Switching power supply problems

[Charlie86]

I have a problem with Microset PCS140 switching power supply.  The power supply fluctuates and a buzzer sounds.  More in the video.  Can anyone give me some advice on how to fix it?

https://youtu.be/r0XsFqVtPfg

[Refrigerator]

Do you have a schematic for it?

[Charlie86]

unfortunately I don't have it, nor do I get it from the manufacturer.  All I found on the net is the document below.  What could this buzz be?  When it hums, the voltage jumps as well.  I can't determine which component is buzzing the transformer or capacitor.


https://www.wimo.com/media/manuals/MICROSET/PCS-125_Schaltnetzteil_Power-supply_IT.pdf

[Riotpack]

Have you had a look at the underside of the PCB for dry joints, broken traces etc.

[Charlie86]

Hi,

Just resolder all joints. Didn’t fix the problem. Any other idea ?

[pbarton]

A squirt from your favorite aerosol can, into that voltage control pot, that you are twiddling?

[Charlie86]

i cleaned it up and it works ok.  from 1 ohm to 1M ohm

[Refrigerator]

That clicking would be the switching circuitry turning on and off sharply, imo.
I think there might be a problem with your feedback circuit, maybe the optocoupler has gone bad.
Hard to tell without any schematic, because then it would be more clear how the regulation works and if there's a standby power supply etc.
If there's a standby power supply then it could be that it is overloaded and is power cycling, i say that because the voltage potentiometer does seem to have an effect on the output still.
You should find if there's a standby power supply and if there is one you should measure it to see if it's stable. Poking around in these PSU's is dangerous, don't get zapped.
A detailed picture of the board would help us a lot.

i cleaned it up and it works ok.  from 1 ohm to 1M ohm

1MOhm? What does the writing on the potentiometer say?

[Charlie86]

first of all I would like to thank you for your help.  I was wrong with the potentiometer, it is adjustable from 1ohm to 1k ohm.  I am also attaching some additional pictures.  Please write if you need a more detailed picture of a particular part.

[pbarton]

The date code on the top of the isolating transformer appears to be “HI-POT 3808”. So, I’m assuming that the power supply is 12 years old. One or two of the capacitors may be past their ‘use by’ date.
The frequency of oscillation appears to be much lower than your supply frequency (50/60Hz) almost like ‘motor-boating’. Which is often indicative of a capacitor problem.
There appear to be many capacitors inside your power-supply and its not a simple task to check them individually for capacitance and ESR or, to replace them all.
Neither is it easy to diagnose a failing capacitor when the power-supply is in its failing operational mode, as that oscillating voltage will probably appear in several places inside the power-supply circuit (including the output).
A schematic would help. You could try adding another capacitor in parallel with an existing capacitor (say 50% of the value of the existing capacitor) for each capacitor in turn, to see if that identifies or resolves the problem. You may find that easier than removing each capacitor to measure it and then have to re-install it back in circuit. However, this is still a lengthy procedure.

Just looked at the photos again. I see 4 surface mount capacitors on the vertical daughter board and 7 radial electrolytic's?

[Technomaniac]

If that pott has DC on it, better to NOT spray it with anything that stays behind.  It will cause minute arcing between wiper and track  and microdamage to one or the other.  OK to use a  quickly evaporating one and wait for it to dry before switchon.

[Charlie86]

That's right, the power supply is 11 years old.  I replaced all electrolytic capacitors on the main board with new ones, except for the two largest 680 micro 300v, which I didn't get in the local store.  I will order them online.  Looking at the daughter board, I am attaching a more detailed picture. One more thing: the power supply is adjustable from 5-15v, when the potentiometer is at minimum the voltage is 4v, when I turn it towards maxiumum the voltage is up to approx.  10v stable, from 10v onwards it starts jumping.  Sometimes it also happens that the overload light comes on and the power supply shuts down by itself.  I also inspected it with a flir camera and noticed that some resistors have 60 degrees Celsius, and the IC chn lm358n kwg838 has aprox.  50 degrees Celsius.  I also notice one suspicious transistor next to the chn lm358n kwg838.  I am attaching a picture.

