Home made benchtop power supply

This article also available in Portugese

Any electronics tinkerer has a few benchtop power supplies. Growing up as a teenager in the 1980s in Northern Ontario, I didn't have money to buy a proper benchtop power supply. So, as usual, I built my own. I first built this one around 1985, but then re-built it with a nicer case and a volt meter in 1987. it had to be fairly small because I wanted to take it with me when I went to university.

The enclosure is out made of wood of course, because that's what I had handy. It's a nice box made of oak. Oak is sturdy and less likely to catch fire than lighter woods would be. The top just slides over the front and back panels, with a groove to hold them in place. That way, with the cover off, I can get at the internals more easily.

For the front panel, I stuck laminated paper in front of a piece of plywood, with the switches and such mounted through the laminated paper and plywood. This looks really good, but I didn't expect the laminated paper to last all that long. That was in 1987, and it still looks good today, 28 years later. But that's partly because it's been in relatively unlit areas most of its life. The front panel lettering was done in pencil. My Commodore 64 computers and homemade plotter wouldn't have produced an output as nice looking, and would have been much more work.

If you use this sort of technique today, I recommend using photo paper on an ink jet printer. I made a new scale for a meter movement that way once. Took a few tries, but it came out really nice.

It's an unregulated supply, with no semiconductors other than the bridge rectifier.

At the heart is a filament transformer from an old tube tester from 1951. The heating filaments of tubes in the cheaper transformer-less TVs were all wired in series, so they used to come with all kinds of filament voltages. The tube tester had a many-tapped transformer for selecting a wide range of voltages.

The primary also had two taps on it. There used to be a rheostat between them, so that the voltage of the unit could be tweaked by what the line voltage was (the tube tester itself was also unregulated). I added a switch to select between the two taps, doubling the 17 voltage steps to 34. Switching taps changes the output by 12%. I couldn't think of what to label the switch, so I called it 'Boost', because the voltages on the front are for it in the low position. People have often made fun of that label!

I added a meter movement with ranges for 15 and 30 volts DC, and a push button for ampere reading. The meter doesn't need to go past 30 volts because I wired it up so that past 30 volts, the DC side of the supply cuts out. The switch conveniently had a tap for this. The filter capacitors and rectifier were only rated for 35 volts. However, AC comes straight out (above the DC terminals) and will go all the way up to 140 volts.

The nice thing about using this supply is turning the knob. It's a very satisfying click switching between voltages. And voltages can be adjusted very quickly, if not precisely. But I can usually get close enough for what I need. Very different from tweaking a multi-turned knob and watching a digital panel meter to hit the right voltage. Granted, I don't use this one when I need a precise voltage, like for for charging lithium ion cells.

The AC output often comes in handy - like when checking out small AC motors, or for powering a degaussing coil. One time, I even brought it in to work to blow out some shorts we had in some prototype printed circuit boards. Without any electronics between the transformer and the terminals, it will put out over 20 amperes for short durations.

I also put two outlets on the back of it, connected to the power cord. I always found myself short of outlets when playing with electronics, so two extra receptacles are always handy.

I re-drew a schematic for it from looking at the circuitry. The beauty of this power supply is that it's so simple that I can understand everything in it. Even drills and flashlights contain fancy electronics these days, but not this power supply.

I also learned an interesting thing about the meter movement when I built it. For the 3-amp current range, I used a steel wire as a shunt resistor, which I tapped at the exact length I needed to get the movement to read full scale at three amperes. Full scale deflection requires about 57 millivolts across the meter terminals. The problem with connecting the meter directly to the shunt resistor is that the movement of the meter's coil through its magnetic field induces quite a bit of back-EMF as the needle moves. Without a few k-ohms of resistance in series with the meter movement, the back back-EMF from the needle makes it move like it was under water. The meter settles at the correct value eventually, but it takes about two seconds to get there.

Anyways, this power supply is still my favourite benchtop supply and still the one I use by default.

Repairing it, adding a breaker

A "feature" of this power supply is no overcurrent protection. This is quite useful at times. But one time I had the power leads touching each other, turned it on, and turned my attention to something else. A minute later, I realized my mistake, and the power supply needed a simple repair. I also added a breaker.