Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

[Th3_uN1Qu3]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

Either way I think it's ridiculous that these TVs fail after 2-3 years, it must be possible to design something that actually lasts......

Sure it's possible. When i was a kid our TV set was older than me by 10 years, but why do that anymore?

But then again no-one would buy the latest model if the old stuff still works perfectly good!

Not really... but people who buy new ones would sell their older sets, which means someone gets a cheap and working TV, the previous owner gets some money back, and the company making the TVs doesn't get sh**. Not a very profitable business model.

[tom66]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

So I found a 1N5820, along with two other SMD diodes the right size but of unknown voltage/current.

The 1N5820 has a smaller reverse breakdown voltage but the same current, but I figure the max reverse voltage applied across it will be the supply voltage, 12V, so it should be okay. It's through hole, but I've soldered a through hole diode to SMD pads before, it just requires some care.

[PlainBill]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

So I found a 1N5820, along with two other SMD diodes the right size but of unknown voltage/current.

The 1N5820 has a smaller reverse breakdown voltage but the same current, but I figure the max reverse voltage applied across it will be the supply voltage, 12V, so it should be okay. It's through hole, but I've soldered a through hole diode to SMD pads before, it just requires some care.

That is incorrect. The maximum reverse voltage will be at least TWICE the output voltage. With an inductor in the circuit, transients would be even higher. A 1N5820 would be marginal at best.

PlainBill

[tom66]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

Hmm. I'm a bit worried now. I just noticed the class-D audio power amp, TPS3002D2, has several holes in it.

I checked the datasheet. It's rated for a 15V supply, absolute maximum. The design has it running off an 18.5V rail. However, I would have thought any failure brought about high supply voltage would occur within hours of using the TV due to some semiconductor breaking down.

So, with bad caps, does the output of a PSU tend to increase? Maybe it had very high ripple on it which killed the amp chip, or maybe a power surge killed it, but I've never seen that happen before.

This chip is on the "slave board", as I call it. The slave board has the amp chip, sound processor (MSP4311), and the two analog tuners, supporting PIP (picture-in-picture?), POP(?) and PIP-POS(??). Honestly I don't see why they put two main boards in it...

[tom66]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

That is incorrect. The maximum reverse voltage will be at least TWICE the output voltage. With an inductor in the circuit, transients would be even higher. A 1N5820 would be marginal at best.

PlainBill

Well, I lost the diode anyway, so I'm making an order for the right one.

I've used 40V Schottky diodes in small power supply designs (rated to go to 30V) with no problems, but it might depend on the particular topology or how the power supply switches (current mode, or something else.) Higher voltage diodes are bigger, and these power switchers don't look like they are particularly size optimised, my application is in model aircraft so size is a big factor in the design.

[Th3_uN1Qu3]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

Hmm. I'm a bit worried now. I just noticed the class-D audio power amp, TPS3002D2, has several holes in it.

You'll need to use a heat gun and lift that off the board before you try and start the TV again, because it is likely to take out the power supply again if it's shorted. And it obviously is - it wouldn't have blown holes in it otherwise.

So, with bad caps, does the output of a PSU tend to increase?

The average output voltage stays the same because it is regulated by the control loop. The peak voltage however gets much higher. High enough to make silicon break down in your case.

I've used 40V Schottky diodes in small power supply designs (rated to go to 30V) with no problems, but it might depend on the particular topology or how the power supply switches (

It depends on the transformer and inductor. Sometimes it works, sometimes it doesn't. Better safe than sorry.

[tom66]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

You'll need to use a heat gun and lift that off the board before you try and start the TV again, because it is likely to take out the power supply again if it's shorted. And it obviously is - it wouldn't have blown holes in it otherwise.

The chip does appear to be failed. However, the supply is switching on and off and all rails are present. No shorts. So it's very odd; most semiconductors like that fail shorted.

I'll try it and see if by some miracle the sound still works. I'm hoping it does, but I know there's slim to no chance that it will.

I do have a hot air gun but removing those types of ICs is not easy because of the large thermal pad under the chip.

Was considering building a small class AB amp as a little project to replace the class D which has failed.

[Th3_uN1Qu3]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

I do have a hot air gun but removing those types of ICs is not easy because of the large thermal pad under the chip.

Heat it gradually and leave the heat on for longer than you would normally would. Burning the PCB a little is still better than having the temp too low and ripping pads out, or worse, having the chip slide instead of coming straight up, and wiping out like 20 tiny resistors and caps.

[PlainBill]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

I think on something like that I would try a hybrid approach. Heat each pin up with a soldering iron until the solder melts, then lift the pin using a dissecting pick. A little ChipQuik would be a good idea, too. Alternatively, use the Chipquik, then heat it with a hot air gun.

PlainBill

[Th3_uN1Qu3]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

I think on something like that I would try a hybrid approach. Heat each pin up with a soldering iron until the solder melts, then lift the pin using a dissecting pick.

It's too small for that. But ChipQuik is a good idea.

[tom66]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

I think on something like that I would try a hybrid approach. Heat each pin up with a soldering iron until the solder melts, then lift the pin using a dissecting pick. A little ChipQuik would be a good idea, too. Alternatively, use the Chipquik, then heat it with a hot air gun.

PlainBill

In theory, that might work for a plain TQFP, but it's got a large thermal pad underneath. That's not going to come off without a lot of heat.

It might sound crazy, but I'm either going to sell it without sound (I'm nowhere near experienced enough to replace an IC like that) or make my own little amplifier for it based on a TDA7269 or something similar on a bit of stripboard... would be some good experience . I may have to isolate the IC if it's causing problems.

