reasonably possible? I've got a Fluke 36 which looks to me to be something of a horse for this application, but I'm more than willing to give it a try if the result might be revealing or to look for someone with a more appropriate instrument and the knowhow to use it.

[mlinchits]

I am having the same problem of the screen not staying lit for more than a minute at a time. Reducing brightness help the screen stay lit for much longer. Perhaps the issue is the lcd and not the inverter (see link below):

http://www.pluralsight.com/community/bl ... 18114.aspx

[sbjoe]

I can commiserate. Guess I'm just trying to find a bit more in the way of certainty before I decide to either spring for a new screen or a new machine. I'd seen that thread before and I have to say it doesn't inspire me with a lot of confidence.

[Brad]

If you like I can send you a 100% working LCD assembly so you can continue your troubleshooting.

Brad

[sbjoe]

Just got off the horn with Lenovo technical support. The girl there was insistent that the correct replacement part for 93P4161 is 39T0359. I pressed her on this a bit but she wouldn't budge. Online parts ordering page, brings no joy with either number but does list 39T0361 at a not unreasonable price. I'm back planted pretty firmly on the fence not having a clue which way to jump. ;(

[ajkula66]

Hmmmm...I suggest you take Brad up on his generous offer...

At this point I'm fairly certain that you need a new LCD...or a backlight replacement procedure...

There's another not-so-bright scenario but to eliminate that possibility you need a known good LCD...yes, I am talking about the motherboard...

[Troels]

Look into checking that the LCD cable is not damaged in any way ... the Y junction, where one parts goes to the inverter board is extremely easy to tear/rip...

But as CCFLs get older they draw more and more current. Actually, if the inverter was not current limited, the CCFL will definitly heat up to the point it could destroy the LCD, or melt the glass at least. So, CCFLs actually do work forever.. it's just a question of how much current they are allowed to draw

When your screen lights up, and has done for for a while, does it get hot near the bottom right and left of the LCD, where the CCFL would be at? If this is the case, the CCFL is definitely defective. The red hue which means that the CCFL is "about to die" is mercury which has more or less depleted in the phospor layer near the electrodes of the CCFL. When you apply a high voltage to mercury, it will emit light in the red and UV sprectra. Phosphor lights up bright white when it is "hit" by UV rays.

[sbjoe]

One of the reasons I've not yet sprung for a replacement panel is that there is absolutely no reddish hue at all to the screen, which btw has been rock solid for the last couple of days at a reduced brightness setting. The monitor's been powered on for the last hour or more and I don't notice any warming anywhere around the periphery. I'd bet the ranch though that if I were to turn the brightness up to max it would go dark within a couple of minutes.

[sbjoe]

Can anyone confirm a backlight failure that was not accompanied by a reddish tinge to the screen? I just put the bezel back on and can't say there's any evidence of the backlight overheating, at least no evidence in the form of noticeable heat coming from anywhere around the periphery of the screen. I'm thinking of trying another inverter at this point. Is the 39T0361 that you mentioned having available, George, or had I just as well order from IBM? New from them is not at all costly.

[ajkula66]

I've seen the LCDs where the CCFL just died overnight, with no red hue warning. Not that often, but it does happen.

Inverter is still there, new in box, and I'm certain that I can give you a better deal than IBM parts...

Now, let's reverse the whole thing: do you have access to another T4x/R5x unit that you could test the LCD with?

[sbjoe]

If the backlight were totally dead I'd probably be more inclined to dive into lcd replacement, but it seems to work perfectly well most of the time as long as I hold the brightness down. OTOH, I'd just as soon not dally until it is totally dead. If I had ready access to a compatible test bed I'm sure I'd have explored that option already. You seem fairly convinced that it's likely the backlight anyway, but I'm not too inclined to buy a used pig in a poke when the premium for new on fleabay isn't really all that great, assuming that there's decent quality to be had in that sty. I don't see any of what I take to be the usual signs of a detached gpu. So trying out the correct inverter sounds like fairly cheap insurance against it being something else. Unless you can think of a better option I'd be pleased if you'd PM me with the particulars on what it'll take to get you to part with your nib inverter.

BTW, when I had the machine apart, I didn't have any thermal grease. I just tried to redistribute what was already there. Was that particularly shortsighted of me? The bottom of the case doesn't seem to get particularly hot, just warm adjacent to the cpu.