Need help with a LED TV. Don't want to trash it Good knowledge Great test gear

[t1d]

I am a hobbyist with a medium skill level and a nice lab. The equipment is way smarter than I am.

I am troubleshooting my neighbors TV, to see if it can be saved and to learn something about TV circuits.

DUT = Onn 32” LED TV
Model# = ONC17TV001

Main Board – SMPS and LED Driver = 3MS553D   TP.MS3553.PB780

Controller Board = MT3151A05-9-XC-2

Symptoms
Failure was a black screen on power up. No smoke, flash or bang was reported.
Presently, there is a black screen on power up, but the ghost menu can be seen with a flashlight. That makes me think that the LED driver might be dead.
I also checked across the LED wire (single pair) with the diode function and the continuity function. I got nothing. That makes me think that all of the LED strips might have at least one dead LED.

First Look
-The boards don’t show any signs of failed components… E-caps look good. Nothing scorched, or zapped.
- I powered the unit up with my Isolation Transformer/Autotransformer@120V/Dim-Bulb=60W. The bulb does not blink/bright to dim. And, it does not glow enough to be seen in ambient light. My guess is that the circuit just doesn’t draw enough amps to shine it, at all?
- I forget the Mains VDC after the rectifier diodes, but it seemed reasonable, IIRC.
- The main board only has 42V on the LED driver output, with the LEDs unplugged. That seems a little low, but I have no idea how many LEDs are behind the screen to do some rough math.
- The main board also has a red indicator LED shining inside what I think is a connector?
- The main board also has a free “PWR Upgrade” socket with 12VDC measured on it.

There is not much labeling in the way of Test Points. Forget finding the schematic, I suppose.

Is there enough information here for you to have some idea of what direction I should go in, next? I guess opening the screen and testing the LEDs can not be avoided… Any way to (reasonably safely) inject voltage through the LED wire?

How do I spec a universal driver board, if we think that is the solution. Where is a legit place to buy such?

Thanks for any and all help.

[james_s]

I've worked on several LED backlit LCD TVs with backlight problems and in every case it was a fault with the LEDs themselves. Note that the LEDs are normally in series strings with a forward drop of around 60-100V per string so it is expected you will measure nothing with your multimeter doing diode check or continuity. My 65" TV actually has all LEDs in one big string with a forward voltage of over 300V.

[t1d]

Thanks, James, for your participation and input. The mystery continues and I will post along the way, in hopes that it helps other folks.

[t1d]

It looks like the "Universal LED [TV] Drivers" are nothing more than constant current circuits. If that is true a-n-d that turns out to be my problem, I see no reason that I could not repair that on the existing board. If I knew what the current load was supposed to be, I could disconnect the LEDs and test the current supply with my DC load... Ooo, I would need to do that v-e-r-y carefully, I think...

I found some writing on the Main Board at the Controller connector. Left side is writing = Right side is the DMM measurement
GND = Continuity to frame
Key = 3V3
Key/5V = 3V3
LED = 3V3
GND = Continuity to frame
IR = 3V3
3V3_IR = 3V3
So, it looks like all the Control Rails are there...

[shakalnokturn]

Same here: I haven't repaired that many because where I source them they usually have shattered screens but a lot more often defective LEDs than drivers.

If you have a high enough voltage source you could try to light the LEDs to confirm.
If you don't have better voltage source could be Variac + isolation transformer diode and current limiting resistor.
Monitor the voltage as you crank it up.

Typical single string currents are in the 200-400mA range.

[t1d]

Same here: I haven't repaired that many because where I source them they usually have shattered screens but a lot more often defective LEDs than drivers.

If you have a high enough voltage source you could try to light the LEDs to confirm.
If you don't have better voltage source could be Variac + isolation transformer diode and current limiting resistor.
Monitor the voltage as you crank it up.

Typical single string currents are in the 200-400mA range.

I have multiple DC Bench PSUs @ 0-30V/0-3A. And, a DIY AC PSU, good to about 360WAC... I could throw a Full Bridge Rectifier on that and bracket the current with multiple 40W Halogen Bulbs.

