"Dick Smith" 55" LED backlight issue - model GE6880

[budm]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

The FB pin is for feedback regulation, just about any SMPS IC has the feedback pin function. So if the LED+ drops to 0 then that means U103 (OB2263) went into shutdown so we need to find out if the PFC Voltage feeding the LED power supply is still present or not.
I slo need to find out what the FB1 (pin 8 VFB) is for.

[rockit711]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

OK, got a little more info:

1) the PFC_OUT stays at ~385V (to ground) after the backlights shutdown. Which in hindsight was likely seeing as the 12V and 5V rails are derived from this also and they continue.

2) It appears I may have been mistaken yesterday about the FB pin of chip U103 being ~0V. Today after startup it seems to go to, and stay at ~4.8V. not sure If I had a bad contact with the multimeter probe or had perhaps done something silly there (like measured the ground pin not the FB pin) but it does seem to be at 4.8V now. Again still no idea if this is good or bad!

3) VDD pin of U103 is ~16 V so it appears to be getting a good power supply.

4) It appears that Q111 (the other transistor on the FB pin of U103) isn't actually implemented on this board. So at least that's one path that cant have failed!


I also probed the GATE pin of U103, and it seemed to stay at 0V, of course thats a PWM output, but I'd be expecting a some sort of reading if it were actually attempting to switch the power MOSFET.


I'm a little confused about this now. It seems that there is *some* switching going on, as there's a brief voltage spike at startup (hence the flash of the backlights) but there doesn't seem to be any sign of life from the U103 Gate pin. Or is it just that even if this were switching I wouldn't expect to see it on a multimeter voltage setting?

[rockit711]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

Apparently I'm a n00b at using a multi meter and stuffed up the GC1 measurement that I had in this post earlier. looks like its a much more sensible voltage, stays below 50V. And across the capacitor EC108 its only seeing ~6V but it seems to vary a lot between power cycles.

I also re-read the OB2263 datasheet and discovered I'd missed a paragraph
before:

"At overload condition when FB input voltage
exceeds power limit threshold value for more than
TD_PL, control circuit reacts to shut down the
output power MOSFET."


Am I right in interpreting this as: If FB pin is above power limit threshold value (VTH_PL Power Limiting FB Threshold Voltage =3.7V from page 5) for (TD_PL) period of time the chip shuts down?

if so, seeing as FB is 4.8V that would explain why the chip has shutdown!

which basically puts the direct source of the "fault" at anywhere in this area (and its attached voltage feeds):


note: Ive edidted the schematic to remove components that aren't implemented here to make it a little less cluttered

so BUDM may have been right about 10 posts ago with stating that we need to know what the OB3354 VFB pin is!

[dick_barton]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

I believe that the chip is in shutdown and therefore no gate drive to the mosfet.

[rockit711]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

I believe that the chip is in shutdown and therefore no gate drive to the mosfet.

100% agreed. Definitely the U103 (ob2263) chip is in shutdown mode.

I just cant figure out *WHY* it is - Obviously the optocoupler isn't "turning on" and pulling the FB pin down below 3.7V as needed. But again *WHY* I cant see.

I do have some more voltage readings if it'll help anyone assist with figuring this one out.

all are pin references/ test points are as marked on the schematic

1) GC ~5-6v (varies over cycles), very briefly.
2)Pin 2 of U203 (optocoupler in question) ~0V
3)"reference" pin of U204 (tl432 shunt regulator) spikes breifly to ~.6v then drops to .4V
4) the "BASE" pins(BASE1, 2 etc) of U5 - OB3354 all go to ~3V (coinciding with the backlight flash) and seem to stay there. I interpret this as meaning that the OB3354 is *trying* to drive the LEDS still (still switching the BJTs to drive the LEDs) even after the LED+ rail has lost power. - could someone please correct me if I'm wrong here?

5) and very confusingly (for me anyway) if I measure the voltage across D123 I get a positive result BOTH ways (ie. if I swap the multimeter leads around I STILL get a positive reading on the DMM)



Im utterly confused about this area of the circuit (the optocoupler drive area).. I have no idea whats going on here. From the similar area in the 12/5V area I assume the "reference" voltage of the shunt regulator is expected to be around 2.5V, but again *why* it isn't 2.5V is confusing me.

