TH-P50G15A - 6 Blinks- Bad TNPA4782

[tw2005]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

Mine is the AB model, I got from a TCP50S1. I did the initial replacing parts and heard a pop and burnt smell which turned out to be D618, and also leading to D608 blowing ,Realizing I missed IC521 and Q551. after replacing everything again, 7 blinks turned to 6 blinks.

If the SC is repaired and with SC50 jumpered and SC plugged in, the set should stay on solid and no blinks correct?

yep and i strongly recommend you test it that way until it does becasue I'm convinced I killed my buffer boards.

[tw2005]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

Everything I found shorted so far:

Q441 - 45G127 replaced with a 45G128
IC521 - replaced with the 1737
Q551
Q661 - RJP6065 replaced with RJP63F3
Q601 - RJP6026
Q701
Q702
IC773

D608
D618
R521
R522

I also checked all the parts that you had shorted or suspect and none seemed shorted, but I replaced some of them just in case.

Does not seem quite fair does it.

I don't see and 2sk3018 or 78L05 regs replaced

have you tripled checked all smd diodes(large ones) and if 45G127 failed is the didoe in front of that ok and the 2 TO220 diode packs next to it?

[Glitcher]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

Mine is the AB model, I got from a TCP50S1. I did the initial replacing parts and heard a pop and burnt smell which turned out to be D618, and also leading to D608 blowing ,Realizing I missed IC521 and Q551. after replacing everything again, 7 blinks turned to 6 blinks.

If the SC is repaired and with SC50 jumpered and SC plugged in, the set should stay on solid and no blinks correct?

[tw2005]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

Just having a look at this spare dead 4782AF, straight away I see D618 is burnt. I'm starting to think i missed this as it's in the shadows of the heatsink .

What's unusual is this too being revision AF , however it has a Vset cct.n So i have one which has and another not both AF. This board does look a little older so maybe mine was the last of a production change. Both TVs are P50G15A

[Glitcher]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

Everything I found shorted so far:

Q441 - 45G127 replaced with a 45G128
IC521 - replaced with the 1737
Q551
Q661 - RJP6065 replaced with RJP63F3
Q601 - RJP6026
Q701
Q702
IC773

D608
D618
R521
R522

I also checked all the parts that you had shorted or suspect and none seemed shorted, but I replaced some of them just in case.

[tw2005]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

Buffers are in. (thanks for Overnight express) Perfect,Perfect Perfect. Fantastic image and SC is now silent with a load .

10500 Hrs

Now have another TNPA4782 to repair. A lot learnt here as far as what parts you can get away with , with some lateral thinking and cross relating knowledge of the 5081 boards.

[tw2005]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

I thought I also had this but nope. Found two resistors R521 and R522 open circuit near IC521. I initially had IC521 shorted which was replaced. replaced the resistors but still getting a short flash and then 6 blinks.

Can you put a summary list of what you've replaced?

Did you check those 2 diodes (marked 88) that I just had shorted ?

[Glitcher]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

I thought I also had this but nope. Found two resistors R521 and R522 open circuit near IC521. I initially had IC521 shorted which was replaced. replaced the resistors but still getting a short flash and then 6 blinks.

[tw2005]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

Which buffer ICs failed? tried removing a few failed ones and see if it will fire up with a pic?

Don't have the gear. brought everything except the kitchen sink and the acetone

Not worth the

All connectors on the SD. 8 ICs on these too.

[tom66]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

Which buffer ICs failed? tried removing a few failed ones and see if it will fire up with a pic?

[tw2005]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

I'm thinking it would be a useful experiment if you ever get a set you don't care too much about breaking (older set maybe) try a largeish diode reverse biased across the FGND-VSCAN portion. If all goes tits up this diode SHOULD conduct before the buffers get a chance to blow. Obviously SC will be damaged but failure was going to occur anyway.

I'm convinced that we only see modern sets with failed buffers because they're less tolerant of such a reverse bias event. Older buffer ICs with 48~68ch per IC and larger silicon were very tolerant, apparently. KUROs use 68ch per IC (16 for FHD panel). 2007/2008 FHD Panny used the same. 71" FHD LG used 36 buffer ICs (2006 model, 30ch/IC) !



Good, looks like you're good to go there

Actually check the FG5V too (should be 5V above FG, for buffers. can't remember what Panny call it, LG use FG5V) if 5V goes missing, maybe buffers will fry as logic signals invalid? could be a case of turning on both switches inside an IC. Buh-bye when that happens...

