inverter problem in LG 1934S LCD flatron

[PlainBill]

Originally Posted by stevenps

The scope is an Owon PDS5022S (http://www.owon.com.cn/eng/pdsSeries.asp). The probe is a P6060 with both scope and probe set at x10. I just switched to DC coupling. I was attaching the ground clip to the case of the monitor, as the center screw hole is attached to cold ground and the board is screwed in to the case. I can try using another location, though it may be a little tricky.

I took new readings but I don't have time to post the values right now; I'll get that done later this afternoon. I moved my ground position to the negative leg of C501 and used DC coupling. I have two sets of measurements, one with and one without shorting the end of R506 to ground to keep the inverter running long enough to properly adjust the scope and let it get a good reading.

Great! I'll be looking at them later. Today was supposed to be my goof-off day, after running errands. It's now 10:20 AM and it'll be at least an hour before I can start on the errands. So much for a goof-off day.

PlainBill

[stevenps]

Originally Posted by PlainBill

Great! I'll be looking at them later.

Here are the results:

• D501: Pk-Pk = 1.760v, Freq = 45.72KHz, RMS = 4.064v
• D506: Pk-Pk = 2.400v, Freq = 45.56KHz, RMS = 663.3mv
• D510: Pk-Pk = 7.440v, Freq = 45.68KHz, RMS = 2.675v
• IC501 pin 11: Pk-Pk = 480.0mv, Freq = 91.41KHz, RMS = 780.0mv

Again, this time with R506 shorted:

• D501: Pk-Pk = 1.880v, Freq = 130.0KHz, RMS = 4.182v
• D506: Pk-Pk = 2.240v, Freq = 66.25KHz, RMS = 700.9mv
• D510: Pk-Pk = 7.920v, Freq = 45.63KHz, RMS = 2.889v
• IC501 pin 11: Pk-Pk = 700.0mv, Freq = 134.2KHz, RMS = 794.2mv

Images are again attached for reference. The view settings are not all the same, if you're looking at the waveforms; I think the offset ones are only offset by 1 div. I'm not sure what to make of the differences in frequency measurements between having the end of resistor R506 shorted to ground, and leaving it alone. The scope seems to have trouble measuring the frequency of the short pulses the backlight stays on for when the resistor is not shorted, but it's possible (maybe?) that the frequencies actually are different during that very short time right after the inverter turns on.

[PlainBill]

These numbers seem to confirm what I suspected, D506 (OP1) and D510 (OP2) should have approximately the same voltages. Compare the resistances across R524 and R527; they should match to within 5%.

I'm going to have to rewrite the data in a table form. Right now I'm like three phase AC - jumping between suspecting a bad CCFL, a bad diode, or a bad transformer. THAT, however is going to have to wait for a day or two.

PlainBill

[stevenps]

Originally Posted by PlainBill

Compare the resistances across R524 and R527; they should match to within 5%.

I checked in-circuit, and R524 = 469 ohm, R527 = 467 ohm (both within 5% of 470 ohm).

Originally Posted by PlainBill

I'm going to have to rewrite the data in a table form. Right now I'm like three phase AC - jumping between suspecting a bad CCFL, a bad diode, or a bad transformer. THAT, however is going to have to wait for a day or two.
PlainBill

If it would help, I have a second monitor of this same model with the same symptoms. I only brought one home, but I can grab the other set of parts to see if swapping CCFLs changes anything. I also have a kit for a Dick Smith shorted turns tester that I could build - I just found all the parts that disappeared after my last move, except for the top cover. Let me know if either of these things would be helpful in diagnosing this problem and I can try to take care of it tomorrow.

[PlainBill]

Originally Posted by stevenps

I checked in-circuit, and R524 = 469 ohm, R527 = 467 ohm (both within 5% of 470 ohm).



If it would help, I have a second monitor of this same model with the same symptoms. I only brought one home, but I can grab the other set of parts to see if swapping CCFLs changes anything. I also have a kit for a Dick Smith shorted turns tester that I could build - I just found all the parts that disappeared after my last move, except for the top cover. Let me know if either of these things would be helpful in diagnosing this problem and I can try to take care of it tomorrow.

By all means build up the shorted turns tester, I the problem is a bad transformer, but I ALWAYS suspect shorted turns.

The light just went on. Let's assume a simplified version of this inverter circuit. It has multiple transformers, CCFLs, etc. Each transformer has one end of the secondary grounded. The other end goes to a voltage divider whick allows measuring the output voltage and to a CCFL. The other end of the CCFL goes to ground through a resistor, which allows measuring output current. We will further assume that the inverter system is shutting down due to a fault.

