Measuring components on the hot side of a SMPS when turned off

[ddewaele]

I am troubleshooting a switch-mode-power-supply from an old computer that keeps blowing fuses and destroying components on it, and I am trying to see what I can safely measure and test on the hot side without the circuit being turned on (and capacitors discharged).

The main components I seeing are the following (including how I checked them).

• Fuse : checking continuity
• EMI filter (coil) : no idea
• Safety capacitors (X/Y class): check if there are no shorts, checking capacitance using multimeter in-circuit looks good
• Bridge rectifier : check if there are no shorts, diode check using multimeter in-circuit failed, out-of-circuit was ok.
• Bulk capacitors : check if there are no shorts, checking capacitance using multimeter, check ESR value using ESR meter both in-circuit and out-of-circuit was ok
• Varistor / surge absorber : no idea, reads as open circuit in-circuit using multimeter
• Thermistor  / inrush current limiter : in-circuit reads about 15 ohms at room temperature, changes resistance when temperature increases / decreases.
• Resistors : checked if there are no shorts, measured them and checked color codes to see if they correspond
• Diodes : check if there are no shorts, diode check using multimeter in-circuit and out-of-circuit
• Capacitors : check if there are no shorts, checking capacitance using multimeter, check ESR value using ESR meter both in-circuit and out-of-circuit
• Power transformer : check if there are no shorts, diode check on base / emitter / collector out-of-circuit
• Opto-coupler : can only test the LED using diode test, no base exposed for transistor. removed if out-of-circuit and built a small test circuit
• Transformer : no idea how to test it. Always read this is unlikely to fail but you never know.

There is no IC to do the switching on the power supply so I assume that this is done by a combination of discrete parts (that I haven't figured out yet how this is done).

I did carefully check the voltage of 1 bulk capacitor and it read 165v, I am assuming this is 230V AC times sqr 2= 320 V DC divided by 2 capacitors = 160V

Is there anything else I can check on the input side, and should I get over my fear of poking around with a multi-meter on the hot side ? Been thinking about using electrical tape to cover parts of the circuit I don't need at that time, and being super careful when sticking my probes not, making sure I never touch both leads while I am checking voltages.

[TimNJ]

Is this an ATX style power supply? There is really no controller? Does it have a daughter/baby board, mounted 90 degrees to the main board?

You say the unit keeps blowing fuses, but also measured 165V on the bulk caps. When does it blow its fuse(s)?

Do you have any banana to mini-grabber/alligator clips? If you are afraid to poke around with multimeter probes with the unit on (understandably), you can try soldering in little flying wires with the unit off (make sure to *fully* discharged bulk cap!!), attaching the mini-grabber clips to the wires, and then powering on the unit once all probes are in place. This is usually how I set up a measurement with an oscilloscope, but equally appropriate for multimeter.

Also, as long as you are on the volts jack, the input impedance of the meter is very high (usually 10 Meg-ohm), so you don't really have to worry about blowing up your meter. The only thing to worry about is a probe slip, where your hand slips a bit and you wind up shorting out something across the probe tip.

[ambrosia heart]

HOW TO TEST :
use a stepdown transformer  to input 36V a.c. Into the EMI circuit   and test if 36V a.c. comes out of the EMI filter circuit,
including EMI coil and X/Y cap.  If yes, they are OK.
    Fuse : checking continuity
    EMI filter (coil) : no idea
    Safety capacitors (X/Y class): check if there are no shorts, checking capacitance using multimeter in-circuit looks good

In real situation,  after fuse blowed, bridge rectifier will be destroyed and smoothing cap. capacitance must be checked to see if it is  dry.  If yes, will make fuse blowed. In half bridge SMPS, however, fuse and bridge rectifier  are ok even one of  smoothing caps dry out.
    Bridge rectifier : check if there are no shorts, diode check using multimeter in-circuit failed, out-of-circuit was ok.
    Bulk capacitors : check if there are no shorts, checking capacitance using multimeter, check ESR value using ESR meter both in-circuit and out-of-circuit was ok

    Varistor / surge absorber : no idea, reads as open circuit in-circuit using multimeter. 
It must be checked, located at and connected to the primary  side of the SMPS transformer . It absorbs back emf
about - 800V, when SMPS power transistor is off, due to -Ldi/dt.
   
Thermistor  / inrush current limiter : in-circuit reads about 15 ohms at room temperature, changes resistance when temperature increases / decreases.  Must  be checked
   
 Resistors : checked if there are no shorts, measured them and checked color codes to see if they correspond.  The starting resistor connected from the large cap to the baase of power transistor must be checked.
   
Diodes : check if there are no shorts, diode check using multimeter in-circuit and out-of-circuit
    Capacitors : check if there are no shorts, checking capacitance using multimeter, check ESR value using ESR meter both in-circuit and out-of-circuit.

