Repairing a Dell U2311Hb monitor

[mcgee]

Hi,

in a recent dumpster dive I was lucky and found a Dell U2311Hb Monitor. At home I plugged it in and it seemed that it worked but after 8,54 seconds (from the time of turning it hard on) the backlight goes off. Using a flashlight I found out that the image is still there so only the light turned off. So I thought that a cap has gone bad. So I opened the monitor and took a look at the power supply board (see photo attached). Visually there is no bad cap (leaked or blown). I don't have an ESR meter so I used a multimeter to check the resistance of the caps and they behaved all normal (rising resistance to mega Ohms). Then I retested everything and found out that after turning it initially on and waiting the 8,54 seconds until it goes black I can make it turn the lights back on for ~2 seconds when I change the input from VGA to DVI or to DisplayPort or vice versa. But this leads me to the conclusion that the power supply board is probably good since it can turn on the lights.

And another thing: on the image (right) you can see the CFL lights inverter and four silver contacts grouped into two groups. If I measure the resistance between any of this four contact points I get a resistance of 0 Ohm (or very, very low). Is this normal?

So do you have any ideas I could try to fix it?

Regards!

[PA0PBZ]

It could be the CCFL tubes worn out, not sure about this model but can you reach them? If they are discolored at the end that could be the problem.

[kripton2035]

you won't check any cap correctly with a multimeter...
build yourself an esr meter see my website below it's easy to build one yourself.
then you will be able to really check caps.
the multimeter checks the DC resistance, to measure esr you need to measure AC resistance.

[RGB255_0_0]

Unlikely all of the tubes will have gone. I just fixed a Samsung that had the backlight go off after 2 seconds. No AC from the transformer. £3 fix. So check the inverter circuit.

[PA0PBZ]

Unlikely all of the tubes will have gone.

Some monitors check the CCFL/inverter current and switch the inverter off when things don't look as expected. Been there, done that.

[RGB255_0_0]

Unlikely all of the tubes will have gone.

Some monitors check the CCFL current and switch the inverter off when things don't look as expected. Been there, done that.

Yeah, but still check the circuit. When I measured the transformer one of the windings was way off what the picture showed on eBay. Changed it and its backlight has worked flawlessly.

[PA0PBZ]

Yeah, but still check the circuit. When I measured the transformer one of the windings was way off what the picture showed on eBay. Changed it and its backlight has worked flawlessly.

Sure, I don't think we disagree, it's just that a couple of times I found that it was the tubes and not the inverter. Just poke your finger at the inverter output and if it starts to smell funny the inverter is fine 
_{(disclaimer: don't do that!)}

[RGB255_0_0]

I'm not really trying to argue you either. I just think it's wiser to check the easy bits first 

[Rasz]

ltimeter to check the resistance of the caps and they behaved all normal (rising resistance to mega Ohms).

doesnt mean anything, you need esr meter, or just swap all the caps blind

into two groups. If I measure the resistance between any of this four contact points I get a resistance of 0 Ohm (or very, very low). Is this normal?

depends on what you measured and how shit your meter is ;-)
primary is usually very low <1ohm, but secondary should be in xxx-xxxx ohm range
my pretty crap UNI-T 70a barely registers 0.1 ohm difference between measuring shorted leads and random primary

what can be bad
-power supply
-transformer. we know its not totally shorted - monitor turns on for a second, might be partially shorted
-mosfets, probably on the bottom of this pcb, again turn on means they still work, but could have bad RDSon due to prolonged overheating maybe?(guessing)
-one of the lamps - when you get monitor to turn on does the backlight look all white or are there any pink/dark spots?

turning off after few seconds is usually due to either lack of power from the supply or overcurrent protection in the inverter controller(again on the bottom side)

[mcgee]

Thank you for your answers.

The backlight looks "normal" when it is on. There are no dark or discoloured spots. I'll check the CCFL bulbs but therefore I need to disassemble the display again :-)

Any suggestions on how to check the inverter circuit?   

[Rasz]

are there any markings on inverter transformer? they are usually printed with black ink on black ferrite

[mcgee]

The only "marking" on the inverter transformer is the product number (I guess): 1J.26051.051 4311R 100511_

Connected to the CCFL outputs is an OZ9933GN on the backside. I attached a schematic from the datasheet which looks pretty similar to the situation I have on my power board. There is a PWM signal which controls the brightness. So I could replace the CCFLs with LEDs and us this for dimming the screen. I've done this in the past with an old 486 notebook where the CCFL has gone bad and it worked flawlessly. So I hope I could apply this "trick" also here. Only problem might be if the controller board gets a feedback (the reference schematic does not indicate this) from the CCFL and turns of the PWM. And I also want - in this case - disable the transformer for power saving reasons. Can I simply desolder it if I decide to apply my LED replacement?

[RGB255_0_0]

Measure the windings and they should be of the same resistance. Unfortunately I cannot find information on what it should be.

If not there are some on ebay/Alie such as https://www.aliexpress.com/store/product/Free-shipping-100-new-original-HS9016-network-transformers-SOP16/925936_32474949163.html

Or you could do your LED retrofit. Personally I'm not a fan of LED backlit screens due to poor uniformity compared to tube  YMMV

[mcgee]

The resistance of both coils is 650 Ohm. There are two lines of the PCB which go directly from the connector to the OZ9933GN and I assume that this are the ENA (enable) and PWM lines. I could measure those. Then we know if the controller shuts down the lights or if the signals stay active there is a problem with the tubes. But I have to stay away from the high voltage stuff.

I also think that the shown schematic in my last post is not quite right although it is in the manual for the OZ9933GN. There is some kind of feedback to the IC so that this schematic (attached) matches better.