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

[tom66]

I have completely recapped this power supply (with Panasonic FM, Chemicon KZE, Nichicon PW & VR): KAS200 inside this Lexsor 32" LCD TV. The PSU is also used in Polaroid TVs I believe. It's a 2-layer board, so a nightmare to debug and repair, as the pads have a tendency to peel off.

Three caps were bulging on the secondary side but I replaced *all* secondary side caps. I haven't replaced anything primary side.

Before the re-cap most rails were measuring way under voltage.

Now only two are out of spec:

22V: measures 18.5V (pretty close - probably okay) - In fact the sticker mentions 18V so maybe it's just incorrect silkscreen?
24V(for inverter): measures 1.4V with inverter connected, 1.8V disconnected

What could cause the 24V to be so low? It's not a faulty inverter loading it down, because it's only 1.8V with it disconnected.

The symptoms at the moment are that the LED glows red in standby. When pressing the power button, it turns blue, but it stays like that until unplugged - no picture appears.

[Th3_uN1Qu3]

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

Bad diode maybe?

[tom66]

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

Bad diode maybe?

If that's the case... Argh! The diodes are right between two heatsinks sandwiched together. (Top right of first pic.) Looks like I'll have to desolder all the diodes, and the heatsinks, just to get at it and replace it.

Would it be possible to solder a diode in parallel (with a good heatsink)? Maybe I can cut the traces to the current diode, solder to them, then mount a heatsink on the back of the LCD panel. It would be ambitious, but it might work. The 24V rail is rated at 6A, so I'd need a good heatsink to dissipate at least 3-4W. I suspect it's a dual common cathode diode.

Is the fact that one half of the heatsink is visibly discoloured evidence that some significant heating has happened in this area?

[Th3_uN1Qu3]

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

The diode may be leaky or shorted, in that case simply putting one in parallel won't work. You do have access to its pins, don't you? You can test it with your multimeter to see if it's good. It's sometimes a pain in the ass but a clean job is better than an ugly hack IMO.

[tom66]

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

The diode may be leaky or shorted, in that case simply putting one in parallel won't work. You do have access to its pins, don't you? You can test it with your multimeter to see if it's good. It's sometimes a pain in the ass but a clean job is better than an ugly hack IMO.

So, do you think a leaky/shorted diode would cause a DC ~1.8V output? Is it possible the diode is very weak, so it's not able to charge the filter caps, if that makes any sense? I've also heard the toroid can be problematic, so I will try and check that - any idea how to? Oh and yes I do have access to the diode pins.

[Th3_uN1Qu3]

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

Then what are you waiting for - use the diode check on your multimeter and test it. If the diode were shorted the power supply would have likely been into protection mode - i'm betting more likely on it being open, and those 1.8v bleeding in from somewhere.

[tom66]

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

Then what are you waiting for - use the diode check on your multimeter and test it. If the diode were shorted the power supply would have likely been into protection mode - i'm betting more likely on it being open, and those 1.8v bleeding in from somewhere.

I'm away from the set at the moment, and may not be able to get there for a week... But I'll post back when I do the test.

[tom66]

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

Just did a diode test, leg 1 -> 2 tests 0.196V forward, OL backwards; leg 2 -> 3 tests 0.196V forward; OL backward.

So the diode looks good.

I had to do a lot of bodging to get this to work, as the pads lifted many times. I just noticed I may have shorted the GND of the filter caps to the diode output. So I'll have to correct it, hopefully that's my problem. (Would be easier with a schematic.)

[tom66]

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

I was able to fix the mistake I made; same problem.

There *is* switching going on with the rail, I can see the ripple on an oscilloscope, but it rises very slowly to 1.8V; whereas the other rails rise almost immediately.

Even though the diode tested good, do you think it is faulty? Is there a sure fire way to test it?

[PlainBill]

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

I was able to fix the mistake I made; same problem.

There *is* switching going on with the rail, I can see the ripple on an oscilloscope, but it rises very slowly to 1.8V; whereas the other rails rise almost immediately.