Those SMPS transformers almost NEVER die.  For Beginners, no need to check them.  For  computer power supply, there are 2 to 3 transformers called voltage output transformer, standby transformer and drive transformer.  For opto-coupler, no need to check.  Its faulty will not make fuse get blowed.

    Power transformer : check if there are no shorts, diode check on base / emitter / collector out-of-circuit
    Opto-coupler : can only test the LED using diode test, no base exposed for transistor. removed if out-of-circuit and built a small test circuit
    Transformer : no idea how to test it. Always read this is unlikely to fail but you never know.

[ddewaele]

Is this an ATX style power supply? There is really no controller? Does it have a daughter/baby board, mounted 90 degrees to the main board?

It is an AT power supply from an old XT computer from 1988/1989 I think. Have some pictures here : https://drive.google.com/drive/u/0/folders/15Jsd2VW87n1EpAHDDRfVNCt2t73Vr7oZ

You say the unit keeps blowing fuses, but also measured 165V on the bulk caps. When does it blow its fuse(s)?

Well it has done it twice but haven't turned it on since ... re-thinking my strategy and going to get myself a load-lamp setup and isolation transformer first. (if I understand correctly that should protect not only myself but also the circuit)

The first time it happened was while the computer was running fine for several hours.  All of a sudden the computer shut down by itself (weird), and about 5sec after that the fuse blew. (didn't have time to turn off the computer... happened pretty fast and was confused as to why the computer shut down).

I did some diagnostics and noticed a shorted power transistor, schottky diode, a shorted electrolytic cap, a blown resistor and a blown transistor.

I replaced all of them (thinking I checked everything) and when I was starting the PSU I measured the 12V dc output rail with my multimeter.
It went to about 10V and then started collapsing and while at 8V or 9V  the fuse violently blew (< 5 seconds after I turned on the PSU). I also had my IR thermometer and noticed the power transistor reaching 60 degrees celcius in a matter of seconds.

This time also the bridge rectifier + all of the replaced components above were shorted again.

Do you have any banana to mini-grabber/alligator clips? If you are afraid to poke around with multimeter probes with the unit on (understandably), you can try soldering in little flying wires with the unit off (make sure to *fully* discharged bulk cap!!), attaching the mini-grabber clips to the wires, and then powering on the unit once all probes are in place. This is usually how I set up a measurement with an oscilloscope, but equally appropriate for multimeter.

Yes, was also thinking about soldering wires onto them and hooking my multi-meter so I can safely turn it on with the equipment attached not having to worry about slipping probes.

Also, as long as you are on the volts jack, the input impedance of the meter is very high (usually 10 Meg-ohm), so you don't really have to worry about blowing up your meter. The only thing to worry about is a probe slip, where your hand slips a bit and you wind up shorting out something across the probe tip.

That's what I was thinking, but even hearing the fuse blow violently from a distance was quite a "shock" (figuratively)... wouldn't want that happening with my nose on the PCB probing it.

Thanks a lot for the reply.

[ddewaele]

HOW TO TEST :
use a stepdown transformer  to input 36V a.c. Into the EMI circuit   and test if 36V a.c. comes out of the EMI filter circuit,
including EMI coil and X/Y cap.  If yes, they are OK.

I don't have a stepdown transformer unfortunately.

In real situation,  after fuse blowed, bridge rectifier will be destroyed and smoothing cap. capacitance must be checked to see if it is  dry.  If yes, will make fuse blowed. In half bridge SMPS, however, fuse and bridge rectifier  are ok even one of  smoothing caps dry out.

First time it blew the fuse the bridge rectifier survived. I checked it out of circuit (checked the 4 diodes in the package) and DC voltage was going into the bulk caps.
The second time it blew it took the bridge rectifier with it.

    Varistor / surge absorber : no idea, reads as open circuit in-circuit using multimeter. 
It must be checked, located at and connected to the primary  side of the SMPS transformer . It absorbs back emf
about - 800V, when SMPS power transistor is off, due to -Ldi/dt.

Still not sure how to check this one.

   
 

Resistors : checked if there are no shorts, measured them and checked color codes to see if they correspond.  The starting resistor connected from the large cap to the baase of power transistor must be checked.

Not really sure what the starting resistor is. I also don't see an IC generating a PWM signal for the power transistor.
have some pictures here : https://drive.google.com/drive/u/0/folders/15Jsd2VW87n1EpAHDDRfVNCt2t73Vr7oZ

   
Those SMPS transformers almost NEVER die.  For Beginners, no need to check them.  For  computer power supply, there are 2 to 3 transformers called voltage output transformer, standby transformer and drive transformer.  For opto-coupler, no need to check.  Its faulty will not make fuse get blowed.

Someone told me as the opto-coupler is part of the feedback circuit, if it providing wrong / invalid / no feedback the transistor might be on all the time (instead of switching) leading to it (or something else) shorting and causing the fuse to blow.

Thx for taking the time to reply.