But, I would need you to give me very exacting instructions (please and thank you,) before I attempted any test... I take it that you are talking about injecting the power through the wire that goes to the LED matrix.

Hmm... Seems like I could get (up to) 64V/400mA, by putting my better PSU into series mode... Again, the LED connector has 42VDC on it, so maybe that is the magic voltage number... If it lights up, the present driver is generating voltage, but not current, I guess... Thoughts?

[james_s]

42V may not be nearly enough. It really depends on the TV but as I said, the backlight in my TV is something like 300V.

[t1d]

Okay, I will investigate the LED supply voltage some more. Thanks for the help. Keep it coming.

[shakalnokturn]

Do you have a megohmmeter around, I find they're a handy solution to quick test the LEDs (not always a reliable test because at low current and higher voltage) but if the LED's will light, although you can't be sure they'll behave under normal supply conditions you do get an idea of the minimum string voltage needed.

Is your DIY AC supply variable?
I don't think halogen bulbs are the best for current limiting here, they have a very low resistance when cold.
What ever test voltage you start with I'd calculate the series  resistance needed for assuming Vstring is 80V and starting at something like 50mA LED current.

[t1d]

Thanks, shakalnokturn. I hope to be continuing the investigations of the LED drive voltage, today. Otherwise, yes, I can measure the LED current draw down to very small levels...

[james_s]

You only need a few tens of microamps to tell if the LEDs are at least mostly ok. A capacitor to limit the current in series with the input to a bridge rectifier will probably do the trick. A high value resistor would also work, you only need to power them for a few seconds.

[t1d]

Here is a schematic of the AC power input to the DC rectifier. I laid it out in the same way as the board (which is awful,) so that I can keep track of things. I am sorry that this approach will make it a little harder to read...

Multimeter Readings (Location = Reading)
VAC @ PCB Board Connector = 115.6VAC
VAC @ Input of Rectifying Diodes = 113.0VAC... That's a drop of 2.6VAC across the NTC and the inductors.
VDC @ Smoothing E-Cap EB1 (meaning after rectification) = 148.8VDC... Here, the math is suspicious.

(113VAC x 1.414 Conversion Factor) - (2 x 0.6 Diode Drops) = 158.58VDC Expected.
158.58VDC Expected - 148.8VDC Measured = 9.78VDC Too Low.
My guess was/is that the smoothing cap is leaking.

I ran out of steam, before I could pull the smoothing cap and measure it. But, that will be the next step.

Suggestions/Comments? Thanks for your continued interest and help.

[james_s]

Normally that layout would be the first draft and then you would rearrange it into a more sensible layout once you have the connections drawn. In this case at least I can see that it's a very conventional input section with a inrush limiting thermistor, common mode choke and full bridge rectifier with a bulk capacitor across the output.

I wouldn't worry too much about the exact number, even if the capacitor is a bit tired that isn't going to keep the backlight from working at all. I don't think it's necessary to reverse engineer the entire driver unless you are just wanting practice. The problem is likely going to be somewhere a little further in.

[t1d]

Normally that layout would be the first draft and then you would rearrange it into a more sensible layout once you have the connections drawn.

Yes, that mess is the way the manufacturer actually laid out the board. This layout is without purpose. There is plenty of room to have arranged it neatly. Actually, this scattering takes up way more real estate than what was needed.

As I mentioned, I was suspicious of a leaky cap. So, I pulled the three e-caps within the SMPS section and tested them. They all tested as perfect.

I will move on to test...
(The Driver IC is an OB3353A, which I think is a common LED TV Driver.)
- The Driver IC PWM Reference
- The Driver IC Gate Output
- The MOSFET Gate Input
- The MOSFET Drain Input verses its Source Output. Hmm... It may just be easier to pull the MOSFET and put it on the tester...
I already checked the feedback optocoupler and it looked okay. But, I did that in-circuit.

I could use some help on what to expect on the output of the secondary side of the transformer. It kind of looks like the board traces connects the five output pins together?

Once the above is completed, and assuming everything tests as good, the issue must be with the LED strips, themselves, I would think.

[james_s]

Yes, that mess is the way the manufacturer actually laid out the board. This layout is without purpose. There is plenty of room to have arranged it neatly. Actually, this scattering takes up way more real estate than what was needed.

The layout of the board and the layout of the schematic are two entirely different things, schematics are laid out to make it easy to follow the function of the circuit. The physical circuit is laid out in a way that allows all of the parts to fit in the available space and interconnect with a minimum number of jumpers, avoiding ground bounce or ground loops, using the thickest traces reasonable while maintaining adequate creepage distances, minimizing interference between parts, allowing adequate cooling airflow, etc. What makes you think the PCB is not laid out logically? I would not be so quick to assume the engineer who laid out the PCB didn't know what they were doing unless you have considerable experience designing power electronics. Schematics almost never resemble the physical layout of the circuit, that wouldn't make any sense at all, in many cases you'd end up with a schematic that is a nightmare to follow like the one you drew, or a circuit that doesn't work properly because it's laid out for form rather than function.

The way the transformer is wired depends on the topology of the circuit, I doubt there is anything wrong with that part.

As I mentioned earlier, or thought I mentioned earlier, in every single LED backlit TV I've worked on that had backlight problems, the problem was one or more faulty LEDs. I have not yet ever come across a bad driver board, which of course does not mean it can't or doesn't ever happen.

[t1d]

Thank you, so very much, for continuing to help and teach me. You are pulling together the things that I have learned and helping me to apply them practically. It is great fun. I appreciate you!

I would not be so quick to assume the engineer who laid out the PCB didn't know what they were doing

That is a point worthy of remembering… That is, that the benefit of the doubt goes to the design engineer.

Schematics almost never resemble the physical layout of the circuit, that wouldn't make any sense at all, in many cases you'd end up with a schematic that is a nightmare to follow like the one you drew, or a circuit that doesn't work properly because it's laid out for form rather than function.

I take it that you meant a theoretical person submitting a theoretical schematic, as oppose to suggesting that what I had submitted had some nightmarish failing on my part. I, of course, did not submit a schematic; I submitted a drawing of the board layout, which I had drawn in situ, for the specific purpose of sorting out the workings of the circuit.

As I mentioned earlier, or thought I mentioned earlier, in every single LED backlit TV I've worked on that had backlight problems, the problem was one or more faulty LEDs. I have not yet ever come across a bad driver board, which of course does not mean it can't or doesn't ever happen.

There are several points in play, in this regard…
- I surely have not discounted your experience, opinion, or counsel, as to the likely fault being the LEDs.
- I chose to investigate the board, first, because I find the voltage which powers the LEDs (41.6VDC) to be low (IMO,) in comparison to what my research suggests is typical. More on this to follow.
- I found it less likely that all of the LED strips had simultaneously developed at least one faulted LED, thereby producing a total blackout. More on this to follow, too.
- It is also my workflow practice to follow the circuit through from beginning to end. This works for me, because I get a better understanding of the working of the circuit (which improves troubleshooting), I am more likely to find the upstream cause of the fault (instead of just replacing a downstream damaged component, powering the repaired DUT and blowing the same component, again,) and I am less likely to miss something along the way. I learn more, too, which is fun.
- Disassembling the screen looks to have pitfalls, which I desired to bypass, if taking it apart is not absolutely required.
- With more experience, I may become confident enough to check the LEDs, as the very first step.

Additional Board Testing
- I took voltage readings on the LED Driver IC (OB3353CP.) It does not have any voltage on the Gate (output) pin. I measured the other pins and can provide those numbers, if they would be helpful.
- Accordingly, I did not find any voltage on the MOSFET gate. As before, the MOSFET Drain is at 41.6VDC.

Additional LED Testing
- I disconnected the LED connector and powered them at 41.6VDC/0.2-0.6A, directly, through the cable. No illumination resulted. The PSU did not register any current draw.
- Because I am working without certainty as to the needed LED voltage and current and thinking the circuit would provide it, naturally via its design, I repeated the test by jumping the MOSFET Drain/Source with a current limiting resistor. Again, no illumination.