Even though the diode tested good, do you think it is faulty? Is there a sure fire way to test it?

Since you have a scope, check the voltage into the diode. If the secondary of the transformer is providing pulses of only 2.5 volts or so there is a short somewhere. If the pulses are about 25 volts, it's the diode.

PlainBill

[tom66]

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

Very small waveform on pins #1 and #3. Seems like a dud toroid. The other rails have quite significant switching waveforms. (see pics) For example, for what I think is the 12V diode, approx ~13.3V is appearing on #1 and #3.

I guess it means I've got to desolder and rewind the toroid? Or buy a new one from... somewhere.

How am I going to remove that toroid as I only have a 35W soldering iron? (it's temperature controlled, though.) I imagine it's going to suck a lot of heat before it budges.

[tom66]

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

Just checked the toroid. *All* secondary pins have continuity to ground. Is that normal?? I'm thinking it must be, because all other rails work...

[PlainBill]

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

Just checked the toroid. *All* secondary pins have continuity to ground. Is that normal?? I'm thinking it must be, because all other rails work...

Yes, that is normal. One side of each secondary is tied to ground.

PlainBill

[tom66]

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

Yes, that is normal. One side of each secondary is tied to ground.

PlainBill

So, do you think there is some problem with the toroid causing no waveform to appear on the 24V tap? Or some other issue causing this?

[PlainBill]

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

So, do you think there is some problem with the toroid causing no waveform to appear on the 24V tap? Or some other issue causing this?

I don't believe the problem is with the toroid, but it is easy to suggest the steps to test this theory.

1. Remove the rectifier diode on the 24V rail. Measure the resistance from the pad(s) the anode was soldered to to ground. It should read less than 1 ohm.

2. Measure the resistance from the pad(s) the cathode was soldered to to ground. It should read over 10 ohms.

3. Hook up the power supply and measure the waveform on the pad the anode was connected to. The pulses should rise to over 25 volts.

4. Hook the anode of the diode to a source of over 12 volts. Connect a 10 ohm load from the cathode of the diode to the negative terminal of the power supply. Measure the voltage across the diode. It should be .5 volts or less.

5. Apply a current limited sourve of 24 volts to the pad the cathode was connected to. Measure the current.

PlainBill

[tom66]

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

Hehe... you make it sound so easy. :O Anyway, I can't directly remove the diode (without completely removing the heatsink, and thus the other power rails), but I was thinking I could just wick away the solder and break the joint, however I'm not sure how well that will work.

I really need a schematic of this PSU to debug it, but no luck thus far...

I have just noticed something. The output filter caps (820uF/35V Pana FM) pos trace goes to the filter coil, but there is no continuity to that pin. In fact, one of the coils has no continuity to any other part of the secondary. I'm thinking that I've lifted a trace, because it's two layer, I didn't notice it, but it's gotta be what connects the 24V to the diode through the LC. Otherwise there's no path for the current. Would this be likely to cause no switching waveform though? With no output (nor any load caps for the inductor) I'd expect there to be more switching than anything else, but I'm probably wrong...

[tom66]

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

Bam! Fixed my error with the lifted trace (bridged it with some red wire.)

Output on 24V is now 24.16V (no load) - perfect This should now work in the TV. (Before re-cap the 12V rail was at 8.5V, and other rails were way out of spec. I'm guessing the main reason was bad capacitors.)

I'm guessing the 1.4 - 1.8V was coming from some sense circuit or bleeding in from other rails.

I *really* need to improve my desoldering technique, because I lifted so many traces, and had to bridge them and do loads of nasty stuff! Luckily the board has a plastic shield inside the TV so none of my bridges can short out to the casing though I'm very tempted to hot glue them down.

Often I find the holes plug themselves, so I used wick, it works okay, but it requires a lot of heat before they will completely unplug, and some don't do it without some extra solder. So it's a bit hit and miss at the moment. I've tried topcat's dental pick idea (but with a needle), it works, but again it's hit and miss, always a bit too much in there for the cap lead.

