Getting a LG 60PZ850 tomorrow - anything to watch for?

[lookimback]

Re: Getting a LG 60PZ850 tomorrow - anything to watch for?

Yup, that's the thread that told me why I kept blowing IGBTs. I lost count how many, but I was off and on with that TV for months.

[Th3_uN1Qu3]

Re: Getting a LG 60PZ850 tomorrow - anything to watch for?

I have built some new coils with 20 strands of 0.3mm wire each. They get about as warm as the originals, so there is a significant current going thru them. The issue is still there.

I have fitted a heatsink to the control board BGA. It's still doing it after a while.

However, i think I'm on to something here, finally. When the picture gets messed up, the Vzb voltage goes up to 120 volts from 110. Trimming that back down does not get the picture back in good shape though.

The Vzb circuit is regulated by a 817B type optocoupler and a surface mount TL431. The voltage reference of that 431 also seems to go to another circuit, so if it's the reference of the TL431 which goes bad, that would explain why just trimming Vzb doesn't get the picture back to normal. There is a jumper on the top of the board where I can monitor that reference voltage.

I've put the board back in the TV and i'm waiting for the picture to go bad again, so i can measure if there is a change there. On power-up it is 2.469 to 2.470V, which is slightly low but normal. It would have to go up to 2.695V for Vzb to jump to 120V, all other things staying equal, we'll see if that's what's going on.

[tom66]

Re: Getting a LG 60PZ850 tomorrow - anything to watch for?

Black level on these is awful. Almost LCD like!! Panasonic's 2009 sets are better than this 2011/2012 LG panel. Had an opportunity to keep a 60" LG but decided not to because of this. You can tweak the black level by adjusting the Vset_up and Vset_dn. However these pots can affect the discharge quality at lower subfields so you may see some crushed/maldischarge on shadows and the adjustment may not last as the set ages. Can't cause harm by trying, though.

The Vzb issue is likely caused by a bad MOSFET in the regulation circuit. This is the same fault that afflicts the PS3000/PS6000 LG plasma series using the PDP50H3 and PDP60H3 panels. Replace the MOSFET and Vzb will no longer creep up.

If the Vzb voltage gets too high the picture will overdischarge and the PSU will shut down or something will pop. It's quite an amusing fault on the PS3000 sets. The picture looks absolutely fine for 10 minutes but as the voltage creeps up the intensity and buzzing increases to the point at which the panel is pulling darn near 800 watts from the wall and sounds like it's about to explode. Plus it's probably dosing everyone with a fair share of UV light. After a while, the PSU will shut down. I've not seen anything pop as a result, which is credit to how well LG built the sustain boards on these TVs.

These panels (and all LG and some Samsung PDPs) are unfortunately very prone to vertical line defects. It may be worth adjusting Va voltage to as low as it will go without causing maldischarge, to reduce the stress on the address drivers.

The white limiter is also pretty aggressive. Is this just done to limit power consumption, or is it also a means of protection? Can it be adjusted in any way? If possible, i would like to make it less obvious.

Nope. It's hard coded ABL limit on the panel controller. Unless you want to reverse engineer that. I have a firmware dump from a Samsung plasma control board (PS43D450) which I've been studying, but it's pretty undocumented and very hard to understand what's going on. Most of the magic seems to occur in a custom ASIC built into the same chip; there's a dual core ~400MHz ARM + an ASIC on one die, as far as I can tell, and the ASIC has certain registers which can be written to, but there's no documentation whatsoever (well, duh). See attached for a screenshot; only 10% so far, hard to go much further based on context only.

Also, here's something i did not really expect: It flickers! Like CRT-type flicker. Not painful, but noticeable. Does this really run at just 60Hz, or could there be something wrong? Interesting to note is that it does not flicker at all in 3D mode, not even when I take a look at the screen without the glasses.

Try switching the picture mode, as this alters the number of subfields. It's likely the owner has set it to a Cinema-type mode that gives you an excellent subfield count (10-11 SFs per frame) so high bit depth, but much greater flicker. Depending on how much dither you are willing to trade off on, you might set it to the Dynamic/Standard mode that may flicker less but give you more dark-field dithering. Input frame rate will also affect flicker. Most LG sets will drop to a 550Hz SF rate with 50Hz input so 50Hz refresh but a few will do 100Hz refresh with alternating 4/5 subfield drive for less flicker, whereas they *all* do 600Hz from a 60Hz input (it's really odd). They will also usually do 48/72Hz refresh from 24Hz input to reduce flicker. It's worth playing with source refresh rates here and picture modes to see what you can get out of it. Unlike LCDs which tend to run at a fixed frame rate and fixed performance, plasma panels perform differently with different refresh rates.