More To Follows
- LED Voltage; 41.6VDC… I think that it is typical to assume that each LED requires about 3V. If so, that would mean there are maybe 12 to 14 LEDs. Because a matrix would require a certain arrangement, I am guessing there are three rows of four LEDs. Does that seem reasonable for a 32” TV?
- Is 41.6V a reasonable voltage to find driving the LEDs, even though it does not seem to fit my research?

Other Considerations
The owner reported a full failure, without any flash, bang, or boom. She turned the TV off. When she turned it back on, there was only the black screen.

As said, all of the LED strips dropping at least one LED, all at the same time, does not seem likely. So, we have several possibilities…
- All of the LEDs are on just one, single in-series string and just one failed LED was all that was necessary to black out the whole screen?
- The OB3353CP has lots of monitoring functions and is not outputting the Gate voltage, because it has detected some LED fault? Or, it has failed, simultaneously, also?

Next Steps
I imagine that you are going to suggest that there is no way to avoid disassembling the screen. If so, I would watch some YouTube videos on the process, but what tips do you have? Is there anything else I should check, before proceeding with the screen?

Please hang in there with me. I am very confident that we will find the problem. Thanks, again.

[RayRay]

IMO, first thing you should do is get a professional backlight tester to test the LED strips:
https://www.amazon.com/Mworld2-Backlight-Without-Disassemble-Application/dp/B08CKCSH92
There are numerous videos on Youtube on how to use one. Anyhows, If it's all good in that department, than that leaves you with the electronics, and you have two different ways to troubelshoot that part, first way would be to simply replace the boards (starting with the power one) by sourcing a replacement from ebay, and second one would be to troubleshoot em manually (on the component level) but this requires having the proper testing gear, soldering equipment and experience (some components can give you a false reading in circuit, or even none, and there are cases where you'd need to desolder a component in order to properly test it or verify that it's gone bad) also do note that just because a component looks physically okay, it doesn't mean that it actually is (and the same goes for detachable cables & PCB traces) so you can't take anything for granted and you'd have to test things fully. Which approach is best for you really depends on what kind of gear/experience you have, and how much time you have on your hands.

[t1d]

Thanks, Ray, for joining in and contributing.

This particular type of circuit is outside of my experience. James and Shakalnokturn have been helping me and they have been amazing.

As I posted, I have descent knowledge (hobbyist; no EE degree) and an excellent lab and equipment. So, as folks give me directions, I can usually get the tasks done in a straight forward manner.

I do want to find and correct the fault(s) at the component level. I learn more, that way, and learning is fun, IMO. Plus, it is always a kick to repair something expensive with a part that only costs pennies. I am retired, so time is not a factor.

I have seen those Backlight Testers. Even though the one you linked is not particularly expensive, I don't think I want to buy one for this one-time project. My guess is that they are just some sort of voltage booster with a constant current circuit and a MCU to sweep the power. I can do that with my PSU. Maybe I will see if I can find a circuit and build one on perf board.

I posted that I did try powering the LEDs directly. I used my bench PSU. I set the voltage to the 41.6VDC that I found on the drain of the MOSFET and swept the current from 0.2A to 0.6A. Regrettably, nothing illuminated and the PSU showed a zero current draw, throughout the entire test.

I do wonder, if those voltage and current settings were sufficient. That's where the experience with this type of circuit comes into play. If someone confirms that the voltage and amperage that I used were not correct, and they tell me what values to use, I would be glad to repeat the test.

I look forward to hearing more of your tips and tricks. Thanks!

[t1d]

I have researched the light strips, in preparation for opening the screen to test the LEDs. Here's what I think is the correct part number, but I will need to verify it, after I open the case.
Onn ZX43ZC332M08A1 LED Backlight Strips (3)
The best price that I have found so far is $19.99/free shipping/no tax @ the DirectTVParts website. That's cheap enough to not bother to solder in individual all new LEDs.