[tom66]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

Okay, well it took me a while to get the diode but I did replace it... and... nothing.

+5V (Vcc) still measures 0.22V.

Then I checked the new diode, shorted at 0.04V - again...

I removed it, and it wasn't shorted...

So... I checked the AP1501 switcher, and it is shorted. Not quite shorted though; just 33 - 34 ohms to ground. The other switcher on the board measures 550 ohms to ground, which is still a bit low. But I'll tackle that one later if it's faulty.

How common is it for these switchers to get shorts to ground?

[Th3_uN1Qu3]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

Anything that has a power MOSFET in it is likely to die shorted... But 33 ohms on a 5v regulator is far from being called a short.

Do you have a power supply with adjustable current limiting? Sounds like a job for one. Apply 5v and slowly bring the current up. If it goes over 3A (which is the limit of the AP1501 regulator in question) and the voltage still reads way under 5v, the shorted part will quickly start releasing that familiar smell. On the other hand, if the board comes up with no other issues then it is safe to say the regulator is bad.

If you don't have a current limited supply, grab a bunch of old power transistors, a big heatsink and a potentiometer and i'll help you build something you can wire on the output of a PC PSU for this purpose.

[tom66]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

Anything that has a power MOSFET in it is likely to die shorted... But 33 ohms on a 5v regulator is far from being called a short.

Do you have a power supply with adjustable current limiting? Sounds like a job for one. Apply 5v and slowly bring the current up. If it goes over 3A (which is the limit of the AP1501 regulator in question) and the voltage still reads way under 5v, the shorted part will quickly start releasing that familiar smell. On the other hand, if the board comes up with no other issues then it is safe to say the regulator is bad.

If you don't have a current limited supply, grab a bunch of old power transistors, a big heatsink and a potentiometer and i'll help you build something you can wire on the output of a PC PSU for this purpose.

I do have a 0-35V 0-3A CC/CV supply but it weighs a ton! I suppose I could try the smoke test but umm... honestly I want to see if there's a safer way of doing it.

Problem is, this TV is at my dad's office. My scope, power supply, soldering gear etc. are all at home. So it's a bit difficult to test the TV without bringing it back to my place and then I'll be stuck for space.

Here's the odd thing. There doesn't appear to be a short on the 12V side, because that measures ~11.93V. The short is probably after the buck converter and that must be current limiting. The converter is barely warm though, if even warm at all.

Could I try removing the inductor, and seeing if the short persists on one side? Then I would have identified if it is the converter or the main board somewhere.

I could always buy a new main board - I see them going for about £20 on eBay, so it may be worth it. However, removing the main board (LVDS) connector is going to be difficult, for some stupid reason it's glued onto the board, I will have to be careful in removing it (it is in a connector, but then they've added some glue.)

[Th3_uN1Qu3]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

Could I try removing the inductor, and seeing if the short persists on one side? Then I would have identified if it is the converter or the main board somewhere.

Yes, that sounds like a good idea.

[tom66]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

Well, I decided to order a new main board anyway (only £21, not bad for TV parts), but just for the sake of satisfying my own curiosity I will test to see where the short is when I exchange the boards.

[tom66]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

Arrhhhh

I ordered the main board but the PSU connector is ever so slightly different, one pin too long. Everything else fits.

Looks like I will have to hack something together to get it working. I suspect I'll just have to route 12V, 5V, GND and STBY to the right pins. Everything else should be okay.

[tom66]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

Okay, so on my old board, I lifted the pads for the inductor and the short is on the switcher's side. 33 ohms to ground. Well not really a short but it shouldn't be there anyway. Might be a sign that the chip inside is toast.

So, I will either fix the old board, or try and get the new one working...

[tom66]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

Okay. So I tried removing the AP1501, no dice. It's rock solid there.

But I wanted to see if the TV actually worked at all... if it was worth fixing it. So I found a 7805 in my component bin. And I soldered wires to it, and soldered those to the board. So I had a freestanding (no heatsink) 7805 attached to the back of the board. Reassemble it - and it works... but only for a few seconds, which I expected, because the regulator quickly overheats.

When it works, briefly I get the text "Searching" then "SCART2 No Signal". Then the display freezes, dots appear randomly, then it goes blank for a second or two and the screen slowly fills with a red/rainbow array of colours. Very interestingly, the person I purchased this from reported that they were watching their TV and it suddenly froze and then the screen turned into an array of pretty colours. Sounds very similar to what happens when the 5V reg overheats - I'm thinking the switcher popped (possibly due to bad caps in the PSU) - and this took out the 5V line, causing the effect described.

So, I just remembered I have a little "battery eliminator circuit" (BEC) for model aircraft, which puts out 5V for a 5V-23V input, at up to 3A. Not currently in use because I have one built into my motor speed controller (I fly model aircraft as a hobby.) It is a buck converter so shouldn't overheat. Guess what I'm thinking...

Also, due to my idiocy, I threw out a lot of the screws holding the case on for this TV. I have a few lying around, anyone know where I can buy the type of screws used in these TVs? I want to assemble this and probably keep it as a nice computer monitor.

[tom66]

Re: Lexsor 32" LCD (Model Unknown) w/ 24V bus measuring 1.4 - 1.8V

BEC installed. Works great!

Well, as expected the speaker audio is dead, but the headphones output does work.

The first picture shows the original mod. (I didn't get a picture of the BEC mod.) The second picture shows the TV working. I need to screw it together properly and fit the stand, but it's looking good.