If you're seeing "sparkles" in black, you're likely seeing dithering. This isn't a fault. If the sparkles appear after white or other bright content, that may be a defect, but it usually cannot be fixed by voltage adjustments alone (sadly only a firmware tweak will do it, and LG only did this for their chronically maldischarge-plagued 2005-2009 gen PDPs).

Just noticed i've been reading the spec sheet wrong - this panel is supposed to scan at 600 not 60 Hertz, so visible flickering, in any mode at all, isn't something that's supposed to happen with this set.

Common misunderstanding of how a PDP works.

Each 600Hz refresh is only one subfield of which there are anywhere from 8 to 12 in a given frame. In your case with a 60Hz input the TV will have 10 SFs.

Each subfield is not a full field, it is a discharge of some of the image data to build the intensity map of the frame. The darkest subfields are on for the shortest amount of time, and encode the lowest bit depth data. The brightest subfields are on for the longest amount of time and encode the most significant bit data.

A plasma panel has no fundamental intensity control(**) so the brightness is set only by the ON-TIME of a given subfield. And to produce intensity-graded image data you need many subfields per second.

See this video: https://www.youtube.com/watch?v=FMlpl7AmvNo and here https://www.youtube.com/watch?v=yWecqe04DTI

So the panel will flicker at 60Hz. Panasonic got around this by mapping the subfields onto a 120Hz refresh rate, and alternately skipping subfields between major fields. This significantly reduced flicker & buzzing. The tech that LG was shipping in their 2011 panels was roughly that of a 2008 Panasonic PDP but I think Panasonic's method is patented so LG could not use it. (Samsung had licence on some of Panny's PDP patents but I don't believe LG ever did.) So LG panels will flicker more.

Bad caps will not cause this effect. They can cause pinking-in-white if the high-voltage Vs caps go bad, though I've never seen that. It certainly wouldn't affect picture flicker under normal modes of operation.

(**) disclaimer - Panasonic do have a patent on using a partial-reset discharge to produce an effective 1/2 SF discharge, to add one more bit level to their panel, but I don't know if they ever used this. In any case, it's far too inefficient to use for any more than the lowest subfield.

Went to the hardware store, sadly they did not carry any freeze spray or duster...

Freezer spray/air duster cans are literally just butane, at least the cheap ones... Just buy a can of lighter fluid with an easy aerosol dispenser or swap it from another one.

Picked up a can of contact cleaner though. I also found all the screws needed to assemble the stand (the former owner had lost them) Put it onto the stand and placed a room fan behind it on lowest setting. Laid it down so it blows in the area where the control board is.

I guarantee your fault is not caused by the control board.

Failure of that will cause many bad things but not maldischarge in dark areas.

You may want to look at the panel waveforms with a scope but you will need a good digital scope to do this.

5 hours later, the sides where the sustain boards are, are still getting hot. The control board area is really cool since it has the fan blowing on it. Picture noise did not show up at all, it is still perfectly clear after running this long. So I guess i found my problem. The BGA on the control board does not have a heatsink on it, and I remember it almost burning my finger yesterday when i touched it after i had just unplugged the set. There's plenty of room for a nice heatsink there, guess i have to go thru my stack of dead boards.

FWIW, I've never seen this chip go bad on an LG, but heatsinking it probably isn't the worst idea.

[Th3_uN1Qu3]

Re: Getting a LG 60PZ850 tomorrow - anything to watch for?

Black level on these is awful. Almost LCD like!! Panasonic's 2009 sets are better than this 2011/2012 LG panel. Had an opportunity to keep a 60" LG but decided not to because of this. You can tweak the black level by adjusting the Vset_up and Vset_dn. However these pots can affect the discharge quality at lower subfields so you may see some crushed/maldischarge on shadows and the adjustment may not last as the set ages. Can't cause harm by trying, though.

