Simple Siwtch Mode Power Supple Repair tips (single UC3843 chip)

Hi

Im new to repairing switch mode power supplies - so excuse my terminology.

I have two SMPS, one looks to be fairly complex and before i try to tackle that one i'd thought i'd start on the easier one.
The story is - it came out of a external USB drive. The SMPS supply looks pretty simple. (because there's only 1 switching transistor)

The modem says Asian Power Devices Model APD-5030-70A2.

It has only one switching transistor and a UC3843 PWM chip. The outputs at 5 & 12v.
The fuse is in tack. I can measure ~334V across the large cap so i know all the initial AC to DC circuitry is working.

However, there's no power to the PWM. My datasheet says 5v should be on pin 8.(vref)

In fact i cant measure any power on the secondary side. ie; PWM is not turned on.

The switcher(transistor) doesnt seem to be firing up and thats about how far i've got. Visually checking the caps all look good. The main 150uF 400V cap works. All the diodes are working. Even pulled out a couple low value resisters that looked to be short only to find that they were 0.25ohm resisters below my multimeter could detect. The solder joints look fine.

I've read a couple posts that suggest to power up the PWM chip independently or even send an external PWM signal to the base of the switch and stand back . (after placing a light bulb in series with the mains power)

I guess im looking for tips or gotcha. I dont readily have an external spare PWM signal ready to inject into the switcher/transister. Is that what is normally done? Also how do you tell if my smps requires a load to start up? Im also lef to believe its normally the 5v rail that is used to sense the voltage?

Any insight is appreciated.

Thanks in advanced.

If it needs a load to start up it would have to provide at least some power to detect whether a load was there or not! PSUs that need a load will generally start and stop cyclically, making a distinctive ticking noise.

In my experience, most SMPS that have primary supply but no output are lacking the 'kick start' power. I can't tell you which components provide it without seeing a schematic but most SMPS provide the working voltage for the UC3843 from the transformer itself, of course, when you first switch on there is no power to the regulator from the transformer so it can't work. They use an alternative 'weak' power source to get the circuit working, then the main source from the transformer takes over. The usual problem is the 'weak' power source is missing so it never has chance to start running.

Some SMPS use a capacitor to provide the initial power, it would be connected between the high voltage line and the regulator so it's charge current flows through the regulator and lifts it's supply up for a moment. If your power supply uses that method, look for low value (typically < 10uF) electrolytic capacitor that has a high working voltage, usually > 200V rated. There are likely to be other components in series with it, most likely a resistor to limit the surge current and maybe a diode to make sure it discharges quickly when the power is turned off.

Sometimes, but rarely, there will be a high power resistor > 2W between the high voltage and the regulator. It's less common because they run constantly hot and are more expensive than a small capacitor.

I would say 90% of the time the problem is the low value capacitor.

Brian.

Thanks for the reply - couldn't see any electrolytic caps.
I've traced from the PWM pin 7 - which according to the datasheet is VCC.

From there i can see an PCB latch plug. I guess its unpluged at manufacturing but does give me "easy access" to powering the PWM externally.
On the otherside of the plug i can sort of make out a zener diode and also several resisters.

NB: I found R19 (1.5Kohm) measuring at 7Kohm - i replaced that. I can see that was around the Rx/Cx side of the PWM - so my hopes werent up when it still failed to deliver power.

Anyway i can see the PWM on the primary side. I can see 920K resister which one side gets 334v and the other looks to be diode and thats about it for now.
Im calling it a night. I think the plan is to trace back from vcc on the PWM and write out the components - ie: create my own schematic. :-(

Just reporting in -

I found a clue seems to be a short between vcc and gnd.