The kit contains three strips with each having eight individual LEDs, if the picture is correct. This number of strips and LEDs does not seem to fit well with the voltage available on the drain of the MOSFET = 41.6V.

The connector for the LED wire has multiple pins for multiple pairs of wires. However, there is only one set of wires going to the LED matrix. This makes me think that all 3 x 8 = 24 LEDs are in series. If so, the forward voltage would be approximately 3V x 24 LEDs = 72Volts. That would explain why no LEDs illuminated during testing at 41.6V. Yes, there is the possibility that parallel wiring could exist inside the case.

If the LEDs are not wired in a single series, then the three strips would likely be wired in parallel. That would make for a forward voltage per strip of 3V x 8 LEDs = 24Volts. 41.6V might make sense, for the parallel scenario, if there is current limiting.

I could repeat the LED test at 72V/0.02A, but, at this point, I am willing to risk opening the case to test the individual LEDs, directly.

It also occurred to me that I might be able to breadboard a LED matrix with common white 5mm LEDS, to test the LED Drive circuitry. I would think that the drive circuit would output a constant current and that no resistors are needed on the strips. If that is not the case, then I would need to know whether the LED strips use a single resistor, per strip, or individual resistors, per LED. And, I would need to know what the correct drive voltage is with certainty. But, maybe that effort can be postponed, until I have checked the individual LEDs.

I have not slept. For clarity of mind, I will not proceed, until I have rested.

Thoughts and suggestions continue to be very welcome.

[RayRay]

Seems you've already figured out most of the things I've wanted to mention already.
Yeah, the specific board is associated with the ZX43ZC332M08A1 V0 led strips (which seems to be 3V) and the overall LED voltage is suppose to be equal to all of em combined (3 x number of LEDs) so 41.6v isn't enough, and also, you may have done some damage with your testing (depending how it was done). If you've tested a single strip (of 8 LEDs) with 41.6v, you've most likely fried it, but if you've outputed it to all of em, that's a different story. In my opinion, it's always best to use the proper tool for the job (and a backlight tester is a must have for TV repairs) but since you currently don't have one, and that replacing the LED strips would cost the same as getting the tester, getting the LEDs replaced would be more cost-effective. However, if the power board isn't outputting enough voltage to the LEDs, that's another thing that would need to be addressed, and I'd suggest you start by checking the diodes, capacitors (capaciatance wise), the voltage regulators, and the resistors (for standard thru-hole resistors, look up "resistor color code calc" on Google) or alternatively, you could source a replacement board from ebay instead. Anyhow, hope you'd have a good rest

[james_s]

I take it that you meant a theoretical person submitting a theoretical schematic, as oppose to suggesting that what I had submitted had some nightmarish failing on my part. I, of course, did not submit a schematic; I submitted a drawing of the board layout, which I had drawn in situ, for the specific purpose of sorting out the workings of the circuit.

The schematic you drew is good enough for this particular case because the circuit is a very typical arrangement that I have seen before many times so I was able to recognize it and rearrange it in my head to make sense. If it was something more complicated or less familiar I would consider that as a rough draft and draw it again in a more logical layout for a schematic, this time ignoring the physical layout of the board. You seem to be mixing and matching terms here. A schematic is a diagram of the circuit wiring using standardized symbols to indicate the components and has nothing to suggest physical layout. A board layout is a drawing of the physical traces, pads and components of the PCB and does not utilize schematic symbols. A schematic laid out like the board layout is a nightmare to follow, if the circuit is anything not easily recognizable.

[Labrat101]

Hi it sounds like you have a couple of open leds .
As you said you have 42 volts and what i could see from the photo that's 42v on 2 x 2 rails being 32 inch would make it 4 rows with 7 leds per row . which is correct .
the led rows are normally any were from 6 plus depending on make Model.
The voltage you will read though a normal multimeter will not be the full load voltage . due to the meters high resistance .
As quoted by RayRay the correct tools are the best a Led Tester
I think you are going to have to open the screen . which is a little tricky .
I have done this many times . If only 1 or 2 led are bad I remove just the bad leds and replace with stick over replacements that sit on top of the strip and solder 2 wires.
I buy boxs of 50 these kits as they work with 99% of most tvs / monitors
If you are going to open the screen, the panel on the bottom must not be touched . so cotton gloves are required also for handling the 3 polarized sheets should not be separated and removed  together . I believe there are some videos on this operation on YouTube