Thanks for chiming in Tom, I did notice the panel performs differently at different refresh rates, but i'm sticking to 60Hz for now, because the firmware kicks it out of "PC Mode" at anything other than 60 or 59Hz, and the built-in sharpness controls cause issues on very fine details, regardless of their setting.

I have noticed that the black level isn't terribly good. I have tweaked the SET_UP and SET_DN some, it improved. I don't really mind it, it is good enough. Color reproduction and uniformity are great and that's all I care about right now. These large plasmas are really hard to come by here in Romania so I took what i could find.

The Vzb issue is likely caused by a bad MOSFET in the regulation circuit. This is the same fault that afflicts the PS3000/PS6000 LG plasma series using the PDP50H3 and PDP60H3 panels. Replace the MOSFET and Vzb will no longer creep up.

The voltage does not creep up. It jumps up suddenly from 110 to 120 volts and stays there.

These panels (and all LG and some Samsung PDPs) are unfortunately very prone to vertical line defects. It may be worth adjusting Va voltage to as low as it will go without causing maldischarge, to reduce the stress on the address drivers.

Thanks. Va voltage doesn't seem to do anything to the picture, i have set it to its sticker value of 55 volts. Will try lowering it.

Nope. It's hard coded ABL limit on the panel controller. Unless you want to reverse engineer that. I have a firmware dump from a Samsung plasma control board (PS43D450) which I've been studying, but it's pretty undocumented and very hard to understand what's going on. Most of the magic seems to occur in a custom ASIC built into the same chip; there's a dual core ~400MHz ARM + an ASIC on one die, as far as I can tell, and the ASIC has certain registers which can be written to, but there's no documentation whatsoever (well, duh). See attached for a screenshot; only 10% so far, hard to go much further based on context only.

Thanks. I wanted to dump the firmware while i had the control board out but was too lazy...

If the sparkles appear after white or other bright content, that may be a defect, but it usually cannot be fixed by voltage adjustments alone (sadly only a firmware tweak will do it, and LG only did this for their chronically maldischarge-plagued 2005-2009 gen PDPs).

It's in white and light colors. But only after it's been running for a while (20 minutes to 1 hour, maybe 1h30 after it is turned on) and room temperature is above 29C.

I guarantee your fault is not caused by the control board.

I guarantee that it IS. Failing to solve anything by messing with the VZB regulator on the Z-Sus, and since i discovered that there is a path to the control board from the VREF pin of the TL431, so the ctrl board could be doing these adjustments (when the picture noise appears, VZB jumps by exactly 10 volts and stays at that level, it does not drift up, or down further, so it's likely that it is TOLD to do that!) i turned my attention to the ctrl board again. I was looking at a replacement board on ebay and this struck me:



The rectangle at the top-right where the ground plane is missing, to the left of the screw hole, with the little chip in the middle. That had a corresponding thermal pad making contact with the backside of the panel. If that isn't a temperature sensor, i don't know what it is! I have removed the thermal pad and mounted the control board back in. I'll report back.

[Th3_uN1Qu3]

Re: Getting a LG 60PZ850 tomorrow - anything to watch for?

It still acted up without the thermal pad coupling it to the screen. So I took the CTRL board out again, put some flux on the sensor and removed it with my hot air station.

The sensor was certainly doing something useful until it would mess the picture up as i had to retweak the pots a bit. But 1h30, some youtube and a TV show later, picture still stable. Room temp crept up to 30.5C in the meantime. I'll keep you posted but i think we can call this solved now!

I'm thinking of putting the sensor back in (before I lose it!) but cutting the VCC line and wiring it thru a switch on the backside of the TV. Maybe it will be useful in the winter.

One more question for you, tom66.
I had to adjust -Vy down to achieve more range for SET_DN and as such get better blacks without maldischarge on black and dark colors. The panel rating is -178V, i lowered that to around -170. An additional benefit of lowering -Vy is that the horizontal trails that would show to the right of high contrast black on white or light colors have almost vanished. Upon doing so, the Vsc adjustment no longer reaches the 150V that's mentioned on the panel sticker even when it's maxed out, these two adjustments interact with each other and Vsc is 146.8V at the pot's maximum. Do you think I would have anything to gain by modifying the circuit to get Vsc back in spec, or does 5 volts less on Vsc make little difference? Thanks.

