Hisense 65H6570G - Color inverted after replacing LED backlight

[ojmc]

Background:

• Hisense 65H6570G stopped working and I root caused to a burnt out LED in the backlight. (For anyone interested the power light pattern when the backlight has gone is 1) two long flashes, 2) a very short third flash, then 3) back to solid red https://photos.app.goo.gl/VWve47GqPxLdApJ48 )
• I replaced all 6 of the backlight LED strips with these replacements: amazon
• The TV worked again perfectly for ~2 weeks.
• After that 2 weeks the colors became "inverted". The TV powers on and shows the boot screen in correct colors initially (aqua hisense logo on black background, pic on the left) but a few seconds after boot the colors invert where black becomes white, red becomes green etc.

I have tried re-seating various connectors and found that if the left quarter of the panel is disconnected then the remaining 3/4 of the screen shows correct colors. I tried applying pressure and re-seating all the ribbon cables for this left quarter but that did not improve anything.

Is there any chance of saving this or does this color inversion suggest that I damaged the panel during the backlight replacement?

[Diah]

try log in service menu and do " clear chip"
also check the LVDS cable between the MB and panel from both side... use magnifier to inspect all pins ... magnifier to use don't trust your glasses or eyes. lol

[Davi.p]

Hi, my suspect the panel is ok, but you have processor to be reballed, i wanna you make this experiment (never tried) , remove heatsink, check in volts every choke inductor, when you find the one for the vcore around 1v, connect in parallel an electrolytic cap low esr of around 470uF 6,3v or more to the caps connected to that inductor, then try the tv..

[Sam Scarbo]

If left quarter ribbon cable is causing the color issue, i would use a small piece of cellulose tape or kapton tape to insulate certain pin(s) on the ribbon cable to see which one is causing the issue.
Good luck.

[EazyBone]

If you have to resuse the grounding tape, make sure you stick a piece of clear tape over it so it keeps contact. Did you reapply the tape on that side good enough?

As you say 3/4 works fine so it's likely a panel problem created by the replacement. Don't reball your processor and be his guinea pig lol, the main board is obviously working. Sam gave a good suggestion and diah.

[ojmc]

Thanks for all the suggestions. The results I have are:

• The TV was off for a few weeks and upon powering on the colors remained correct for ~2 minutes before switching to inverted. Other than that initial power on, the color now inverts almost immediately upon booting up from powered-off (and from standby).
• Checking grounding tape - the tape on the far left (on the quarter that causes the color inversion) was semi-peeled off although I think still in contact. I applied clear tape to firmly affix but that didn't change anything (I am reusing the grounding tape). After seeing this I am slightly worried that lack of grounding may have cooked some component?
• Covering ribbon pins/checking ribbon cables -
  • The only way to remove the color inversion is to completely disconnect the left quarter of the panel by either a) disconnecting the white ribbon cable (E248682 AWM 20941) for the left half of the panel from the main board, or b) disconnecting the brown ribbon cable (ST6451D02-B-FPC-6) that connects the left most quarter with the inner left quarter.
  • Isolating either half of the pins on either the white or brown ribbon cable results in some part of the image not working but the colors stay inverted in the rest of the panel.
  • The same brown ribbon cable is used on both sides to connect quarters together -- I tried swapping the left and right cables and nothing changed (disconnecting the left quarter still fixes colors even with newly switched brown ribbon cables)
  • I can't see anything suspicious with a magnifying glass. Can't see any breaks in the traces and the pins on the cable and socket look normal.
  • 
• "clear chip" in service menu - I couldn't find anything resembling that in service menu, am I looking in the right place?
  
• Reballing the processor - I don't have equipment to reball the processor but I popped off the heatsink and took a look. The only inductor I could find that was ~1V from ground was L3. (The 8 pin IC next to it is labelled MSH6103A which I think is the 1V regulator?). I tried holding a 470uF 100v electrolytic cap (labeled low esr but it is from aliexpress) parallel to C42/C43/C44 but didn't seem to do anything.
  