Now if we compare voltages and currents and we find one transformer has a higher output voltage than the other, and the current through one CCFL is higher than through the other, we can come to several conclusions.

1. If the transformer which has higher output voltage has a higher current it is probably operating normally.

2. If the transformer which has a lower output voltage has a lower current, the transformer is probably defective.

3. If the transformer which has a lower output voltage has a higher current, the CCFL is probably aging and drawing a higher current.

4. If the transformer which has a higher output voltage has a lower current, the CCFL wiring may be bad.

Of course, there are several problems - we don't know 'normal' voltage or current, and this assumes all parts of the measuring system are equivalent. A further complication is the fact that one transformer drives two CCFLs. Swapping the two CCFL leads on a single transformer doesn't really change anything - the CCFLs are in series.

Definitely, a second monitor, even one with the same problem will be a big help.

PlainBill

[stevenps]

Originally Posted by PlainBill

By all means build up the shorted turns tester, I the problem is a bad transformer, but I ALWAYS suspect shorted turns.

I built the kit and it seems to be working. I'm winding up the sample transformer to test it with now. I have a board from an old power supply I can steal transformers from to test with also.

Originally Posted by PlainBill

Definitely, a second monitor, even one with the same problem will be a big help.

I brought the other monitor home from work. I'll hook up the inverter board I've been working with to the lamps from the other monitor and see if shorting the same resistor keeps the backlight on.

[stevenps]

Originally Posted by stevenps

I built the kit and it seems to be working. I'm winding up the sample transformer to test it with now. I have a board from an old power supply I can steal transformers from to test with also.

The test winding and the old transformer both light all LEDs. I tried using the inverter transformers in-circuit and found that the high-voltage side lights the 3 red LEDs but not the yellow or green ones. I'm having trouble getting a good reading on the low-voltage side though - I usually get either no LEDs, one red LED, or up through one of the green LEDs depending on which of the many pins I pick. What should I see on the shorted turns tester here? Can I get an accurate reading in-circuit, or should I desolder the transformers?

[PlainBill]

Originally Posted by stevenps

The test winding and the old transformer both light all LEDs. I tried using the inverter transformers in-circuit and found that the high-voltage side lights the 3 red LEDs but not the yellow or green ones. I'm having trouble getting a good reading on the low-voltage side though - I usually get either no LEDs, one red LED, or up through one of the green LEDs depending on which of the many pins I pick. What should I see on the shorted turns tester here? Can I get an accurate reading in-circuit, or should I desolder the transformers?

Desolder the transformer. The ring tester I have uses an analog meter so it's possible to say something like 2.5 rings. With the LEDs it's a little more difficult.

Do label them so you know where they came from.

PlainBill

[stevenps]

Originally Posted by PlainBill

Desolder the transformer. The ring tester I have uses an analog meter so it's possible to say something like 2.5 rings. With the LEDs it's a little more difficult.

Do label them so you know where they came from.

PlainBill

Swapping panels and backlights didn't change anything. Shorting the end of R506 to ground allowed the backlight to stay on again. I'm running the monitor with the panel disconnected, if that matters; the (LVDS?) ribbon cable won't reach.

I desoldered both transformers and tested them with a multimeter and the ring tester. They both measured the same on the ring tester, with the high voltage side of both returning 3/8 rings and the two windings on the low voltage side of both returning 5/8 rings. There is little resistance (< 0.6 ohm) between pins 1-4 and 3-6 on both transformers. T501 measured 621 ohm between pins 7-8, and T502 measured 642 ohm. All other pin combinations read open.

When I put the transformers back on, I can swap them to see if the problem follows a transformer instead of the remaining parts (assuming that I don't have two bad CCFLs in the same position on both monitors). Should I do that?

[alexanna]

By chance is it the upper bulbs/connectors that you’re having problems with?
I have had quite a few widescreen monitors that have had problems with the upper bulbs giving a problem, usually a return wire just barely making contact, I guess it’s because of the heat rising inside of the panel. Could you try plugging in the bottom bulbs of your two panels to your inverter and see if it makes a difference?

[PlainBill]

Originally Posted by alexanna

By chance is it the upper bulbs/connectors that you’re having problems with?
I have had quite a few widescreen monitors that have had problems with the upper bulbs giving a problem, usually a return wire just barely making contact, I guess it’s because of the heat rising inside of the panel. Could you try plugging in the bottom bulbs of your two panels to your inverter and see if it makes a difference?

Great idea!!!

PlainBill

[stevenps]

Originally Posted by PlainBill

Great idea!!!

