Announcement

Collapse
No announcement yet.

Panasonic TH-42PD50U random pixels, some slowly disappear

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    Panasonic TH-42PD50U random pixels, some slowly disappear

    I got a Panasonic TH-42PD50U 42" plasma TV that has random pixels that overlay the image (including OSD) all over. Over time they disappear as you watch, letting the image come through where they were, though most stay stuck. The image in the areas underneath them looks sort of right, maybe tinted. Any ideas on where to start? I might open it and try applying slight pressure to various boards, and see whether anything looks amiss. Thanks.

    Also, problem occurs moment it's powered on, no change over time. Previous images seem to linger slightly in these messed up regions. And occurs for analog and digital inputs.

    One other thing, when I ran in high picture mode with snow screen, it seems to improve somewhat. But then switching to lowest picture (brightness), pixels start to become unresponsive again.
    Attached Files
    Last edited by blargg; 11-03-2012, 08:11 PM.

    #2
    Re: Panasonic TH-42PD50U random pixels, some slowly disappear

    verify voltages coming out of psu

    Overall I suspect D board then A (unless it is integrated, not familiar with PD)

    Comment


      #3
      Re: Panasonic TH-42PD50U random pixels, some slowly disappear

      I measured all output pins from P board (switching supply that connects to line voltage). Nothing looks really out of whack. A couple of pins seemed floating which should be power pins, marked with ******, but it may have been that I couldn't get the needle into the plug far enough. I'm going to re-measure those two tomorrow. On the schematic, these are connected to the adjacent ones with the same voltage, so it's unlikely that this is a sign of anything.

      I noticed that there's a separate DC-DC converter board with some low-power switching converters. I'll check the outputs of that tomorrow as well.

      VDA is spec'd at 67V +/- 1V
      Vsus is spec'd at 175 +/- 1V

      First col is pin #, second is my reading (?? = floating, couldn't get stable reading), third is name from service manual. It leaves out a lot of pins, which I'm assuming are ground if not listed. I put ? for ones not listed in the schematic.

      P12
      1 67.00 VDA
      2 15.77 +15V
      3 00.00 ?
      4 13.70 STB_PS

      P23
      1 ??.?? +15V ******
      2 15.77 +15V
      3 00.00 ?

      P11
      1 ???.? ?
      2 176.5 Vsus
      3 ???.? Vsus ******
      4 000.0 ?
      5 000.0 ?
      6 000.0 ?

      P2
      1 176.5 Vsus
      2 176.5 Vsus
      3 ???.? ?
      4 000.0 ?
      5 000.0 ?
      6 000.0 ?

      P5
      01 11.78 +12V
      02 11.78 +12V
      03 00.00 ?
      04 15.77 +15V
      05 00.00 ?
      06 11.79 +12Vs
      07 11.79 +12Vs
      08 11.79 +12Vs
      09 00.00 ?
      10 00.00 ?
      11 00.00 ?
      12 15.77 ?

      P25
      01 11.78 12V
      02 11.78 12V
      03 ??.?? FAN CONT
      04 00.00 ?
      05 00.00 ?
      06 00.00 ?
      07 04.78 +5V
      08 00.00 ?
      09 04.78 +5V
      10 04.88 STB5V
      11 00.00 ?
      12 00.00 ?
      13 02.65 F STBY ON
      14 00.00 FAN SOS
      15 00.00 ECO_ON
      16 00.00 PS SOS
      17 03.20 PANEL MAIN ON
      18 00.00 ALL OFF
      19 ??.?? ?
      20 ??.?? ?

      P10
      1 13.59 F STB 14V
      2 13.59 F STB 14V
      3 13.59 F STB 14V
      4 00.00 ?
      5 00.00 ?
      6 00.00 ?

      Comment


        #4
        Re: Panasonic TH-42PD50U random pixels, some slowly disappear

        Looks like sc board issue specifically in the erase pulse drive.
        Did I leave the soldering iron on?

        Comment


          #5
          Re: Panasonic TH-42PD50U random pixels, some slowly disappear

          DC-DC converter board voltages looked fine.

          I tried adjusting some of the voltage pots and varying the 67V, 156V, 12V, etc. a bit, but no visual effect.

          Today when I turned it on with lowest brightness, it was almost all black except for a little bit at the bottom and I thought it was damaged even more. But turning to high brightness, it slowly raised like a curtain, finally showing snow over most of the screen, with about 5% random pixels not working, and some black areas near the left and right sides.

          I thought it might be heat-related, so I let it cool, then ran a hair dryer over various boards. No effect.

