Wine cooler flickers at startup and doesn't cool enough

[edugimeno]

Hi there!
I have an old wine bottle cooler, it's a basic peltier cell system with a power supply built in with the controller board, with a simple display at the front door and 2 push buttons to select the desired temperatura.
Symptoms are it doesn't barely cool down, and makes a chirping noise at the board.
Both fans on each side of the peltier are (now, after much cleaning) running smooth and silent
When you turn it on, the front 7-segment display flickers for a variable time, usually like 1 minute but sometimes nothing or 3 minutes... Then turn on solid (but still some minor intermitent dimming) and it technically works..it allows you to select a different temperature and it shows it onscreen
But the inner temp is mild, the back radiator is not even warm and the noise at the main (and only) board is loud. There is a chirping all the time, which varies depending on the load (disconnecting fans changes it)
I found one of the 2 HV input side large caps bulky, so I replaced it and expected it to be fixed, but no. Then I swapped the other one with no result.
Then I swapped both larger caps at the lower left side, close to the outputs, same effect. None of the other capacitors (expect the first 220v) had any external sign of damage or aging.
Should I keep swapping the smaller ones too?
Also the tein radiator at the HV side (top) get too hot. I measured the fan&peltier voltage and it flikers between 2.5-3.3V, which should be 12V, The fans turn but slowly. The peltier cell won't do much with 3v...
Is this a cap problem or am I missing something else? Hopefully the noise it makes sounds familiar to someone.
https://www.youtube.com/shorts/bAWuz...ZDaYgTsP9rT3si

Thanks!!

[stj]

slow startup is caps,
given how they put some of them next to big resistors they are probably all cooked.

[edugimeno]

ok so I should change those smaller ones by the low voltaje side right?
Would that relate to the high temp.at the choppers heatsinls?
Also when the screen has already settled, the chirping sound keeps gong and still 3v present at the fans and peltier output...makes sense?

Thanks

[CapLeaker]

Well… definitely change all caps near a heatsink or any other heat source like resistors etc. These would be primary suspects and those I would go for first. Then look at anything that is heavy and can get hot like Mosfets and diodes on a heatsink, transformers etc and check them for cold solder joints.

edit: your video with that buzzing sound… sounds to me a transformer is arcing.

[stj]

use low esr caps like panasonic FR

[edugimeno]

edit: your video with that buzzing sound… sounds to me a transformer is arcing.

That could be, I thought It was a heavy switching noise but it could be arcing. I Will check in full darkness
If so whats the common source of arcing? Bad solder?

And posible solution?
Thanks!!

[CapLeaker]

That could be, I thought It was a heavy switching noise but it could be arcing. I Will check in full darkness
If so whats the common source of arcing? Bad solder?

And posible solution?
Thanks!!

Sounds to me a transformer or something is arcing on the high voltage side of the PSU. Take a wooden or plastic big cooking spoon and take a listen. Could be bad isolation between the windings for one reason or other. First you got to find out what is making that arcing noise.
It's not a ticking noise, indicating something wrong in the secondary,

[edugimeno]

Sounds to me a transformer or something is arcing on the high voltage side of the PSU. Take a wooden or plastic big cooking spoon and take a listen. Could be bad isolation between the windings for one reason or other. First you got to find out what is making that arcing noise.
It's not a ticking noise, indicating something wrong in the secondary,

Thanks! But I got a bit lost on the wooden spoon usage for this...
Could you please elaborate?.
Thanks!!!

[CapLeaker]

Thanks! But I got a bit lost on the wooden spoon usage for this...
Could you please elaborate?.
Thanks!!!

Put one end of a non conductive spoon like a wooden or plastic cooking spoon on a transformer and the other end put your ear to it. Almost like a stethoscope. Sound travels.

[edugimeno]

Put one end of a non conductive spoon like a wooden or plastic cooking spoon on a transformer and the other end put your ear to it. Almost like a stethoscope. Sound travels.

Gotcha!! I understand now. Ok yeah I will 1) try to see if I can spot the spark under no light, and 2) see if I can hear what part is having that noise with that method, thanks!

[edugimeno]

Updates... I fu**d up... Went there (I have it at my country house), checked for visible sparks at full dark. Nothing on either side
Used the wood spoon method that

CapLeaker

suggested, and found the noise coming from the larger transformed in the middle of the board.
Took it out of the chassis and started redoing some solders around the transformers area. Put it back in place and turned on. Bang...At least it was quick and felt like whatever exploded was instant and damaged no more.
Found one of the 2 (probably) chopper transistors split into 2 pieces. You can see the pictures. Turned around and found a short in two of the transformers leads I made while resoldering...
Don't you love to find that your friends wine cooler is way worse now than when he gave it to you for repair...?