PlainBill

I'm already doing that. To get access to the back of the board while powered on, I had to flip the back of the monitor over.

[stevenps]

Originally Posted by stevenps

I'm already doing that. To get access to the back of the board while powered on, I had to flip the back of the monitor over.

I think I read the suggestion a little too fast. I haven't tried using the bottom CCFLs on both monitors, but I need to reattach the transformers to the board first and I was waiting to hear if PlainBill thinks that swapping them would be a good idea before I solder them back on.

I also gave my second monitor (all parts, including the IP board) another try this morning and much to my surprise, it seemed to work fine. I had swapped the CCFLs though as I had inverted the board to gain access to the components, so I could very well have at least one CCFL in less-than-perfect shape.

[PlainBill]

Originally Posted by stevenps

I think I read the suggestion a little too fast. I haven't tried using the bottom CCFLs on both monitors, but I need to reattach the transformers to the board first and I was waiting to hear if PlainBill thinks that swapping them would be a good idea before I solder them back on.

I also gave my second monitor (all parts, including the IP board) another try this morning and much to my surprise, it seemed to work fine. I had swapped the CCFLs though as I had inverted the board to gain access to the components, so I could very well have at least one CCFL in less-than-perfect shape.

Yeah, go ahead and swap them. I see three possible results - It now works the same, it works better, or it works worse. Each of those will tell us something.

It would have been nice if it turned out to be a bad solder joint on a SMD cap, but it's beginning to look like a CCFL. At least it isn't an unobtainable transformer.

PlainBill

[stevenps]

Originally Posted by PlainBill

Yeah, go ahead and swap them. I see three possible results - It now works the same, it works better, or it works worse. Each of those will tell us something.

It would have been nice if it turned out to be a bad solder joint on a SMD cap, but it's beginning to look like a CCFL. At least it isn't an unobtainable transformer.

PlainBill

Swapping transformers didn't change anything, and shorting the end of R506 to ground still kept the backlight on.

After all that diagnosing and measuring, it looks like one of the simpler explanations was the right one. I tried alexanna's suggestion of using only the lower CCFLs from both panels, and the backlight stayed on! Now to find a source for 19" wide CCFLs...

While thinking about this, I was wondering if there's a better way to test CCFLs to see if they're any good. I have a case mod kit that has a CCFL and an inverter that is probably good enough to diagnose a completely failed CCFL (if it's safe to use one of these on LCD CCFLs), but other threads I've read suggest that maybe these aren't suitable for this kind of testing. Is there a better way to diagnose bad CCFLs than replacement? If not, I'd better order a couple of extras for testing in the future. Also, how do I determine which of the 2 upper CCFLs is the bad one? Is this going to be trial and error also?

[alexanna]

Maybe you will be lucky and just find a wire you will need to re solder.
I have had very good luck just testing with known good bulbs, and if the 2sec to black goes away with the test bulbs it’s time to go inside the panel.
You will probably be able to see the failed bulb,If you cannot,Plug them in after there out of the pannel
Al.

[stevenps]

Originally Posted by alexanna

Maybe you will be lucky and just find a wire you will need to re solder.
I have had very good luck just testing with known good bulbs, and if the 2sec to black goes away with the test bulbs it’s time to go inside the panel.
You will probably be able to see the failed bulb,If you cannot,Plug them in after there out of the pannel
Al.

I took the whole thing apart (which was not particularly easy) and found that one of the tubes is a bit more black on the HV end than the others. Its return wire also has signs of being hot as there is what appears to be a scorch mark on the rubber cap, and the wire was loose. I bought some new CCFLs on eBay, and I'll update again when I get them installed.

When you say "test bulbs", do you have a spare set with the connector and such on them? I had to completely disassemble the backlight (unstuck the wires, removed the connectors, cut off the heat shrink tubing, pulled the rubber caps off, and I will have to unsolder the bad tube and solder a new one in and then put it all back together), and it seems like a lot of work to test one lamp.

Also, I didn't see an answer so I'll ask again: Is a case mod cold cathode kit sufficient for most testing of inverters (using the kit's tube) and CCFLs (using the kit's inverter)?

[PlainBill]

Originally Posted by stevenps

I took the whole thing apart (which was not particularly easy) and found that one of the tubes is a bit more black on the HV end than the others. Its return wire also has signs of being hot as there is what appears to be a scorch mark on the rubber cap, and the wire was loose. I bought some new CCFLs on eBay, and I'll update again when I get them installed.

When you say "test bulbs", do you have a spare set with the connector and such on them? I had to completely disassemble the backlight (unstuck the wires, removed the connectors, cut off the heat shrink tubing, pulled the rubber caps off, and I will have to unsolder the bad tube and solder a new one in and then put it all back together), and it seems like a lot of work to test one lamp.