[pbarton]

Oooh, you have already replaced the main board electrolytic's. I doubt that those large 400V rated caps on the HOT side (which have not been replaced) are the problem. You are further along with this investigation than I imagined. Ooooh you have more test equipment than I imagined. Those surface mount capacitors on the daughterboard have a reputation to weep their clear electrolyte onto the PCB. However, I expect that you have inspected the daughter board for visual contamination of the clear electrolyte. How to proceed? I would identify the control chip on the daughter board and look up the datasheets and schematics to find a typical application circuit for that chip. Your implementation will be very similar to others that use the same chip.
Edit: Can you check the ESR of the daughter board capacitors without removing them?

[Refrigerator]

Both of your primary side filter capacitors are mismatched, they are in series and should both be the same value so maintain balance.
Also looks like the gray block with the zip tie going over it is your standby transformer.
This PSU seems to be set up similarly to an ATX PSU. The daughter board has an isolation transformer to drive the gates (or bases) of the switching transistors.
But it looks different to your usual TL494, because i only see two wires coming out of the GDT on the secondary side.
And what's weird to me is that i see no signal shaping circuitry on the gate drive. What's the part number on those switching transistors?
And you might want to test them while you're in there.

[pbarton]

I ran the video again and I note that those two large high energy capacitors on the HOT side are very different to the ones shown in the photos. Probably nothing to do with the fault though.
I also noted during the video, at time marker 00:34 that the meter jumps and the “overload” lamp flickers at the same time. With no PSU output load applied, the overload detection circuit appears to have undue influence on the PWM controller.

[Charlie86]

hi, yes the capacitor on the HOT side I replaced because with different values ​​because I didn't have the right ones.  I put the old ones back now.  Unfortunately, I don't have an ESR meter, but I removed daughter board and noticed that someone befoure me was already repairing it.  It can be seen that someone has replaced the SMD resistors with PTH. Next to these two resistors, there are also two diodes.  Both let the voltage in both directions.  In one direction 0.5v in the other 0.65v.  Is that OK?  As you've probably already noticed, I'm not an electrician ,One more thing I'm going to buy SMD electrolytes tomorrow, but I don't know the value from the labels.  Can anyone write me the values ​​I need to buy.

[pbarton]

99% of SMPS use an opto-isolator to provide feedback from the output back to the PWM controller. Yours appears to be the exception, as I can’t see any opto-isolator anywhere. I can only assume that the small transformer on the daughter board is performing this function.
The RHS of the daughter board (which has 8 pins) is low voltage and provides the variable output voltage, the output voltage regulation and the overcurrent detection. The LHS of the daughter board (which has 4 pins) is high voltage and is a 2 transistor PWM controller (the main switching transistor is on the main board next to those big caps). The two ends are separated by a safety barrier. Only one component is bridging this gap, the small feedback transformer.
The previous repair is in the PWM controller section. Unfortunately we are not in a good position. 1. We don’t have the schematics and 2. Components have been changed (we can’t be certain as to their original values).
Having seen the prior repair I feel less confident that replacing those surface mount capacitors is going to fix the problem (3x 10uF at 35V and 1x 1uF at 50V). I suspect there is another problem in that two transistor PWM circuit.
Option 1, may be to contact Microset (http://www.microset.net/categorie.php?mid=89&catid=89&lang=en) to see if they can supply a daughter board (or schematics). But as its 12 years old it may no longer be supported?
Option 2 is to check and replace components in the PWM controller on the daughter board. You mention measuring voltages across the diodes. Normally we measure resistance across the diodes, so I’m a bit confused.
I’m having difficulty finding a typical schematic for a SMPS with transformer coupled feedback.

[Refrigerator]

the small feedback transformer.

It's a GDT and in no way has anything to do with feedback.

