Today we take a look inside an old power supply made by a Polish company from Gliwice. I recovered the whole thing from scrap metal, so I'm going to check how this power supply has coped with years of use in an industrial device, examine whether it's still in working order at all, and finally show what heats up the most in it. The power supply shown here is the PWS-100RM model and is mounted on a TS-35 rail. It comes in different versions differing in output voltage, these include 5, 12, 24 and 48 V, although the documentation shows that the manufacturer is also happy to make a version with a different voltage.
The whole thing is designed to work as a stand-alone module. The label on my copy seems to suggest that this is piece number 4153 from 2001, so potentially this equipment is a good 25 years old! That could be right, as the manufacturer's company, as far as I can see, started in 1991.
Externally the whole thing is quite modest, one of the terminals (terminals) is broken off in my copy. Apart from that, there is only an LED indicating the presence of power. The hardware is fully passively cooled, which in an industrial environment, however, can be a plus, as it eliminates the problem of fans seizing and clogging up over time.
We remove the casing. You can immediately see that this is no budget Chinese power supply. No one has spared the copper here, the EMI filters seem to be fully present. You can also see significant dirt and traces of moisture.
It's been a long time since I've seen a Polish "caution" sign on a 400 V mains voltage filtering capacitor.
The board is single-sided, simple but fulfilling. The thicknesses of the current paths and the appropriate insulation spacing instantly inspire confidence. There is a manufacturer's logo on the description layer.
I like the fact that the PE is connected to the chassis by a connector rather than soldered. It looks like the whole thing has been thought out for servicing. In cheaper power supplies they would just solder it on.
There is, of course, adequate protection on the input - a T2A/250V fuse and varistors to protect against surges from the mains:
The power supply is based on the immortal UC3844 chip and is realised in flyback topology.
On the primary side, the RT, or thermistor resistor (NTC) limiting the inrush current, is conspicuous; it is soldered after the rectifier bridge, before the primary winding.
Feedback is provided by CNY75GA optocouplers, and I also see a TL431 (reference voltage) circuit there:
On the secondary side we have a small choke, a Schottky diode in a TO-220 housing and as many as four electrolytic capacitors. A potentiometer is used to adjust the output voltage over a small range.
It remains to determine the keying transistor - it appears to be an ST Microelectronics H8N80 N-MOSFET (800 V, 8 A).
Out of curiosity, I checked the most fail-safe component of such power supplies - the capacitors. They are not swollen, but perhaps they have lost their performance? I used my ESR70 for this.
All the capacitors are fine, they can be soldered in place. By the way - that fix, the fine decoupling capacitor, is not from me, it was like that from the factory.
Time then for a more serious test. The power supply is scrap, I didn't expect it to be in working order, and I wasn't prepared to test it, so I loaded it with an exorbitant 5.44 A and saw what would happen. The no-load voltage is 24 V, with four 1 Ω power resistors each connected it drops to about 21.75 V.
The most heated ... schottky diode and the thermistor (NTC) resistor. The keying transistor itself heats up slightly less.
The thermal imaging camera images above clearly show the temperature distribution after load. The main transistor heats up significantly less:
Finally, it is still worth taking a look at the power supply parameters from the original manufacturer's documentation:
It is well apparent that this is no ordinary no-name power supply from China. The manufacturer specifies its parameters in detail, including ripple levels, radio interference and the electrical insulation strength of the mains-to-housing, mains-to-output and output-to-housing.
In summary, it appears that the power supply, despite its poor visual condition, is fully operational. A longer test would be useful, already with the appropriate heatsinks in place and at a lower current, but I do not currently have such a load at hand.
The design seems very solid and well thought-out, confirming the good reputation of older, industrial devices of Polish manufacture. If, after years of service, and probably some time spent on the scrap heap, the equipment is still working and holding its parameters, this testifies to the selection of good quality components and an adequate safety margin in the design.
My plans are to clean the board thoroughly, apply thermal conductive paste and perhaps use this power supply in one of my projects.
And have you guys had the opportunity to use this type of power supply in your factories or home workshops? How do you rate their failure rate?
Cool? Ranking DIY Helpful post? Buy me a coffee.
I once came across a switch mode power supply from a Polish TV receiver Jowisz 04 it was an even bigger monster.