I swapped the shunt regulator over with a good one and it's still all the same as before, so that isn't the issue.

[rockit711]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

So, finally got *somewhere*

I desoldered the VDD pin of the OB3354 chip and raised it slightly so it wasn't contacting anymore.

I now get a stable output on the LED+ rail at 165 V

the "REF" pin of the U204 (tl432) has gone to the expected ~2.5V
the optocoupler is now working and pulling the feedback pin of the U103 down to 1.2V.

Of course the LEDs don't even try to turn on anymore.


So I guess this leaves me with a couple of possibilities.

(1) the OB3354 chip is bad.
(2) the LEDs are bad.
(3) the Diodes/transistors in the cathode return are bad.


Is there anything I can do to test any of these and narrow it down? Really would like to rule out (2) before investing in replacing the chip (well its only $2 but itll take a month or so to ship in from China) so its more of a nuisance if i wait all that time and it doesn't work anyway!
I'm about 99% sure that (3) is not the case as I've checked them all with the DMM and they appear to work, and would also be very odd for all 4 transistors and diodes to die at the same time! I suspect it is (1) the LED driver chip has failed, simply because if the LEDs had failed surely it would trigger one of the OVP/SCP etc and stop driving the transistor gates not just pull VFB to ground and shutdown a *different* chip!

[martyparish]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

Perhaps test every LED like this: https://www.youtube.com/watch?v=eEJoPCa7Kt0

I'd even take it a step further and add the meter in the loop to measure current on each one. If they are all close to same spec then I think you could rule out LED strips being bad.

[dick_barton]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

Reconnect everything as it should be and if you breifly apply a short across C219 which I think will overide the fault detection (but at your own risk) to see if it will bring up the backlights even breifly. If it does then you must have a led(s) fault.

[rockit711]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

I did a few more tests and am now 99% sure the OB3354 chip has failed. It looks like the VFB pin is being pulled to ground regardless of the operating conditions. Neither OVP nor SCP have actually been triggered and the chip even appears to continue operating (probably not as intended) even after killing the LED+ power.

I've ordered a replacement chip, and hopefully in a month or two I will be able to confirm that the chip was indeed responsible for this fault.

[rockit711]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

so got a few replacement OB3354 chips finally. Soldered one in and its all exactly the same as with the other one

So now am at a bit of a loss as to what to try next.


In desperation I also tried shorting across C119 as suggested above, and again absolutely nothing. I am a little confused by this, as as far as I can tell this should bring up the LED+ rail, but it appears to do nothing at all...

[dick_barton]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

If you disconnect the led connector does the 165V come up?

[rockit711]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

With the LED connectors unplugged it responds exactly the same as if they were plugged in. LED+ briefly goes ~100V+ then the LED+ rail drops down to 0. As far as I can tell all other pins respond in the same way when LEDs are unplugged or plugged in.

This would normally get me thinking its an open LED somewhere, however as far as I can tell the "open string" fault isn't being triggered (FAULT pin stays high).

pin 9 of the OB3354 chip is "FAULT" - and the best data I can find on the OB3354 chip indicates that this pin is the "Fault output (LED STRING OPEN OR SHORT)" seeing as this pin stays high it would tend to suggest to me that the LED open fault isn't EVER being triggered, even with the LEDs entirely disconnected!

I've included a copy of the pin descriptions of the OB3354 chip that I found. Unfortunately I can't find a datasheet for this chip anywhere so I don't have any idea whats actually supposed to be happening with this chip (for all I know this is entirely as expected!).



On a side note, when I swapped the OB3354 chips over I also went back to all the "original" parts (I had originally swapped over the opto and shunt-regulator, thinking they were faulty).

I did notice that with these original parts in and with the OB3354 chip power pin pulled off its pad the LED+ string was only going to ~80V, not the expected ~160. when I swapped in a new opto and shunt regulator it again went up to the anticipated 160V.

Does this assist in any way?

