Power supply failing to start up

Question

I'm trying to diagnose why a power supply isn't working. It tries to start up, and then fails, and repeats.

The power supply is based around the HFC0100 switching controller. It's the power supply to an Arbitrary Function/Signal Generator (A Tenma 72-14110, which is a re-badged Uni-T UTG1005A). These power supplies are notorious for failing, but no one seems to know why.

The basic circuit in question, from the chip datasheet, is:

The chip is initially powered by the HV pin. When the voltage on VCC reaches about 10.5V it switches over to being powered by those pins instead. As you can see that comes from the secondary auxiliary winding. All pretty normal, really.

What actually happens is, the voltage on VCC rises to 10.5V (or sometimes very slightly higher) over the period of a second or two. Then it immediately drops down to about 5-6v. It then repeats, over and over again.

I have, so far:

• Replaced the HFC0100 - no change
• Replaced the diode in the auxiliary power signal - this was originally dead (short) and I replaced it before the chip
• Tested the drive MOSFET (IRFBC40) and with my small crude tester it claims to be fine.

I've not yet cracked out the 'scope to take a look at any of the PWM signals, but will be doing soon.

I'm really hoping someone will be able to say "That behaviour is indicative of ..." and point my probes in the right direction...

Answer 1

In common flyback supplies, the electrolytic capacitor on the VCC dries out and you get the fault you describe. This is usually 22 to 100uF and surprisingly rarely a low ESR type.

The main input electro can fail, but with flyback style supplies this can cause the integrated control/switch chip to explode.

With most switchedmode supplies, the electrolytic caps are prone to failure due to the high ripple currents they experience. For a given supply, it will only last a given amount of hours. Better quality power supplies have more conservatively rated parts and last longer.

It could be there is an overload causing the psu to shut down and go into ‘hiccup’ mode. It is not uncommon that the secondary diode(s) go short circuit. Measure the voltage on the output with a multimeter. If it rises then falls, the diode is most likely ok. If there is only a single output voltage, you can use a known good power supply to power the rest of the electronics to rule out an overcurrent situation.

I’ve had things like a leaky diode in the snubber circuit causing the hiccup mode. Again, this is on flyback style circuits. High voltage film/ceramic caps have also been known to die - the result is usually the switch device exploding. Your circuit is a resonant type and doesn’t look to have a snubber, so this doesn’t apply.

As Spehro mentions, high value resistors can fail due to the voltage across them causing electromigration. Usually this causes the circuit to not power up.

I’d be putting my money on the electros. For the cost involved, just replace all of them. If you have an ESR meter you can identify the ones that have failed or gone out of spec to avoid the ‘shotgun’ solution.

Answer 2

Perhaps a long shot given the symptoms, but check the value of the resistor to HV. Resistors in that position have a tendency to increase in value under the stress they experience, causing unreliable starting.