Switched Mode Power Supply (SMPS) Problem

Thread Starter

daveLNER

Joined Apr 25, 2020
6
I'm looking for some advice on fixing a SMPS (LAD6019AB5) based around the 3842ADM chip. It's supposed to run a small 19" TV.

The fault is that essentially it doesn't work. Which isn't strictly true in it's entirity.

To elaborate..... With the TV connected to the SMPS PSU, the PSU is turned on/powered up, the TV doesn't work - (almost) no output, only around 3v at the output. On the other hand, disconnect the TV from the PSU, power up the PSU and I get the normal 12.25v. Usually when the TV is then hot-connected to the PSU (i.e. still powered on), the voltage disappears back to the 3v mark. Occasionally, however, the TV sparks up and works normally. Once we get to this situation of the TV/PSU working together, the TV can be powered off and on (in/out of standby) without problems - the PSU continues to work normally.

Turn off the PSU at the mains and when it's powered back on with the TV connected, nothing again.

When all is working, the TV takes 0.1A in standby and 3.85A when powered up/working. The PSU is rated at 5A output @ 12v. Incidentally, the TV is good (tested on another power unit).

The PSU has no apparent visible damage (burnouts, etc. anywhere), no aroma of anything previously getting hot.

These SMPS power supplies are a mystery to me (I get the basic principle, but not the detail of why they can go wrong - other than blown diodes/FETs, etc. for which I've checked - if anything such as these components were defective, I'm sure the PSU/TV combo would never work).

Is anyone able to give me a clue as to where to look / what possible components to change?

Thanks in advance - and in hope of a fix.
Dave
 

dendad

Joined Feb 20, 2016
3,913
Firstly, be very careful when working on these.
Then, the #1 suspect in SMPS faults are the electrolytic capacitors.
Remove them and measure the capacitance.
If low, replace them, but ensure you use low ESR types.
Ensure they are discharged before working in the supply.
 

Thread Starter

daveLNER

Joined Apr 25, 2020
6
So it's Not the Psu then, it must be the TV with a fault.
Thanks Dave. How do you work that out then?
The TV is fine, no issue, works perfectly, time after time, after time, after...... with another PSU.

I've tried simulating a "load" onto this PSU output - guess what - nothing, dead-ish, output ranges from 1v to 3v with differing load values.

It's the PSU.
The PSU's cream-crackered! Unless there's no load on it - which is pretty useless if it won't run anything, er, like a TV.

But thanks to dendad for the suggestion of the electrolytics - I did have my suspicions about them but wanted to ask the audience just in case there was something else I should be looking at as well (or instead of). These closed-loop things can give odd results with a fault remote from what is suspected.

I'll have a dig around the spares box and see if there's anything suitable - if not I shall have to go get some from elsewhere.
 

Dodgydave

Joined Jun 22, 2012
10,050
Ok if it's the psu it will be usinfg the UC3842 chip, for what its worth i would just buy another 12V 5A psu rather than messing around with trying to fix it, could be capacitors, or over current limiting.
 

Thread Starter

daveLNER

Joined Apr 25, 2020
6
Thanks guys.

All the capacitors "look" fine visually - not domed, not leaking, not lifting, but as well all know, looks aren't everything. The PSU is quite old (the 3842 chip is now obsolete so that should give an indication as to its age) and has been left unused/unpowered for some while. I'll remove the caps and see if I can glean anything regarding value on a meter. Looking at the webpage it shows the CapXon brand - same as I have in the PSU. So maybe it's worth swapping them out anyway - even as a precaution. For the few coppers it'll cost, not worth messing.

As regards just going out and buying a new PSU - I agree that would be the simplest option, probably the quickest - but I like a challenge! If I can fix something then I've learnt something in the process. And part of the reason the question was raised on here was to get ideas as to where the problem might lie if not the capacitors, i.e. as you indicate, an over-current issue. Now how or where to find that and fix it is a step too far for my current level of understanding of how these things work. But I'm wanting and willing to learn these things. Even at my time of life I still want to learn.

I'm off now to source some low-ESR caps and bung 'em in and see if she flies again.
 

Dodgydave

Joined Jun 22, 2012
10,050
Here is an example circuit of the UC3842 in use, pin 3 is current sense on the Source pin of the mosfet to monitor overload on the transformer.


UC3842 Flyback Converter.gif
 

DarthVolta

Joined Jan 27, 2015
521
How's the feed back network ? With the PSU unpowered, and 12V connected to the CATHODE side of the SR304, what DC value does the TL431 have on it's pins ? I'd check that too. Some protection is kicking in, turning off the chip

Vref from the chip, at least when active, should be 5.00V, what does it read, what about when unpowered and you apply 5V, it should barely pull more current than the opto collector/emiter, if on. Yo can turn than on from the other side too.

You can check the 18V zener too, do you have an adjustable PSU ? I'm used to thinking most people on here do for some reason, I just got 1 last year. Help's a ton for checking unpowered circuits, by carefully applying voltages here and there, and seeing what voltages are where, and the current draw too.