[james_s]

The glass is also very thin and freakishly fragile. I've repaired a handful of sets from 40"-65" with success but a friend of mine just broke a 65" the other day when he was too impatient to wait for a time I could come over and give him a hand. 32" is much easier to handle than the larger ones though, anything over about 55" and the glass will flex and bow like crazy even if you're careful.

It actually really annoys me how these things are built. In every one I've worked on, both back and edge lit, it would have been trivial to design it such that the LED strips could be easily removed without disassembling the entire panel, it wouldn't have to cost more to build even, maybe the cost of a few screws. They're all built so that it is an expert level project to service the LEDs though.

[t1d]

Seems you've already figured out most of the things I've wanted to mention already.
Yeah, the specific board is associated with the ZX43ZC332M08A1 V0 led strips (which seems to be 3V) and the overall LED voltage is suppose to be equal to all of em combined (3 x number of LEDs) so 41.6v isn't enough, and also, you may have done some damage with your testing (depending how it was done). If you've tested a single strip (of 8 LEDs) with 41.6v, you've most likely fried it, but if you've outputed it to all of em, that's a different story. In my opinion, it's always best to use the proper tool for the job (and a backlight tester is a must have for TV repairs) but since you currently don't have one, and that replacing the LED strips would cost the same as getting the tester, getting the LEDs replaced would be more cost-effective. However, if the power board isn't outputting enough voltage to the LEDs, that's another thing that would need to be addressed, and I'd suggest you start by checking the diodes, capacitors (capaciatance wise), the voltage regulators, and the resistors (for standard thru-hole resistors, look up "resistor color code calc" on Google) or alternatively, you could source a replacement board from ebay instead. Anyhow, hope you'd have a good rest

Thanks, Ray. Good info.

Yes, I tested all of the LEDs at once with 41.6V, because that is the voltage present. My thinking was that level of voltage was the most logical number, based on the information available. No smoke, flash, or smell.

The economics of making the repair are not looking good... Strips, tester and unknown board repair that is seeming more likely... So, this may just end up being a learning lesson, but that is well worth the time and much fun. Talking electronics with good folks like those of you helping me is a bonus.

The neat little testers seem to be not much more than a high voltage source, a constant current source and a voltage meter. If I understand correctly, the voltage settles, naturally, to whatever is needed, based on the number of LEDs, because the current is fixed and that's what matters. My bench PSU will do those functions, but I have read comments that (maybe?) they don't do as well as the testers. I forget the magical difference.

I do need to investigate the voltage supply more in depth... Verifying the components, as you say, looking at the voltage reference and driver PWM... That means I will get to play with my scope = Woot!, as the kids say.

Putting my scope on high voltage SMPSs makes me a bit nervous, but I have a proper isolation transformer. The whole TV is floating, for that matter... No earth pin on the plug.

I did get some rest, thanks for the well wishes... Long term health stuff. No fun. Prayers appreciated.

[t1d]

The schematic you drew is good enough for this particular case because the circuit is a very typical arrangement that I have seen before many times so I was able to recognize it and rearrange it in my head to make sense. If it was something more complicated or less familiar I would consider that as a rough draft and draw it again in a more logical layout for a schematic, this time ignoring the physical layout of the board. You seem to be mixing and matching terms here. A schematic is a diagram of the circuit wiring using standardized symbols to indicate the components and has nothing to suggest physical layout. A board layout is a drawing of the physical traces, pads and components of the PCB and does not utilize schematic symbols. A schematic laid out like the board layout is a nightmare to follow, if the circuit is anything not easily recognizable.

Good points, James. Please forgive me for mixing terms. I guess that those kinds of communication issues are inherent to a website that mixes hobbyist with engineers. But, three cheers for EEVBlog/Forums for the information, help and fun that is provides, right?