          I took a picture of the pattern, turned it off for a while (so it could cool), then turned it on, and it was still the same pattern. Something has a memory effect in it, and it's not temperature. It seems that the only possible places for this memory is in the plasma panel itself (capacitance of each element? mercury distribution/vaporization?) or the RAM for the frame buffer in the digital board. Since the effect is highly dependent on brightness, I doubt it's the RAM.

          I'm trying to explain the pattern on screen and why it tends to affect certain areas more. My best guess is that a driver board isn't putting out the right signals, so is running the panel in a way that tends to expose slight differences between each pixel, even though these differences aren't a problem normally since they're being run away from a region that amplifies these differences.

          Much else I can do with only a DMM?
          Attached Files
          Last edited by blargg; 11-04-2012, 02:16 PM.

          Comment


            #6
            Re: Panasonic TH-42PD50U random pixels, some slowly disappear

            Plasma panels do have memory. The nitrogen in the plasma panel creates hysteresis which can lead to temporarily stuck pixels. It definitely looks like the erase pulse isn't working. There's also a vague possibly of an SS fault, though it doesn't show all the characteristics.
            Please do not PM me with questions! Questions via PM will not be answered. Post on the forums instead!
            For service manual, schematic, boardview (board view), datasheet, cad - use our search.

            Comment


              #7
              Re: Panasonic TH-42PD50U random pixels, some slowly disappear

              When you say erase pulse not working, do you mean the signal to the SS board, or just the final pulse to the panel itself? And when you say SS fault, do you mean SS board? It's my understanding that the SS board is where the erase pulse is handled, along with the other "blunt" pulses (ones that go to the entire panel at once).

              I tried grounding each of the four UEH, USL, USH, and UML (after some resistor buffers, so I wasn't shorting the signals) to see what would happen, to get a better idea as to what the SS board was doing. After the UEH one, the whole screen got "erased" to black and only a few vertical lines at random points started "showing through". Even after a while, these only grew some and I was worried. After an hour or two, they grew vertically and expanded some horizontally, but still weren't filling in lots of areas. Waited overnight, same. I was worried that I had damaged it more. Today I messed with the various voltage trimmers, including some unlabeled on the SC board, and over time was able to fill in the whole screen to working how it was previously. I think I found that with the SC voltage higher (my guess as to what they adjusted), it got rid of almost all the dots for a while. I also found that lowering the erase voltage (VE) didn't seem to affect things.

              I've read a bit about how a plasma works. My basic model is that each cell is basically a capacitor that acts as a binary memory. To display something, it first charges the respective cells to around 67V, which won't cause any plasma. Then, it pulses a higher voltage across all cells at once, such that only the charged ones will reach plasma state. This also keeps the charged ones charged and the off ones not charged. It flashes them several times in a row for some duration (which varies for the cycle, depending on the binary weight it's currently at). Then it erases them by discharging them with a strong negative voltage ramp.

              Since many are sticking at black, this suggests that it's the charging that's not working, rather than discharging (erase). Or maybe if they're erased wrong, they can't be charged correctly either?

              I also don't have an explanation for the extreme memory effect, and how long it took to unstick the pixels from black so they all generally worked again. I noticed that this unsticking effect seemed to only run upwards along a column, virtually never sideways. If a column was working near the bottom, it would slowly grow upwards one pixel at a time. I'm wondering whether this could at all be due to panel fatigue. I found the service mode and this has 5982 hours on it and 2089 power cycles (which is quite a bit over its 7 years of life).

              I found an SS board on eBay for $13 shipped, but I'd like to have some confidence that this is really the cause, and not say the SC board. I was thinking that I might be able to verify the waveforms into the SS board using an old monochrome CRT or even just an LED swung quickly back and forth as an o-scope, since the signals appear to be in the millisecond-timing range. It seems more likely that any failure would be in one of these boards, given that they do all the heavy lifting.

              Comment


                #8
                Re: Panasonic TH-42PD50U random pixels, some slowly disappear

                Originally posted by blargg View Post
                When you say erase pulse not working, do you mean the signal to the SS board, or just the final pulse to the panel itself? And when you say SS fault, do you mean SS board? It's my understanding that the SS board is where the erase pulse is handled, along with the other "blunt" pulses (ones that go to the entire panel at once).
                The SS board is responsible for both the alternate sustain waveform (the waveform out of phase with the SC), and some energy recovery functions. It doesn't generate the erase waveforms itself, but it is necessary (in some way) for the panel to see all of them.

