Switching power supply unreliable startup.

[naive1073]

I've got an octopre mkII with a switching power supply that more often than not takes its time to power up - i.e. no lights until a minute or so. Sometimes switches on immediately and once it's warmed up it can be switched off and back on without issue. When lit up it works as expected. Any clues, procedures that might help me? Even a google search tip would be welcomed ("slow start switching power" gets you loads of not applicable stuff) Any further info I can supply? Thanks in advance.

 

[Walrus]

Possibly a failing varistor (if it has one). Taking too long a time to heat up and drop in resistance.

 

[JohnRoberts]

Switching power supplies often require start-up circuits to accommodate the initial current draw to charge up large output reservoir capacitors that could current limit or trip normal protection circuits. Switching supplies usually deliver many thousands of small current pulses, so start-up can be problematic if it involves a huge current draw initially.

The last switcher design I worked on (10 years ago) had a cap on one of the controller IC pins that set a soft start-up delay time.

JR

 

[radardoug]

Switchmodes commonly have a cap across the power supply pins of the controller ic, 10 - 100 uF. This is usually fed by a high value resistor from the rectified mains. Commonly either the resistor fails high, or the cap dries out.

 

[naive1073]

Excellent, thanks all. That gives me a few start off points and makes my googling more useful than I thought. Had time today to pull  the cheap 85 degree 10uf 50v electrolytic at the IC and replaced it with a Panasonic FM 10uf 50v, now there's no power up at all. I'll investigate further tomorrow. Thanks again.

 

[Andy Peters]

Excellent, thanks all. That gives me a few start off points and makes my googling more useful than I thought. Had time today to pull  the cheap 85 degree 10uf 50v electrolytic at the IC and replaced it with a Panasonic FM 10uf 50v, now there's no power up at all. I'll investigate further tomorrow. Thanks again.

Shotgunning a part without any evidence that the part is bad is often fruitless, and sometimes makes matters worse.

Can you get the schematic? That would help.

 

[naive1073]

Thanks Andy, yes this is definitely evidence for the shortcomings of that strategy, unless the failure is a useful diagnostic symptom. No schematic has come to light for me, on first appraisal it looks like a really simple through hole single sided board but remove it and turn it over and the smd comes to town.

 

[JohnRoberts]

I have repaired a lot of circuits without having schematics.

Some times it is easier when component catch fire... but early on as a junior technician I learned how to identify failed components using a VOM. With switching power supplies I've seen some early low impedance electrolytic caps fail as a short circuit. switching devices can fail open or short, short is common too. 

Control chips are hard to troubleshoot simply.

JR

 

[s2udio]

Replace the Power Fet, feedback opto,small electros and The PWM controller chip.
And check all the diodes and resistors on the switching side,....only way to be certain of a working repair on an SMPS.
Schematics unlikely !

 

[Andy Peters]

Thanks Andy, yes this is definitely evidence for the shortcomings of that strategy, unless the failure is a useful diagnostic symptom. No schematic has come to light for me, on first appraisal it looks like a really simple through hole single sided board but remove it and turn it over and the smd comes to town.

Of course it's SMD It was made in the last 30 years.

If you can find the part number of the switcher controller chip, pull its datasheet. Production designs are often close enough to example designs to give you a good starting point for probing.

Be very careful if it's a non-isolated off-the-line switcher.

-a