On this model F_5v_A, & _B

pins 3 & 5 on SD42. Both produce 5V with respect to Vfg

[tom66]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

I'm thinking it would be a useful experiment if you ever get a set you don't care too much about breaking (older set maybe) try a largeish diode reverse biased across the FGND-VSCAN portion. If all goes tits up this diode SHOULD conduct before the buffers get a chance to blow. Obviously SC will be damaged but failure was going to occur anyway.

I'm convinced that we only see modern sets with failed buffers because they're less tolerant of such a reverse bias event. Older buffer ICs with 48~68ch per IC and larger silicon were very tolerant, apparently. KUROs use 68ch per IC (16 for FHD panel). 2007/2008 FHD Panny used the same. 71" FHD LG used 36 buffer ICs (2006 model, 30ch/IC) !

Vfg - TPSC1 = 143.1V

Good, looks like you're good to go there

Actually check the FG5V too (should be 5V above FG, for buffers. can't remember what Panny call it, LG use FG5V) if 5V goes missing, maybe buffers will fry as logic signals invalid? could be a case of turning on both switches inside an IC. Buh-bye when that happens...

[tw2005]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

I'd say since MID is good the sustain portion is ok. Check voltage between FGND and Vscn, if it's ~140V (positive, FGND < Vscan or buffers fry!) you should be good to put in new buffers.

Vfg - TPSC1 = 143.1V

[tw2005]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

So you just replaced the shorted diodes and that seem to have fix the problem?

Hard to say if I missed a diode or what happened. definitely missed IC521. I even tried without buffers and still had 7 blink. I did try bypassing the SOS7, maybe that's when I killed the buffers( I think I did)

Basically doing a whole lot of stuff wrong, rushing as I have to leave this TV on Friday to head home.

I don't normally test SCs with live buffers. I only just clicked that maybe the state of the buffers had changed and that's when I did the resistance at Vfo and both dead short.

SC50 bypass and pleasantly surprised.

No I had to replace the newly shorted stuff Q660,601 , 661 as well. Nothing had blown so maybe I damaged the buffers and the next cycle that's when those IGBTs popped and D618 went bang and the track also vaporised.

[tom66]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

I'd say since MID is good the sustain portion is ok. Check voltage between FGND and Vscn, if it's ~140V (positive, FGND < Vscan or buffers fry!) you should be good to put in new buffers.

[Glitcher]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

So you just replaced the shorted diodes and that seem to have fix the problem?

[tw2005]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

interesting update. I've made what appears a real cluster of this but I now have a running SC, no blinks, no picture.. SD/SU are shot.. both shorted and I'd like to think I checked them properly as not but only gave them a quick continuity check between vfg and the outputs.

anyway, I'm not going to be exact on what exactly was faulty but have a fair idea.

Originally

Q661 RJP6065, shorted_..I've used a RJP63F3A which is the newer version .
replaced Q818 2SK3018, used 2N7002
Q702 2PD601 was short, IC773 TC7SH14F was shorted Vcc-gnd(41ohms),
Replaced Q701 MMBTA92

Replaced all 78L05 regs( they appeared ok) IC724,725,771
Replaced all TC7SH14F _ IC773 bad , others looked ok, IC464,465,564
Replaced TC7SH00FU - IC691 NAND
Replaced IC521 M81707FP, used M81737FP which is same pinout higher current output
Replaced Q551 ( looked ok)

Replaced Q660,601, RJK6026 ... I had none and used STP10NK60ZFP 600V 10A N FET.

Replaced D618,652 .. RF101L4S 400V 1A super fast rectifer.

SC50 buffer bypass(buffers now shorted), Powers on, Sc is a bit noisy and yet to see how the picture is.

voltages are now:

TPVsus 198.7V
TPVscn -41.1 (- 40V ± 6V)
TPVAD -186.1 (- 185V ± 2V)
TPVe 135.2

MID is now 99.7V as predicted by Tom about 50% Vsus.

[tw2005]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

Did you check if D806 is okay?

Diodes range from 811-875, did you mean 860?

Just going through the bigger stuff again

but currently shorted

Q661,660,601, D618,652

[tw2005]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

basic theory is, MID is charged by panel capacitance on HIGH-LOW sus, and discharged by panel on LOW-HIGH sus.

