June 2007
This project involves the use of AC mains current, which is more than enough to kill a human being. Be safe when working with high energy sources.
This is a bench power supply with regulated DC outputs -12V, -15V, +5V, +12V, +15V, and variable and one +35VDC unregulated output. I based this bench supply almost exclusively upon Andrew Kilpatrick's design. I did add a few things, but it's essentially the same and I am not writing here to claim any shred of originality. I spent somewhere around 80$ US on the project (including the GIANT transformer), which isn't bad compared with commercial supplies boasting similar features. Please note that I DID find a few minor errors in Kilpatrick's schematic, which have been fixed here. For details, read Kilpatrick's original page.
The first part of the circuit is the rectifying/filtering board. The idea here is to rectify and smooth the positive and negative sides of the AC current from a large transformer (must be center-tapped) to produce +35V and -35V DC with a common ground. These two lines are then fed to regulators which are connected to the jacks on the front panel.
Since I had a bunch of small prototyping PCBs (cheap from Hong Kong with factory defects), I used two and split the circuit into two parts: the rectifying/filtering board and a regulating board. In the end, however, the regulators needed to dissipate more heat than I thought, so I threw out the second pcb and mounted all the regulators on a hefty Pentium II heatsink I had laying around.
The bases among the TO-220 regulators used in this project are connected to different sources, so I had to find a way to isolate them. In the end, I mounted the 78xx regulators directly to the heatsink with sheet metal screws and sandwiched the LM317 and the two 79xx's with a metal bar (which I also mounted with screws) to the heatsink, isolating them with some heatsink pads I found in the power supply of a 1989 IBM PS/2 Model 70.
Anyway, here are some pictures:
Rectifying/filtering board (left), Pentium II heatsink with attached regulators (right), case with internals exposed (top). Everything just barely fits.
The case design includes 14 banana/screw terminals, 3 DPDT switches, a 10K potentiometer, 3 LEDs, 1 male bananna plug (for the voltmeter), 2 LED panel displays purchased from Hong Kong, and 14 screw/female bannana terminals for the outputs and ground. The diagram below shows how the displays are wired:
The output from the unregulated +36V line, right off the caps. It's a perfect flat line according to the oscilloscope :)
Almost done. The voltage panel is the top one. I haven't yet mounted the LED "faceplates" or potentiometer in this picture. Note that the ammeter is actually displaying milliamps -- the decimal point should not be there. I need to fix that sometime... Notice the fuse mounted to the rear panel. For safety, a 2A slow-blow fuse is in series with the primary coil of the big mother transformer (although there was that one accidental short that blackened my screwdriver which the fuse didn't catch... perhaps a fast-blow fuse would be more preferable?)
All right! It works! Here it is shown powering the Laptop LCD to VGA project (a work in progress) .
You can see that the supply holds up pretty well, coming close to 600 - 700 mA or so in the above photo, accounting for the current consumed by the LED displays themselves (again, ignore the decimal point on the ammeter -- it's just showing milliamps.) I am very pleased with the result. One thing to note, however: 35V down to 5V is a very large drop for the 7805, and thus even when only pulling a few hundred milliamps on the +5V line the heatsink gets quite toasty. For larger currents I switch the LED displays off to reduce current consumption to a minimum. If I were to build it again, I would use a second smaller transformer for the +5V line, so the voltage difference is not so great. Other than that, it's wonderful and I hope to use it for years to come. And that's it!
Questions/Comments may be sent to adwiens AT gmail DOT com.