• Here are some test images -- there aren't any areas of dead or degraded pixels. Red is mapped to cyan, green is mapped to magenta, blue is mapped to yellow and vice versa (ie opposite sides on RGB/CMYK color wheel)
  
• I checked board temps with a heat laser gun - there is a hot spot on the left quarter board attached to the panel next to where the ribbon cable connects (the same quarter where disconnecting fixes the color inversion). It is about 100F at that point while other boards are all 75-80F. Just going by touch it seems like it is around C17 that is the hottest part
  

Any further pointers/ideas are much appreciated. Should I spend more time trying to cover pins on ribbon cables? I tried covering each half at a time but I could try covering 90%+ at a time if that might have a different outcome

[Diah]

what do you have under factory mode ---- option ?

[Davi.p]

I downloaded for the first time the main board pic , and noticed only now there is an integrated tcon to the main board, but the chip that controls it, it is outside the main processor, i noticed the Chot chip that is a panels brand, so the capacitor where you put the cap it is not the right one, there have to be another around the Chot chip, maybe 1,2v, maybe 2-3v, that powers the chip itself, that is the correct one..

[Francesc V.]

As an idea...............could you measure generic voltages at the T-CON in both situations (screen showing right colors and screen showing bad colors)?

Maybe there is a naughty voltage there......

[ojmc]

The problem seemingly fixed itself for a few weeks (~10 hours of watch time) and has now appeared again. Disconnecting the left most quarter of the screen DOES NOT fix the issue anymore. I can't find any permutation of disconnecting ribbon cables that fixes the problem and even when a section of the panel is disconnect that portion is still washed out white.

what do you have under factory mode ---- option ?

• I tried "Reset Gamma" and "Clear Unprotectly" in the service menu. There wasn't any change in the problem after doing either of these.

I downloaded for the first time the main board pic , and noticed only now there is an integrated tcon to the main board, but the chip that controls it, it is outside the main processor, i noticed the Chot chip that is a panels brand, so the capacitor where you put the cap it is not the right one, there have to be another around the Chot chip, maybe 1,2v, maybe 2-3v, that powers the chip itself, that is the correct one..

• I measured all the inductors around the csot cs602 chip

• 	color inverted	color correct
  L6	1.53	1.54
  L18	12.03	12.03
  L21	12.03	12.03
  L28	7.56	7.59
  L27	~0.27mV	~0.27mV
  L20	3.3	3.33
  L31	1.87	1.87

• The first inductor I tried was L28 7.5V with a 470uF capacitor across C315 and this fixed the problem for next boot (may just be coincidence?). The colors stayed correct across power cycling at the wall socket even though the capacitor is removed (ie capacitor only applied for the first boot)
• I can't trigger the problem reliably so I will have to wait until it appears again to do more experiments.
• I'm not sure if the fix with the cap across C315 is just a coincidence. I will keep using the tv until the problem appears again and see if I can reproduce this fix.

As an idea...............could you measure generic voltages at the T-CON in both situations (screen showing right colors and screen showing bad colors)?

Maybe there is a naughty voltage there......

• I've measured all the inductor voltages and all voltages seems to be the same regardless of whether picture issue is present or not. (table above)

[Davi.p]

Ok, the capacitor was just my curiosity idea, but i see often polemics here, the most important here is to fight between techs.. polemics aside.. the correct voltage was one between the 3,3v (most probably) and 1,5v, it needs to be the csot chip vcc, 1,5v usually is for ddr ram that are near there, but the rapid change between normal Logo and reverse logo lets me think of some broken joints around the CSOT qfn soldering, so the capacitor do not help in that case, my guess it needs to be dry out all old tin and put new tin, without remove the chip, if you have compressed air can you can spray it upside down and freeze the chip when defect appears..

[ojmc]

The problem surfaced again and capacitor trick did not work again on any of the voltages.

I don't have any freeze spray or compressed air on hand but I will obtain some. What is the purpose of freezing the chip? Should I expect to see a hotspot where there is a bad solder connection?

Another observation is that various settings have changed between power cycles (audio settings, wifi enabled and home/store mode) so maybe there is some failing eeprom?

[Davi.p]

Another observation is that various settings have changed between power cycles (audio settings, wifi enabled and home/store mode) so maybe there is some failing eeprom?

stop, here the things changes.. a lot.. so now i'm suspecting other components, but which one??
You can blow hot air with air dryer first on the two ddr ram, coverinng the cpu and.the emmc before with aluminum foil, then turn on and see what's going on, if nothing then use same method heating cpu or.emmc...