Ok I will undo the short and find the transistor (labeled 13005.2), seems to be MJE13005-2, easily found in Amazon.

Will be back with news as soon as I get it
Thanks

[CapLeaker]

Well, shit happens… that's why something like a “dim bulb tester” is a must repairing switch mode power supplies. I have these at my bench at work and at home. Obviously there is or was a short somewhere in the primary on that power supply. Take that bad Mosfet out and check the 3 pads on the board for shorts. Problem now is that full power went to the gate drive circuit of the Mosfet including the PWM. There could be more damage there. Check it out.

[edugimeno]

Yes, what looks weird is that the short right now is at the SECONDARY of the transformer powered by that mosfet, not the primary. Would a short in the secondary make the Q driving the primary blow so hard? and quick? Or should I look for something else. Maybe I should try to link all this to the original issue of the clicking noise and only 2-3V delivered at output...

Thanks

[CapLeaker]

A short on the secondary? Check the ultra fast rectifier diode.

[edugimeno]

Are we talking about the guy circled in red in this picture? Testing on-.board it measures 0,5v in one direction and 1.2v in the other. His "twin" on the other transformer has near zero values on tboth ways, but the suspect side has a blown mosfet connected to it..

So you believe the short I made in the secondary should not be enough to blow the mosfet driving the primary?

Thanks!

[CapLeaker]

If you made a short on the secondary most PSU's should shut down. No, that diode isn't it. Look at your second picture, that one from the bottom. You see traces. About in the middle of the board there is about a 15mm space with nothing that separates the primary from the secondary. The rectifier diode is mounted on that big black heatsink. If that diode has 3 legs, anode on each side with a common cathode. That red square on the second picture is the secondary side of the power supply. The transformer is where the primary is separated from the secondary. Look hard on that bottom picture again, it seems there is a broken solder joint and one maybe 2 shorts on traces. Maybe redo the bottom picture, straight shot with some light, high resolution.

[edugimeno]

If you made a short on the secondary most PSU's should shut down. No, that diode isn't it. Look at your second picture, that one from the bottom. You see traces. About in the middle of the board there is about a 15mm space with nothing that separates the primary from the secondary. The rectifier diode is mounted on that big black heatsink. If that diode has 3 legs, anode on each side with a common cathode. That red square on the second picture is the secondary side of the power supply. The transformer is where the primary is separated from the secondary. Look hard on that bottom picture again, it seems there is a broken solder joint and one maybe 2 shorts on traces. Maybe redo the bottom picture, straight shot with some light, high resolution.

Thanks a lot.
The part placed on the black large heatsink is labeled STPS2045C, you are right, it's a schottky diode, I will test it later on out of the board. Also I can't really tell which bad solder point you are refering to, can you explain maybe drawing on one of my pictures?

Thanks a lot!

[CapLeaker]

… just to point out a few. In that red little box there is a solder blob that shorts to pads out. The rest either have circles or just shoddy soldering. You have to go through the whole board and clean stuff up.

[edugimeno]

Ok I see. Yes those pads are looking bad but the connection is ok, I will redo them anyway.
There is a 0Ohm across lead 1&3 of the schottky diode but I believe this is internally created in the diode...oh wait, I just checked the datasheet and it's not a single diode between lead 1-3 to lead 2, its a twin diode so leads 1&3 should not have that short... I will recheck that but someone mentioned that such a short in the secondary wouldn't make the mosfet in the primary burn immediatly...
The new mosfet is in its way but aside from removing the short in the secontary, I need to:
1) Find out what made mosfet to burn, in case it was not the short by the schottky diode
2) Find the original problem, a clicking noise in the supply voltage that would make the screen power cycle for several seconds after plugging in.

I have later on realized that this PS is PWM driven so it won't put out either 0 or 12V to the peltier and fans, I found other posts on this forum and they said it would provide a proportional voltage depending on how far it is from the set temperature, and I had set it to 12ºC, while the room temperature was 13-14ºC, so maybe the 3V I measured at the fan&peltier was not a problem, just a reduced controlled power.

Thanks

[CapLeaker]

Look at that arrow. There is a solder blob between those pins. That definitely needs to be removed. That diode package with 3 legs has 2 diodes (outer pins) and the middle pin is a common cathode. Measure each outer pin to the middle pin in both ways. One way you get an open and the other way you should see something like 0.2V of a voltage drop with your DMM set to diode mode.
Once everything looks good in the secondary and has no shorts on the output, then look at the primary.
Mosfet can burn for different reasons, by either input voltage, too much heat, gate drive circuit of the mosfet stuck on, bad capacitor(s), etc.