[tom66]

Re: Getting a LG 60PZ850 tomorrow - anything to watch for?

So the temperature sensor on these boards is important, because the panel's characteristics are adjusted by the control board according to temperature.

Reminds me of this old 42PQ3000 I had, another LG "beaut". This was released after their issues with the old 42V7/V8 panels and maldischarge so LG took a much more...shall we say, "cautious" approach. If the panel was below 15C the black level was a light grey. Contrast ratio was 100:1 at best. But above 15C, it just "snapped" and the picture got much better; black level dropped to Panasonic PZ80 levels. It was very odd. Unlike Pioneer and Panasonic that used gradual adjustment, LG decided to just wing it with a simple threshold point.

I suppose that could be what you are seeing, but as the panel ages, the transition point will probably change, so it's not going to be reliable. You might need to find out how to adjust that Vzb change, or see if you can set other voltages to make it behave better once it warms up.

I had to adjust -Vy down to achieve more range for SET_DN and as such get better blacks without maldischarge on black and dark colors. The panel rating is -178V, i lowered that to around -170. An additional benefit of lowering -Vy is that the horizontal trails that would show to the right of high contrast black on white or light colors have almost vanished. Upon doing so, the Vsc adjustment no longer reaches the 150V that's mentioned on the panel sticker even when it's maxed out, these two adjustments interact with each other and Vsc is 146.8V at the pot's maximum. Do you think I would have anything to gain by modifying the circuit to get Vsc back in spec, or does 5 volts less on Vsc make little difference? Thanks.

No, Vsc is just the floating scan voltage. It usually has an unregulated output around 130 - 160V. Anywhere in that range is likely OK. Technically, it's closely related to Va, so if you had a low Vsc, you could boost Va to compensate, but a higher Va affects address driver heating, so it's best avoided if possible.

While you're inside the set, double check that the heatsink on the address COFs is well secured!

[Th3_uN1Qu3]

Re: Getting a LG 60PZ850 tomorrow - anything to watch for?

To clarify, along with the 10V boost in Vzb, the control board also does other adjustments when it reaches that temperature threshold, as bringing Vzb back down does not remove the noise. I do recall someone on AVSForum worried that they damaged their set by pot tweaking and described the same kind of junk over the picture that I got, but then he came back and said it was just due to the set overheating for running for a long time in the summer, as it went back to normal after it's been turned off for a while. So this is likely a firmware bug on LG's part and not a malfunction of my particular set...

This set had lived in an air-conditioned room until I got it (which explains why it did not give any trouble to its former owner). So I am not too worried about the COFs right now. But if you could point me where I should look, maybe I will.

Anyway, more progress. I was able to put Va down to its minimum of 50V without any ill effect - in fact, the picture actually improved! With the sensor removed, I was getting some slight noise on light areas again, when I first turned on the TV. I am sure this would have gone away after it warmed up but I was eager to test your suggestions. Putting -Vy all the way down would get rid of them, but putting Va all the way down also did the same thing.

Another thing I did is boosted Vs to the maximum panel rating - it says 204 volts, I set it to 203.8 on a full white screen, it goes up to 203.9 with normal content. The picture is a bit brighter now. Factory calibration for this panel was supposed to be 201 volts. There is quite a bit more range left for Vs but i don't want to blow anything up! After boosting Vs, the -Vy adjustment range has changed slightly, now I need to keep it at -171 volts for an optimal picture, it starts having the misfires on light colors if I turn it all the way down.

[Th3_uN1Qu3]

Re: Getting a LG 60PZ850 tomorrow - anything to watch for?

Hmm. Low Va seems to cause other misfires so i turned it back to stock setting of 55V. -Vy almost all the way down and Vs at 203.6V. I would have liked to be able to keep Vs higher as it makes the screen brighter, but it introduces misfires on light gray that can't be tweaked out with anything else.

Vzb doesn't really seem to do anything on this set, i've tweaked it from as low as 94V to 120V and it did not change the picture one bit. I'll try again tonight to tune the blacks, but anyway, this shows that the increase in Vzb isn't what was causing the misfires with the sensor installed, the control board likely did other timing changes.

I think that i'll try to make a breakout board for the sensor and stick it to the back of the TV, because it does take its time to warm up even now in the summer. But the picture it shows now when cool is still better than when the sensor used to kick in when it got very hot.