[SLK001]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

I believed that you had failed LEDs all the time while reading this. You should test every LED in your strips with a power supply. For each strip, multiply the number of LEDs by 3 and apply that much voltage - ie, if there are 10 LEDs then apply 30V. There are two test points at the connector end for factory testing, one (+) and one (-). If no power supply, take two 1.5V batteries and tape them together so that you have a "3V supply". Apply power to each LED by attaching the leads to straight pins and touching the leads of the diodes. The good ones with give off quite a bit of illumination. The bad ones may just slightly glow. Just be sure to get the polarity correct.

Another way is to use the pins to dig into the white resist layer of the strips near each diode. I don't like doing it that way because it is unprofessional, but it works.

[budm]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

Use the power supply with current limiter set to around 100mA.

[rockit711]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

I believed that you had failed LEDs all the time while reading this. You should test every LED in your strips with a power supply. For each strip, multiply the number of LEDs by 3 and apply that much voltage - ie, if there are 10 LEDs then apply 30V. There are two test points at the connector end for factory testing, one (+) and one (-). If no power supply, take two 1.5V batteries and tape them together so that you have a "3V supply". Apply power to each LED by attaching the leads to straight pins and touching the leads of the diodes. The good ones with give off quite a bit of illumination. The bad ones may just slightly glow. Just be sure to get the polarity correct.

Another way is to use the pins to dig into the white resist layer of the strips near each diode. I don't like doing it that way because it is unprofessional, but it works.

Sorry I should have mentioned earlier, I had already done the 2xAA batteries test over each LED (The LED strips have test points for each LED in the chain which made it much easier than it would have been otherwise!)

Each one illuminated approximately the same (there were no distinct dim or bright ones anyway) and drew about ~80mA

I actually did that test early may, it was one of the "few more tests" I had run before my last post at that time.


So that would seem to indicate that the are all at least somewhat functioning, unless they are failing at a higher current. but as far as I can tell the LED+ rail never even hits a voltage that would get even 80mA through them! - 48 LEDs per chain, means 144V to hit 3V per LED, the LED+ rail doesn't seem to get that high!


also the big big big indicator (in my mind) is that the LED open or short fault isn't being triggered. even when the LEDs aren't physically connected, which surely MUST trigger the open fault if its working "correctly"!

[rockit711]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

Use the power supply with current limiter set to around 100mA.

what do you mean? do you mean an external power supply or to somehow get the LED+ rail going and use that to drive it?

[SLK001]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

what do you mean? do you mean an external power supply or to somehow get the LED+ rail going and use that to drive it?

Set the power supply for 100mA MAX, then set the voltage to 5V. Use this setting to test the individual LEDs (or individual strips, if the voltage can go high enough). With the current limited, the voltage will drop to the voltage needed to force 100mA through the LEDs.

[rockit711]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

I took a movie of the LEDs flashing to see if any didn't turn on or were noticeably dimmer etc.


the only irregularity that I caught was one of the LEDs seems to "turn off" faster than the others. (picture below)

Is it likely/possible that this LED is the culprit? the rest take an extra frame to turn off (the entire bright pulse and fade out is 4 or so frames)

Although if this is the case I still don't know why the fault pin isn't being tripped as surely that would engage the SCP and pull the fault pin low.


I was also thinking, seeing as each of the LEDs can handle ~100mA through the AA battery test, is there any way to lower the output current limit just to test if the PSU actually brings them on (which would confirm a faulty LED at higher current)?

[rockit711]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

Set the power supply for 100mA MAX, then set the voltage to 5V. Use this setting to test the individual LEDs (or individual strips, if the voltage can go high enough). With the current limited, the voltage will drop to the voltage needed to force 100mA through the LEDs.

I just used some fresh AA batteries and the LEDs were drawing ~100mA, I assume the two I used a month ago were old. Doesn't that achieve the same thing? they each work fine with the 2AAs so I suspect if they are failing its higher than 100mA

[SLK001]

Re: "Dick Smith" 55" LED backlight issue - model GE6880

I just used some fresh AA batteries and the LEDs were drawing ~100mA, I assume the two I used a month ago were old. Doesn't that achieve the same thing? they each work fine with the 2AAs so I suspect if they are failing its higher than 100mA

Should be the same thing. How were you measuring current?

As for your movie, make sure that that your observation isn't a camera scanning artifact, by taking several movies with the camera at 0, 90, 180 and 270 degrees to the TV. If the same LED goes off early, then that guy is probably your culprit.