Thinkpads Forum

When troubleshooting a "dead" display, a common question is to ask if there's an image still being shown on the LCD. You can usually tell if this is happening by shining a bright light directly onto the panel. If an image is there, then the attention normally shifts to several other areas: the lid switch, the inverter board, the ribbon cable, the CCFL (cold cathode fluorescent lamp, or backlight), and finally, the motherboard.

The inverter's job is to convert a small DC voltage into several hundred (or even over one thousand) volts AC. If it's not working, then there's no backlight. How do you know if the inverter is working? You'd probably need something more than your regular digital multimeter to be able to safely measure the output voltage. (I know, I tried and the end result was a loud "snap", some smoke and a burnt clip; the inverter survived). So, the usual advice is to swap out the inverter board with a known good unit and see if that helps. As it's a relatively cheap piece (~$10-$20) and generally available from eBay, that's been my suggestion up until now.

For the cost of a few dollars and a small amount of disassembly, you can verify that the inverter is working or not.

- Get a neon lamp from Radio Shack or electronic parts supply. The general part number is NE2. If the lamp is already installed in a housing, it needs to be removed. They look like this:



You can twist the two wires from the neon bulb together to make it easier to handle. Don't worry if you disconnect one or both wires; they are not going to be used anyway.

- Remove the LCD bezel. Generally, this means a couple of screws near the bottom of the panel and carefully pulling it away from the rest of the lid assembly. Some bezels may be secured to the bottom of the panel with double-sided tape so go easy. This should expose the inverter board. If your system locates the inverter somewhere else, consult the Hardware Maintenance Manual for the location.

- With the laptop on, hold the neon bulb with some insulated pliers up against the inverter board near where the connector is for the lamp. If it is working, you should see this:

56KB picture
37KB picture

This experiment was done on an X24 system. I have not verified this procedure with any other system at the moment but I do have plans to check it out on a T23 later.

If the neon lamp lights up, then everything up to, and including the inverter board is properly working. That just leaves the CCFL as the problem. If the neon lamp doesn't light up, then there still can be several reasons: the inverter is dead, the ribbon cable is broken or disconnected somewhere or there's a motherboard problem.

Note: I tried holding the lamp up against the X24 and T23 LCD bezels but could not see it glowing. Only when the lamp is almost in direct contact with the inverter board or the wires running to the CCFL, did it light up. This may be due to the age of the lamp I am using (at least 30 years old; a leftover from my days playing around with tube radios and TVs).

edit: Made the two pictures of glowing neon bulb clickable to larger images

I think this thread deserve sticky!! Hey..Ray.....do you have all this information compiled?? I would love to download them if you have!

Thanks

Ray, great idea ... can you give us links to bigger pics?

I motion to sticky this thread too!

The two pictures now are clickable for slightly larger images (800x600). The originals are 2048 x 1536 and are over 1MB in size. If you really want to see those I can email them to you!

@ambientscape: This is about all of the information I have. The hardest part is locating the neon bulb. Most of the ones I have seen lately are already inside a housing (barrel with a lens). Those are typically a few dollars. If you can find an electronics surplus parts dealer, they might have the NE2 bulb just by itself.

Great idea Ray!
I wish you had come up with it a few months ago.

BTW, these neon lamps come with some of the electric pet collars, such as Radio Fence, for testing the collar safely.

Splendid.

Actually, I have been thinking about this topic since I first noticed Ray mentioning his idea in a thread. Great idea.
I even came as far as to try with an oscilloscope, using the probe as antenna. It worked right away. Got a clear picture of the signal on the scope screen.
Did not even have to remove the bezel.
Sure, I know, that will not be doable for those without a scope. Then it is much better with a cheap neon lamp. I just wanted to mention it anyway.

One question - does the position of the brightness control have any influence on the amount of light from the neon lamp?
If it does, then it might be worth trying the lamp again with the brightness fully on and the bezel in place.

Would also like to know if this will work if the CCFL is broken. A broken CCFL may change the magnetic field from the coil on the inverter.

And just for the record. I do remember the days where repairing electronics, were just a matter of replacing some tubes. Those were the days.
In many years I have known that scopes could pickup various signals through the air, by I have never heard of this trick with a neon lamp before.