                Originally posted by blargg View Post
                I tried grounding each of the four UEH, USL, USH, and UML (after some resistor buffers, so I wasn't shorting the signals) to see what would happen, to get a better idea as to what the SS board was doing. After the UEH one, the whole screen got "erased" to black and only a few vertical lines at random points started "showing through". Even after a while, these only grew some and I was worried. After an hour or two, they grew vertically and expanded some horizontally, but still weren't filling in lots of areas. Waited overnight, same. I was worried that I had damaged it more. Today I messed with the various voltage trimmers, including some unlabeled on the SC board, and over time was able to fill in the whole screen to working how it was previously. I think I found that with the SC voltage higher (my guess as to what they adjusted), it got rid of almost all the dots for a while. I also found that lowering the erase voltage (VE) didn't seem to affect things.
                This is a bad idea -- you're lucky you didn't damage the SC.

                The SC is a "dumb" board. It's got lots of transistors and diodes and such on it. It deals with about 200V, at 1.8A for a 42". If it happens that two transistors get switched on simultaneously, nothing will stop it from destroying those, as well as other circuitry, and potentially the buffers.

                The D board contains the logic which sequences the control lines and it knows when to turn off transistors so they don't simultaneously conduct.

                The Ve is the electrode voltage, not the erase voltage. Don't ask me what is does. It is apparently -vital- that it is lower than Vs at all times.

                Originally posted by blargg View Post
                I've read a bit about how a plasma works. My basic model is that each cell is basically a capacitor that acts as a binary memory. To display something, it first charges the respective cells to around 67V, which won't cause any plasma. Then, it pulses a higher voltage across all cells at once, such that only the charged ones will reach plasma state. This also keeps the charged ones charged and the off ones not charged. It flashes them several times in a row for some duration (which varies for the cycle, depending on the binary weight it's currently at). Then it erases them by discharging them with a strong negative voltage ramp.
                That's mostly correct, except the cells aren't really capacitors (you can't measure their charge), but simple memories. The memory is created by the electrons in the nitrogen knocked free by the Va addressing pulse. The addressing pulse causes no light emission because the actual light emission comes from UV light hitting the phosphors -- and the electrons don't have enough energy yet to cause this. (Side note -- apparently this is a great efficiency loss in plasma displays, hence why Va draws around an amp. It's all dissipated as heat in the panel.)

                To make the subpixels which have been written light up, the alternating sustain voltage is applied to the panel. The SC applies +200V, the SS applies 0V, the plasma cells see 200V across the Scan and Sustain electrodes. Then the polarity flips; SC applies 0V, SS applies +200V. This happens at around 80-200kHz. This AC causes light emission from the pixel.

                Originally posted by blargg View Post
                Since many are sticking at black, this suggests that it's the charging that's not working, rather than discharging (erase). Or maybe if they're erased wrong, they can't be charged correctly either?
                I don't know enough about this part, but I have oddly seen poor erase voltages cause both non-firing and extra-firing of pixels. Maldischarge is the technical term. I believe (don't quote me on this) that the ramp waveform before every subfield is important in both polarities. It's supposed to liberate the electrons and make them more willing to emit light.

                Originally posted by blargg View Post
                I also don't have an explanation for the extreme memory effect, and how long it took to unstick the pixels from black so they all generally worked again. I noticed that this unsticking effect seemed to only run upwards along a column, virtually never sideways. If a column was working near the bottom, it would slowly grow upwards one pixel at a time. I'm wondering whether this could at all be due to panel fatigue. I found the service mode and this has 5982 hours on it and 2089 power cycles (which is quite a bit over its 7 years of life).
                No plasma panel is perfect and they all have leakage occurring between cells. You will often see banding from bright elements superimposed on a grey background. I suggest that maybe if the nearby cells are too sensitive, they are picking up scan data from adjacent rows. Perhaps it could be an erase timing issue.

                Originally posted by blargg View Post
                I found an SS board on eBay for $13 shipped, but I'd like to have some confidence that this is really the cause, and not say the SC board.
                It depends how much you're willing to invest. While I don't think it's the SS board, without a scope, I would perhaps succumb to that low price -- just to eliminate the board from the equation.

                Originally posted by blargg View Post
                I was thinking that I might be able to verify the waveforms into the SS board using an old monochrome CRT or even just an LED swung quickly back and forth as an o-scope, since the signals appear to be in the millisecond-timing range. It seems more likely that any failure would be in one of these boards, given that they do all the heavy lifting.
                The signals are in the microsecond range. You would need a proper oscilloscope to test it. I use a Rigol DS1102E, but it weighs in at $399. You can get a cheap analog scope for $10 or less, perhaps even free if you ask nicely -- check Craiglist, etc. Make sure it has a 10X probe and a 5V/div minimum range.
                Last edited by tom66; 11-05-2012, 08:24 PM.
                Please do not PM me with questions! Questions via PM will not be answered. Post on the forums instead!
                For service manual, schematic, boardview (board view), datasheet, cad - use our search.

                Comment

                Working...