Yesterday I picked up a Dell 30" monitor, model 3007WFP. It looks in good shape except the upper half of the screen has no backlighting. I googled around and found what little I could (no schematics to be had) and then took the covers off to have a look. The inverter board has 8 high voltage transformers on it and only the lower 4 are working. I'm from the mechanical world, not the electronic, but I have some experience. The board seems to be basically two pretty identical halves and I feel very fortunate that I can compare the working half to the non-working half or I would be in over my head immediately.

Logically, of course, I am looking for a pair of something since half the board works and half doesn't. It looked to me like the highest things in the chain are the two BD9886FV DC-AV inverter control ICs so I started comparing things there. I found a number of voltages that did not match from one to the other. All the readings are DC volts using my Fluke meter to chassis ground except the FET drivers where I used my scope. The working chip had square waveforms that looked pretty square to me aside from the fuzzy high-frequency noise I see all over. The non-working upper chip has no waveforms at all on those pins. The inverter board has no caps other than SMD, the power supply to the board is about 24 VDC and appears pretty clean on the scope. No visible clues on the board. I will appreciate greatly any informed opinions as to whether the problem might be internal or external to the chip, thanks in advance. Here is what the readings were with the significant differences highlighted:



Here is a scan of one of the chips and surrounding area (1280x1280 is very limiting!):



Thanks very much for looking.

Rob

 

Looking at R46 R96 and C45 (COMP2).. These are a mirror of R45, R97 and C46 ( COMP1).. There is no voltage detected at Comp1.. So test this area for a loss of the 4.3volts expected on pin 17..

 

Looking at R46 R96 and C45 (COMP2).. These are a mirror of R45, R97 and C46 ( COMP1).. There is no voltage detected at Comp1.. So test this area for a loss of the 4.3volts expected on pin 17..

Thanks for the reply. I am not anxious to start removing these tiny components yet so I did some testing in place just for comparison. By the way, the chip in the photo is the working one. First I tested for an obvious short on pin 17 to ground, it reads about 100k, the same as pin 18 and the same on both chips. One side of C45, C46, R96 and R97 is grounded so it appears no short circuits here.The resistances across R45 and R46 are 78k each and they are not directly grounded. All of these readings are duplicated on the corresponding components at both chips. I tried reading capacitance across C45 and C46, one direction reads 13.9 uF on each and reversing the leads reads 2750 uF. I don't know if it's significant but the corresponding caps on the problem circuit read 9.8 uF and 2750 uF.

Then I tested voltages powered up. R45 and R46 each have about 5.5 v on one side, dropping to 4.3 v on the other, as does the resistor corresponding to R46 on the non-working circuit. On the resistor corresponding to R45 it also has 5.5 v but then drops to .02 v on the other. Aha! I unsoldered that resistor fully expecting it to be open but it tests at 100k, correct by its marking. With it removed, the resistor corresponding to R97 tested at 330k, matching its marking. On the unlikely chance that the resistor had a cold joint I reinstalled it but nothing has changed.

With my limited knowledge it seems logical that with all the resistance readings matching that pin 17 is being held low internally. Any more suggestions, please?

Rob

 

One thing to look into.. The soft start seems to be low as well.. While this is almost certainly to do with the "error" from the COMP1 input, this also could be todo with the cap.. If this voltage doesn't rise, then the voltage reg will not kick in ( pin 21 )..

 

One thing to look into.. The soft start seems to be low as well.. While this is almost certainly to do with the "error" from the COMP1 input, this also could be todo with the cap.. If this voltage doesn't rise, then the voltage reg will not kick in ( pin 21 )..

OK, I measured pin 21 on both ICs, they match at 7.5 megs. I removed the cap for that pin on the non-working side and it reads about 100 nF, so not open. With the cap removed both pin 21's still match resistance. I think it's time to try replacing the IC even though I'm not looking forward to the job. There are no components on the underside of the board, so I plan to heat from underneath with a small heat gun throttled down with a controller to avoid overheating. Or I might just Dremel off the legs from the chip with a thin cutoff wheel for the removal stage, then I can unsolder the legs and clean up for the installation of the new chip.

Rob

 

OK, I ordered a new IC and also some solder paste (which I have never used before but it sounded like a great thing for SMD parts). Today I used hot air to unsolder and install the IC. I tried to use the solder paste but this particular tube had paste that was too thick and did not want to stick to anything either. It was impossible to apply to the tiny legs of the IC without just clumping it all over the place, not working for me at all. I expected it to be much "wetter" and want to cling to things. Anyway, I managed to get the chip on there and checked each leg for low resistance to its pad and no shorting between adjacent legs. Bottom line is: no difference at all, acts the same and all voltage readings are about as before. I tried to check all the legs that lead first to plated-through holes in the board to their eventual component, all ok as much as I can tell. I'm afraid I am at the limits of what I can do but I hate to give up. I'm sure if I had a schematic I could stumble along a little further. One thing that could help me is if someone who understands this chip could help me understand the voltage mismatches I am seeing on particular pins. I am unsure on some of them if the voltage is coming from the chip or going to it. If low voltage on a pin is coming from the board maybe I can work backwards from there a bit.

Rob

 

Solved! It's been a while but I didn't quit thinking about it. It turns out I made a rookie mistake by assuming that a discrete component like a transformer would be good if there were no visible signs of overheating, etc. I eventually stumbled across one of the transformers on the half that didn't work was open on the secondary winding and somehow the IC was sensing a voltage problem and refusing to run. The numbers printed on the transformer meant nothing to Google so I searched for inverter boards on eBay and found one for a different 30" Dell model that had identical-looking transformers. I took one off of it and soldered it onto my board and it works great now! Thanks, Ian, for your time and advice.

Rob