Just splendid Ray. Thanks.

Dont they use those neon lights in the cheap electrical test screw drivers? You could try holding the lamp end of one of those screwdrivers next to the inverter..

rkawakami wrote:<snip> I know, I tried and the end result was a loud "snap", some smoke and a burnt clip; the inverter survived. <snip>

/me chuckles. Does that ever sound so familiar to me! ... JDH

SMA wrote:I even came as far as to try with an oscilloscope, using the probe as antenna. It worked right away. Got a clear picture of the signal on the scope screen.

Ah, excellent thought! I did not think of that, but like you said, not everyone has an o'scope to play with.

SMA wrote:One question - does the position of the brightness control have any influence on the amount of light from the neon lamp?

Good question. Yes, the brightness of the neon lamp is directly proportional to the brightness control. Here's two pictures from a T23:

Bright setting (40KB)
Dim setting (34KB)
Camera exposures were identical between the two shots. You may also notice that I had to put the neon lamp on top of the inverter since the PC board has a copper shield and blocks the RF from the output coil. The only place I could get the lamp to light up on the side facing out is right at the end of the inverter board, next to the jack for the CCFL connector, but it was fairly dim. Again, it could be due to the age of the bulb I'm using. I'll try to get another one sometime this week.

SMA wrote:Would also like to know if this will work if the CCFL is broken.

Another good question! I'll have to try this out on a system with a burned out CCFL. I've got some around here but I used my "test bed" T23 for this try. (I keep a T23 mostly disassembled so I can swap parts into it and see if they work).

SMA wrote:In many years I have known that scopes could pickup various signals through the air, by I have never heard of this trick with a neon lamp before.

This was an old trick I picked up many years ago when dealing with the high voltage sections of TVs. I don't even remember where I heard or read about it.

@kensplace: Possible, but the intensity of the neon lamp glow is related to how close you can get the bulb to the output coil of the inverter. At least for the bulb that I'm using, I have to basically touch the glass directly up against the inverter board. Even the 2 or 3 millimeters that separate the X24 inverter from the bezel was enough for my lamp to not fire up.

SMA wrote:Would also like to know if this will work if the CCFL is broken. A broken CCFL may change the magnetic field from the coil on the inverter.

And here's a possilbe answer... yes, but only for a few seconds.

I unplugged the connector to the CCFL and fired up the system. There was a soft "scratching" noise coming from the inverter board for about three seconds and then all was quiet. During that time, the neon bulb was lit; afterwards, it was not. There appears to be a shutdown circuit in the inverter which will cause it to turn off if no load is present. I suppose the same thing would happen if the CCFL is dead, but still connected to the inverter.

So it appears that using a neon lamp is going to be a little limited in diagnosing a "no backlight" problem.

<Ray goes off to conduct some experiments; comes back with some bad news>

I have (had) three spare inverter boards for the T2x systems laying around here. I decided to plug them in and run them without connecting the CCFL to see if they also shut off. Bad mistake. Two of them lit up with flashing white lights (and I'm not talking about the neon bulb). During this time those inverters were also making the same "scratching" noise, however when I plugged the CCFL back in, those inverters are now dead. I'm sure that they were good before since I tested them several months ago.

So, word of warning... don't try running the inverter without having the CCFL plugged in. It might burn out the inverter like it did for two of mine.

Additional note: If you look closely at the pictures of the neon bulb atop the inverter you may notice that I was using a 13.3" T23; there's a little bit of room between the inverter and the LCD panel. I found another T23 (this one with a 14.1" panel) with a known backlight problem and ran the same test with the neon bulb. Upon power up, and for only about three seconds, the neon bulb lit up. Since there was no room between the panel and inverter, I held the bulb up against the connector at the end of the inverter. That tells me that the inverter was outputting (some) voltage but the CCFL did not light. I swapped inverters with the last spare one I have () and it did the same; neon bulb glowed orange for just a couple of seconds. I think that this still validates this test. If the neon bulb glows for even a few seconds but the LCD does not light up, then this points to a bad CCFL.