(the little quarter watt resistor dangling from the board is to turn the power supply on, it's just wrapped around two pin contacts: PSON and STBY-5V.)

[tom66]

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

Okay-

I put the board back in the TV. Still won't turn on: same problem. LED goes from red to blue, but no sign of life on the panel or with the inverters.

So I opened up the main board, which is encased in shielding. Took some work, but eventually got it open. There looks to be another board, which seems "slave" to the main board. It only has a tuner, speaker and SCART connector on it, and all connectors come from the main board except for one from the power supply (providing 18V.) I'm guessing it adds Freeview/DVB or analog TV to this set plus contains the audio amplifier for the side speakers. Similar to my Goodmans TV I'm repairing - it had a separate board just for Freeview, seemed like an odd design decision.

I noticed across the board, there is only power available in spurious parts, mostly standby. None of the core display processor voltages are there, and most of the LDOs have sub 1V outputs. I doubt all the LDOs are bad, so I am thinking that something which provides core power is bad.

I have successfully located the service manual. The PSU schematic is different but the main board schematic (pp 39-51) is the same.
Available for free on Elektrotanya :

There seems to be a typo across the whole document. "+23V" rail should be "+12V". This is what the power supply is outputting on the corresponding pins, at least.

I turned my attention to the two on-board switching regulators both based around AP1501 chips (integrated 3A buck converters.) They are on page 50.

When the on/off button is pressed, the LED changes from red to blue, the PSU on/off pin goes to 4.8V, and the power supply begins outputting 12V, 18V and 24V. The 12V goes to the main board. 18V seems to go to the slave board, possibly for audio. 24V only goes to the [correction: inverter]. 5Vstby is always provided to the main board. All of these voltages check out and all are within 2% of their rated voltage.

N704 should be converting 12V to 5V, for core logic. Input voltage is 12V. The SD pin measures below 1.3V, so that should switch it on, according to the AP1501 datasheet. But the output voltage is 0.77V, very low; the AP1501 can only output as low as 1.23V, so something is wrong... The schematic shows it set up to output 5V, and shows it is being forced on, the CPU has no control over it.

Checked the diode D701 for N704. Shorted @ 0.032V, conducting both ways. So I'm guessing this is the fault. Inductor tests DCR @ 0.2 ohms. The caps are cheap, but they seem okay from a casual inspection.

N705 is the second buck and seems to provide additional 5V, but there is nothing there (output = 0.0V). It is turned on by the CPU. The SD pin measures at 1.11V. It's lower than the 1.3V threshold, but the chip might be a different revision or have an internal clamp, or the tolerances could be poor on that pin. D702 for it checks out (not shorted), so I'm hoping N705 is just in shut down mode.

How's my troubleshooting?

[Th3_uN1Qu3]

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

Checked the diode D701. Shorted @ 0.032V, conducting both ways. So I'm guessing this is the fault. Inductor tests DCR @ 0.2 ohms.

How's my troubleshooting?

So far it looks your troubleshooting is very good. Report back once you change that diode.

[tom66]

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

It's a 3A 40V Schottky with 0.5V max forward drop, model SK34. I'm wondering why it has failed. Perhaps it was just hot and near the end of its life... but this TV is less than 3 years old, which is pretty poor really. I could probably install a 4-5A SMD Schottky in the same space with newer, lower ~0.35Vf.

This reminds me, I repaired a DVD player with a failed Schottky diode in the power supply but it wasn't SMD. I've always thought of diodes as reliable devices, but I guess under repeated thermal stress they can fail, like any other component.

I'm almost wondering what caused this set to die, I'm not entirely sure the caps were the problem (but would have been soon). I was told by the original owner, who I purchased it off, that they were watching it and the screen just turned to rainbow coloured lines, after turning it off it would not turn on again. That's somewhat consistent with the display processor crashing so it might have just been the time for the diode to pop. Maybe higher peak currents were a result of the capacitors slowly failing causing the output voltages to drop on the supply, which in turn caused more stress on the diode.

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