[tom66]

Re: Getting a LG 60PZ850 tomorrow - anything to watch for?

Ok, yeah, so it looks like your set is switching into a different preset because LG made some miscalculation/guess about panel aging.

Not much you can do about this unless you can find an EEPROM on the board, dump that and see if you can find an obvious hour counter (or possibly observe data from several hours apart)

Pioneer made the same error on their KURO sets with a red mist appearing on black with age, but this was easily fixed by adjusting the various voltages from the service menu.

You can try SET_UP and SET_DN but unfortunately you may not be able to get it perfect.

The brightness difference with Vs increasing is pretty minor. Less than 2% by eye. On my Panasonic set it actually made no difference and a Vs as low as 180V produced similar brightness (there was a difference but it was really hard to see). So, I would be tempted to suggest a lower Vs instead, as possibly a way to fix the issue, instead of a higher Vs.

[Th3_uN1Qu3]

Re: Getting a LG 60PZ850 tomorrow - anything to watch for?

Lower Vs makes the trails on high contrast and image retention really bad. Also, it has a really noticeable impact on brightness. I have really sensitive eyes, maybe I notice it more than most people, but going a bit out of spec with Vs (209V) made the screen a decent amount brighter.

Actually what I ended up doing is setting Vs back to stock too (even if I kinda wished there could be a way to preserve the extra brightness), as there would be a bunch of red pixels running to the left just below the tab bar in certain colors with Mozilla browser tabs. Really minor issue, I know, but stock Vs gives more wiggle room for both -Vy and SET_DN between those red running pixels and dark green splotches on purple, so i'd rather keep it this way to account for temperature drift.

This set has no hour counter in the user menu. I'll dump the EEPROM of the control board before I button it back up.

As for how long it actually takes to warm up, after 2 hours of having the TV off and the room fan right next to it, it started with both minor maldischarge on white and on purple, then cleared up to perfection in 12 minutes. This is with the back off and not close to a wall, so it'll probably warm up sooner when it's in its intended spot. LG's service manual say to do a "heat run" for 10 minutes before doing any voltage adjustment - makes sense now. The issues were not noticeable with normal video content, I think I can live with that. The voltage settings that I had left it at were done about 2 hours from it being turned on and with the panel quite hot, so it's obvious to me that it reaches a stable operating point after it warms up a little and then does not deviate from that with further temperature rise.

I have not felt the desire to tweak it any in the dark, black level is good enough, and i'd tweaked it close to the point of the green spots or purple during the day. If i turned Vs up to get it brighter, I got to pick between the red running pixels and the green snow on purple, i couldn't have both go away, too fiddly.

I think I could live with that, but i'll probably go thru with my plan of either relocating the temperature sensor, putting it on a switch, or both. Just to know what I was coming from with this set, attached is a picture of what it looked like when the temperature sensor messed up the settings. I only have that closeup to show, but on video, peoples' faces were getting horribly solarized. I'll gladly take 10-12 minutes of warm-up vs becoming unwatchable after 30min in the summer.

[Th3_uN1Qu3]

Re: Getting a LG 60PZ850 tomorrow - anything to watch for?

I'll take some pics of the minor maldischarge that occurs now until warm-up, after removing the temp sensor. It's a whole lot better even at start-up.

I did tweak it a little more before bed last night. -Vy was left at minimum this time. This morning, after being off for 7 hours and in lower room temp, it took only around 5 minutes to clear to a perfect picture.

I think i'll take off the Y-Sus again and mess with the divider on the -Vy pot, as it seems to me that it would like that even lower. Odd why they would set it so high from the factory, as high -Vy requires set_dn to be set for really poor blacks to not cause maldischarge on dark colors. It could just be age, but the optimal settings for Vs and Va on this panel are still the ones printed on the factory sticker, so it's odd to me why -Vy (and since it interacts with it), Vsc, would have to be so far off.

[ReeceyBurger123]

Re: Getting a LG 60PZ850 tomorrow - anything to watch for?

Since this is not a Smart TV, do I have to worry about the main processor overheating and change the heatsink or add a small fan? I've had a couple LG mainboards that i needed to reball or replace the CPU.

Best comment I've seen, completely bang on !!