The previous repair is in the PWM controller section. Unfortunately we are not in a good position. 1. We don’t have the schematics and 2. Components have been changed (we can’t be certain as to their original values).

That's the signal shaping circuit i mentioned. It wasn't apparent in the pictures because it was covered with the GDT.

You mention measuring voltages across the diodes. Normally we measure resistance across the diodes, so I’m a bit confused.

Why would the forward voltage of a diode be measured in Ohms? What's the point in that?

Both let the voltage in both directions.  In one direction 0.5v in the other 0.65v.  Is that OK?

This means that your switching transistors are likely BJT's and not MOSFET's and the second direction in the "diode" is the BE junction on the BJT. In that case me calling the small transformer a GDT would be incorrect, because it means Gate (not base) Drive Transformer, but i'll keep calling it that for the sake of continuity.
Anyways, OP, you can count yourself lucky because i see a TL494 and just by knowing that we no longer need the schematic because the rest can be tested or traced back with just a DMM, imo.
TL494 is widely used in ATX PSU's and there are many schematics that can be used for reference. This IC controls the switching transistor from the secondary side by using the GDT to achieve galvanic isolation and therefore needs no optocoupler because it senses the voltage directly.
If what @pbarton mentioned about the overload LED flashing is true then it must mean that your PSU is being falsely overloaded.
I see that the gauge on the front panel also shows current meaning that there has to be a way for the PSU to measure the current. Next step from here would be to find the current shunt and the error amplifier connected to it to test if it's not acting up.
From the pictures i saw two LM358 op-amps, and i think one is for displaying the current on the gauge. By the way what does the gauge show when switched to display the current?

[pbarton]

The three transformer design and the lack of an opto-isolator does indicate an ATX Power Supply architecture. (Main transformer, PWM Control Isolator and Standby transformer).
A good article/tutorial about ATX Power Supplies can be found here http://www.ntrains.org/atx-power-supplies-and-model-railroads/
Regarding the OP’s comment “In one direction 0.5v in the other 0.65v.” Voltages are observed in normal operation, however when the daughter board is removed from the main board, I would have expected to OP to measure the diodes with a multimer, on the Ohms range. Hence the confusion.

[Charlie86]

Hi, I contacted Microset before I wrote on the forum, but I didn't get a response.

About capacitors, I will try to replace them anyway because it is a small cost.

Regarding the overload: yes, it happens that the LED for the overload flashes or even lights up constantly (in this case the voltage drops to 0v if I turn off the power supply for 1 sec. And turn it on again it works again normaly.)

Regarding the measurement of diodes, I apologize, as I said, I'm not an electrician

About LM358; marked in the picture below has a constant approx. 50 degrees Celsius.

Next to it is also the C32725 W 62 transistor, which looked “suspicious,” but I removed it from the circuit and it’s OK.

Regarding the current shunt, I assume this is located at the output connectors?
I google but I can't find how to check if it is working correctly.

I will check Amp meter today after I came from work.

[Refrigerator]

50 degrees Celsius.

Not good, i think this might be your problem. I can't even think of any way to make an LM358 heat up that much.

[Charlie86]

Could the reason be a poor current shunt for heating the LM358 or a bad LM358 itself?

[pbarton]

The LM358 is a low power dual operational amplifier and should never reach 50 degrees Celsius. So we suspect that the LM358 is bad and it should be replaced.

[Refrigerator]

The LM358 being hot may not be the root cause of the fault, ultimately something had to make the LM358 fail, but they're so cheap that replacing it to see if that fixes the problem is worth the try.

[Charlie86]

Hi,

I replaced the LM358 and SMD capacitors.
Apparently, I somehow further damaged the power supply
Now, as soon as I plug it into the mains, it blows the fuse. Something may be in contact, I removed the LM358 as well as the daughter board. I also removed the two largest capacitors, but there is still something in contact. I checked the solder joints for shorts twice or even three times.
Could a transformer that used to emit buzzing sound break down?
Maybe some advice on how to figure out what’s in short circuit?