If you just charged the panel using the SUS IGBTs (high and low) you'd waste tons of energy because you'd draw a helluva lot of current to charge up the capacitor which is basically a short circuit. Then you'd go and dump that all in the SUS DN transistor. Not good. Lots of heat, lots of power...

So, high ER IGBTs turn on before the controller wants to swap sus from high to low. This discharges panel (~100nF capacitance) thru inductors into the blue poly caps. Charges would equalise if there was no inductance so youd have 100V both sides. Problem is you'd still have to dump that 100V somewhere. 1/4 the energy (1/2 volt 1/2 time) but still a lot. Why not make use of it? Inductor creates resonant tank ckt, which is critically damped by the sustain timing, so 90%+ of energy is extracted. On the other cycle, lower ER IGBTs turn on when the controller wants to take sus from low to high. Same deal with resonance. Even though mid is only ~100V, resonance boosts it to 200V.

Hence why the capacitance in ERC (3~4uF) is much larger than panel capaciance (50~200nF). Haven't done the math but the inductor values are cca 1-2uH air wound coil, which sets resonant frequency 500kHz, half the sustain frequency (150~250kHz). Its a little more complicated because the panel capacitance goes up with screen size and changes depending on the image displayed and panel temperature also why the SCs are different between models.

modern designs have got the sustain frequency, energy recovery and panel drive so nailed down they can work without heatsinks on the SCs. Of course some don't quite manage it (2011 panny's being a famous example.)

I had a G10 in before I cracked screen, SOS6 too, MID voltage would hit 55V but no ER activity. Mid voltage is biased via a network of resistors to ~1/2 Vsus. This appears to be a strategy by Panny to reduce the turn on stress. Possibly between "black screen" events. The problem is when the ER caps are empty and the panel wants to start sustaining. It has to go through that energy intensive stressful sustain without ER until the voltage has built up. Once it has built up, the panel can be sustained efficiently. AFAIK only Panasonic do this. Not sure why.

Funny thing, look up Larry F. Weber. Single handidly invented half the stuf modern plasmas used to make things work including the energy recovery circuit. Epic beard.

Well, I'll see if I can use all that theory. i wonder if Larry knows how to fix this

As always, very informative Tom.

[tom66]

Re: TH-P50G15A - 6 Blinks- Bad TNPA4782

basic theory is, MID is charged by panel capacitance on HIGH-LOW sus, and discharged by panel on LOW-HIGH sus.

If you just charged the panel using the SUS IGBTs (high and low) you'd waste tons of energy because you'd draw a helluva lot of current to charge up the capacitor which is basically a short circuit. Then you'd go and dump that all in the SUS DN transistor. Not good. Lots of heat, lots of power...

So, high ER IGBTs turn on before the controller wants to swap sus from high to low. This discharges panel (~100nF capacitance) thru inductors into the blue poly caps. Charges would equalise if there was no inductance so youd have 100V both sides. Problem is you'd still have to dump that 100V somewhere. 1/4 the energy (1/2 volt 1/2 time) but still a lot. Why not make use of it? Inductor creates resonant tank ckt, which is critically damped by the sustain timing, so 90%+ of energy is extracted. On the other cycle, lower ER IGBTs turn on when the controller wants to take sus from low to high. Same deal with resonance. Even though mid is only ~100V, resonance boosts it to 200V.

Hence why the capacitance in ERC (3~4uF) is much larger than panel capaciance (50~200nF). Haven't done the math but the inductor values are cca 1-2uH air wound coil, which sets resonant frequency 500kHz, half the sustain frequency (150~250kHz). Its a little more complicated because the panel capacitance goes up with screen size and changes depending on the image displayed and panel temperature also why the SCs are different between models.

modern designs have got the sustain frequency, energy recovery and panel drive so nailed down they can work without heatsinks on the SCs. Of course some don't quite manage it (2011 panny's being a famous example.)

I had a G10 in before I cracked screen, SOS6 too, MID voltage would hit 55V but no ER activity. Mid voltage is biased via a network of resistors to ~1/2 Vsus. This appears to be a strategy by Panny to reduce the turn on stress. Possibly between "black screen" events. The problem is when the ER caps are empty and the panel wants to start sustaining. It has to go through that energy intensive stressful sustain without ER until the voltage has built up. Once it has built up, the panel can be sustained efficiently. AFAIK only Panasonic do this. Not sure why.

Funny thing, look up Larry F. Weber. Single handidly invented half the stuf modern plasmas used to make things work including the energy recovery circuit. Epic beard.