Also, I didn't see an answer so I'll ask again: Is a case mod cold cathode kit sufficient for most testing of inverters (using the kit's tube) and CCFLs (using the kit's inverter)?

A 'case mod kit' is one method. If you hook up the CCFL from the kit to the inverter in a problem monitor (an adapter would be necessary) and the 'two seconds to black' problem goes away, the problem is most likely the original CCFL. If the inverter from a case mod kit fails to light a monitors CCFL, the CCFL is bad or there is a wiring problem. If the monitor's inverter fails to light a ccfl from a case mod kit, the inverter is most likely bad.

I use 'probably' and 'most likely' because the monitor inverter closely checks the voltage and current through the CCFL. A CCFL with different voltage or current characteristics could cause shutdown.

PlainBill

[alexanna]

Originally Posted by alexanna

Maybe you will be lucky and just find a wire you will need to re solder.
I have had very good luck just testing with known good bulbs, and if the 2sec to black goes away with the test bulbs it’s time to go inside the panel.
You will probably be able to see the failed bulb,If you cannot,Plug them in after there out of the pannel
Al.

What I am about to say is going to open the can of worms but here goes.
When you order replacement ccfls you’re asked the length and the diameter. Is going from a 17” to a 22” cclf going to change the current flow enough to cause shut down of the ccfl circuit,I do not know.
I can tell you what I have had success with though .Its has been using 17” bulbs for testing monitors up to 22” wide.
I use a panel with the LCD removed that has good 17” bulbs in it with the proper connectors to plug in to the suspect inverter. I have been able to identify bad ccfls in a number LG 22w-Samsung 19” to 22”w and HP 19” wide monitors, if a bad ccfl is suspected I do open up the panel and replace/repair the bad ccfl.
I may have been extremely lucky but it has worked for me.
I wonder if a case study may be in order. I will continue to test this way and if I run into a particular monitor that this does not work on I will post a note about it. Maybe other members could help. It sure would save a lot of time on diagnostic.
Al.

[stevenps]

Originally Posted by alexanna

What I am about to say is going to open the can of worms but here goes.
When you order replacement ccfls you’re asked the length and the diameter. Is going from a 17” to a 22” cclf going to change the current flow enough to cause shut down of the ccfl circuit,I do not know.
I can tell you what I have had success with though .Its has been using 17” bulbs for testing monitors up to 22” wide.
I use a panel with the LCD removed that has good 17” bulbs in it with the proper connectors to plug in to the suspect inverter. I have been able to identify bad ccfls in a number LG 22w-Samsung 19” to 22”w and HP 19” wide monitors, if a bad ccfl is suspected I do open up the panel and replace/repair the bad ccfl.
I may have been extremely lucky but it has worked for me.
I wonder if a case study may be in order. I will continue to test this way and if I run into a particular monitor that this does not work on I will post a note about it. Maybe other members could help. It sure would save a lot of time on diagnostic.
Al.

Originally Posted by PlainBill

A 'case mod kit' is one method. If you hook up the CCFL from the kit to the inverter in a problem monitor (an adapter would be necessary) and the 'two seconds to black' problem goes away, the problem is most likely the original CCFL. If the inverter from a case mod kit fails to light a monitors CCFL, the CCFL is bad or there is a wiring problem. If the monitor's inverter fails to light a ccfl from a case mod kit, the inverter is most likely bad.

I use 'probably' and 'most likely' because the monitor inverter closely checks the voltage and current through the CCFL. A CCFL with different voltage or current characteristics could cause shutdown.

PlainBill

Sounds like it might be worth my time to order a couple of extras next time I buy and just keep them around. I usually end up fixing Samsungs with bad capacitors in the power supply but that are otherwise fine, so I don't have very many spare parts hanging around. If I attach wires to a couple of bare bulbs, I could probably use these small jumper wires I have (the kind designed to fit a pin that's about the size of a component leg) to plug them in to the inverter instead of the usual tan plastic connector.

I bought a case mod kit a while ago for this purpose, but I wasn't sure if the results I got would be reliable after reading some other posts. It sounds like I can probably use spare CCFLs to test questionable inverters, and the case mod kit's inverter to test questionable CCFLs. After doing all this troubleshooting, it might be worthwhile for me to compile a list of things to check and some suggestions for how to test them so next time I have a problem I can come with a bunch of information ready to go.

I'm waiting on my new CCFLs now, so I'll post back after I get the replacement installed. Thanks for all the help you've given me so far!