The chip itself can do a lot of funny thing's, probably a whole lot of parts inside it that can go wrong over time.
 
Last edited:

AlbertHall

Joined Jun 4, 2014
11,505
How's the feed back network ? With the PSU unpowered, and 12V connected to the ANODE side of the SR304, what DC value does the TL431 have on it's pins ? I'd check that too. Some protection is kicking in, turning off the chip

Vref from the chip, at least when active, should be 5.00V, what does it read, what about when unpowered and you apply 5V, it should barely pull more current than the opto collector/emiter, if on. Yo can turn than on from the other side too.
First, we don't know that the power supply in question uses exactly that schematic.
Second if you connect 12V to SR304 anode in that circuit it would be shorted by the transformer secondary.
 

Thread Starter

daveLNER

Joined Apr 25, 2020
6
Following my day off from electronics (to do some other chores!) I appreciate the comments and assistance guys - and also the illustrative circuit diagram for the UC3842 chip. Some differences though to the PSU (as might be expected!) given the chip is the 3842ADM version ) whatever the differences may be.

The output smoother caps are each 1000uF 16v in parallel, the slightly smaller cap next to the LED is also 1000uF 16v. The small cap near the 2 diodes is a 47uF 25V. The large cap next to the mains connector is 150uF 400v (needed to lift it to see its value). None of the caps are bulging or leaking (looking good). The FET (Q01) next to that is a (2S)K2996 and the 2 diodes (D03, D07) are LT7210 MBR20100CT which appear to be in parallel, for which I can find no information other than it's a diode of some description, possibly a zener (?).

In-circuit resistance checks:
R01 (next to the bridge rectifier) whilst it looks like a 2mohm (or less likely a 55ohm) measures 186kohm one-way, 196kohm t'other way - which might be about right when everything else is taken into account.
D03, D07: Across the 2 diodes: 229ohm one-way, charging t'other way to 5k2ohm.
D02 (near FET), D?? (near U04): Both dodes appear to check-out ok.
Q01: Across the FET: G-S 927ohm (diode setting-neg to Gate)/8k75(20k setting) - D-S 700ohm (diode setting-neg to Drain)/-.
Surface mounted - Q02 (marked 1AM)
Surface mounted - Q03 (2AN)
Surface mounted - Q04 (1AM)
Surface mounted - Q05 (1AM)
Surface mounted - Q06 (1AM)
No real resistance measurements on the above SM devices, they all seem to have some suspect measurements from normal transistor expectations.
U02: might seem a little high KIA 431A but being a programmable shunt reguator (the datasheet says!), this may be correct.
U05: H. X 1225 72.
U04: H. X 1225 72.

Not much else I can tell you at this stage - not had chance to follow any of the hints as yet - but thought the photos of both sides of the board might be more useful in the short term.

Attached:
example circuit of UC3842 Texas
top side of pcb
underside of pcb

Hope you can find something of use in what might be going on.
Thanks
Typical Circuit.jpg20200427_075501.jpg20200427_075417.jpg
 

AlbertHall

Joined Jun 4, 2014
11,505
R01 (next to the bridge rectifier) whilst it looks like a 2mohm (or less likely a 55ohm) measures 186kohm one-way, 196kohm t'other way - which might be about right when everything else is taken into account.
I make R01 (red, black, yellow) 200k so your resistance readings, in circuit, look OK.
 

Thread Starter

daveLNER

Joined Apr 25, 2020
6
R01 - must be the light I was looking at it - looked like the colours were red, black, green, (and green), i.e. 2Mohm, but take your point the greens being yellows. Makes more sense.
 

Thread Starter

daveLNER

Joined Apr 25, 2020
6
Ah well. A few checks made and none seemed to make a lot of sense, except the output.

Without any loading on the o/p I get 12.20v - tick.
On the TL431 I get 2.48v on one side and a variable 9.2>9.6v on the other - ?
On the 3842 pin 2 I have 0.38v (should be according to the datasheet 2.5v)
On pin 7 (Vin) I got 14v - tick (probably)
On pin 8 I have a variable 0.6v where as should have 5v.

That was before it went bang - blew the fuse and died! Quite what caused it I have no idea. The meter probe was nowhere near the chip (or anything else - I had all the HV side masked off).

Now there's nothing on the o/p (after temp bridging the fuse) except 333v on the bridge +.

Something shorted over, nice smell of burning but nothing to show for it except a grey smear from under the fuse.

Think I'll take Dodgydave's original advice and go buy another PSU.
Not happy, but that's life. You win some, you lose some.

Anyway, thanks guys for your inputs into this. In some ways I have learned quite a lot - once I read through all the pages and pages of stuff on SMPS. I'll know more for next time.
 

AlbertHall

Joined Jun 4, 2014
11,505
Think I'll take Dodgydave's original advice and go buy another PSU.
Not happy, but that's life. You win some, you lose some.
These things are difficult to repair even with a schematic and knowledge of previous faults repaired. I spent quite a few years fixing the things.
 
Top