Switch mode power supplies

----- Original Message ----- From: "Glenn" Newsgroups: aus.electronics Sent: Thursday, August 24, 2006 8:57 PM Subject: Switch mode power supplies

Probably a Toshiba TLP621 opto isolator.

Usually when the transistor goes short something else is gone, too. Cause and effect are usually hard to separate. What else is gone depends on the topology of the supply. Is there a shottky diode that may have upgraded itself from semiconductor to full conductor, perhaps?

Cheers, Alf

Its always worth checking if there's a regulation feedback optocoupler (check for faulty or dry joints) often there's an error voltage smoothing electrolytic - usually on the primary side and also smooths the control chip's running supply (dried out or dry jointed. Shorted secondary diodes also cause the chopper Tr to blow at switch on.

Thanks for the tips. The supply is all discrete components, no IC (except for the suspected optocoupler, which is providing a feedback signal to the primary side. I've checked the two electros on the primary side, and the main smoothing capacitor, but they look OK. Also done a look around and resolder for dry joints. I'm attempting to check every component on the primary side - so far, all are showing up fine (except for the chopper).

Unfortunately, I have no circuit diagram so the job is a bit more of a challenge.

Cheers Glenn

ian field wrote:

Possibly the 2 most common causes of SMPSU failure are dry joints and electrolytics gone high ESR, its no good just measuring the capacitance as this will often read acceptable within limits on a faulty cap. Dick Smith do a good ESR meter.

"ian field" wrote

Possibly the 2 most common causes of SMPSU failure are dry joints and electrolytics gone high ESR, its no good just measuring the capacitance as this will often read acceptable within limits on a faulty cap.

***** Ian,your good advice is certainly wasted if we are to believe the OP:-

"I've checked the two electros on the primary side, and the main smoothing capacitor, but they look OK".

You understand,they must be in good working order because they LOOK OK!

Brian Goldsmith.

I was asked last Friday if I could fix a single output 24V, 60W SMPS for a Zebra-Eltron P310C plastic ID card printer. The local support company wanted a mere $AU450 for a new one shipped to my door.

There appeared to be drastic failure of a number of components (turned out that I had to replace 13 componets in all) with burnt resistors, shattered output rectifier, shorted and cracked transorb, dried out electro caps, leaky 1N4148, dead 3842 PWM chip, and of course the power FET (STW13NB60A a TO247 device). There was nothing wrong with the opto-coupler or the TL431 feedback regulator but the resistor which determines the current through the opto diode was burnt so that the value was unrecognisable. A bit of intuitive guesstimation showed that 1.8K was about right and this proved to be so under load testing.

I couldn't see the need for the original specified 13A 600V FET for a nominal 60W PSU because other similar supplies I had worked on used no more than an 8A device. Anyway RS Components only had a couple of near equivalents for the STW13NB60A at between $AU43 and $AU52 each - far too much. I substituted a IRFIB6N60A, a TO220 device rated at 5.5A

600V. Under load testing I had no trouble getting 3.2A at 24 volts (better than 72W) and the FET heatsink only got moderately warm. The substitute FET was much cheaper at around $AU8.

I also had no circuit diagram for the Zebra-Eltron PSU I mentioned in my other post. Luckily it was fairly easy to work out the schematic so I do now have a circuit.

I find the easiest way to do the shematic is to divide the schematic into 3 functional blocks and draw each out on a separate piece of paper. The first block I do, for example, is the secondary low voltage side. Nxxt I do the primary high voltage side and lastly the mains input rectifier part. Many PSU's are hard to work out because of the number of heatsinks and or the desire to achieve highest packing density and the only way to work out the schematic for these is to remove the components which prevent you from reading values etc. An arduous task but if you plan to work on the same type again the work involved is invaluable.

Very often the physical appearance of an electrolytic cap saves me the bother of switching on the ESR meter - bulged top or shrivelled plastic sleeve and out it comes, but I certainly agree not to rely on "they look OK"!!!

Thanks for all the posts and tips. I need to be more careful how I put things. "look OK" is my way of saying "tests OK", although that is a basic test since I don't have an ESR meter. Anyway, I have replaced them just in case. On more digging, I've found a couple of suspect low power transistors that look like they are in the circuit driving the base of the chopper. I'll let you know if I have success.

The set doesn't owe me anything and would otherwise be carted off to the tip. So I don't have much too lose except a few hours of mucking around.

Cheers Glenn

ian field wrote: