Dual POS-NEG Power Supply

This is a great little power supply for breadboarding Op-Amp circuits. Here are the features:

• 1.5 Amps a Side
• Independent Voltage Adjustments
• Compact
• Easy Construction

This is a beginner electronics project so I've decided to add a lot of basic material that will hopefully assist new builders to complete this project. A lot of knowledge that many advanced users take for granted is often left unsaid. I am going to try not to do that here. If you're experienced then all you're going to need is the first couple pages of this article. If you've any questions keep reading and maybe something else I've included here will help you figuring out whatever it is that you do not understand.

Parts the Regulator Board requires

R1, R2 220 (2)
R3, R4 2K 10 turn trimmer (2)

C1, C2 .1uF 50V (2)
C3, C4 10uF 50V (2)
C5, C6 1uF 50V (2)

D1-D4 1N400X (4) (X=2 to 7)
IC1 LM337
IC2 LM317

Heatsinks for regulators
Stranded hook up wire 3 or 6 colors
2" X 4" (51x102mm) copper clad board

Suggested Parts for Unregulated DC Circuit

• AC Power Cord
• SPST Switch
• Fuse & Holder
• Center Tapped Step Down Transformer
• Diodes or a Diode Bridge
• Pair of Filter Capacitors 1,500-5,000uF
• Box, Chassis, Other Mounting Method

I am including a schematic of what I am supplying my regulator with as an example circuit. Many variations are possible. I happen to be using 6800uF capacitors for C1 and C2 but anything from 1500uF on up would suffice. The transformer I am using is a 24 Volt center tapped one. Here again you may choose a transformer with a voltage output more suitable for your intended use. I will discuss this part of the project in more detail in step 4. So if you feel you need some more information to build something for yourself you can find some there.

The second schematic is what I consider to be the project this article is about, although without something doing what this first section does the regulator board is not useful.

This circuit is so simple it may not even be worth etching. It is certainly simple enough to be a candidate for the print it onto plain paper, paste that onto plain board, and drill build method.

When I laid my board out I designed it with using specific heatsinks in mind. You can rotate the LM317voltage regulator and get a cleaner trace pattern on your board. How I laid it out isn't completely horrible and it does work. One thing I'd have liked to have done was to have made my copper lands not as blocky as they came out but I do not know how to do that myself using the software that I have.

You can always get creative with a Sharpie marker I suppose. Etching less copper keeps etching solutions from loading up. Or skip etching a board entire like I suggested. The choice is yours.

What I am calling the input stage of this project is everything up to the circuit board. That would be the power cord, power switch, fuse, transformer, bridge rectifier, and the filter capacitors. Everything depicted in the first schematic on page 1 of this article which I shall include again here to avoid confusion.

Let us consider each component in the power chain in turn beginning with the power cord. Today the choice of a power cord for a project is simple. My project box of choice comes with one built in, a defunct PC power supply. Gut it and you are left with a metal enclosure of handy dimensions and an IEC connector that fits the ubiquitous power cord that I've piles of.

Some IEC connectors even have built in fuse holders which satisfies our next requirement. Whatever you do use some kind of a fuse with your plugged in project. How you determine the correct value fuse to use is a combination of factors based mainly on what you have the most of in your fuse hoard. No not really, but use some common sense selecting an appropriate type and size please.

Junk power supplies even often contain switches, although I find the ones that typically come on PC PSUs difficult to use so I usually don't bother with them. We all collect switches don't we? I know I've my switch stash complete with those "special" switches waiting for just the right project to use them on. This project only requires a Single Pole Single Throw Switch. Other types are adaptable though.

This project needs voltages both above and below ground so a center tapped transformer best satisfies the circuit's requirements. These are not the most common transformer style available. One place you may be able to scare one up is in defunct audio equipment. The general rule here is the higher the wattage output the higher the voltage transformer is in it. So for this project you're going to be looking for something around the 10 watt or less to get voltages in a useful range. Maybe an old boom box that plugged in or a really lousy weak old home stereo has a good transformer in it?

One obvious idea that has struck me scavenging for components to pursue my electronics hobby is the fact that I am best off if I look into devices similar to what I am constructing when I am searching for parts along those lines. Like duh! Though I'll admit this thought didn't dawn on me as quickly as it should have. I bring it up now in case it hasn't quite crystallized for others yet either.

Getting back to our project if you luckily find yourself the proud new owner of a rotten old stereo that contains donor parts for this project hooray! You may find filter capacitors that are past their prime but still serviceable enough. There should even be diodes or possibly a bridge in there that you can use as well.

A word about AC and rectified and filtered DC is in order here though. One thing that surprised me when I first learned it was how voltages rise going through a rectifier, then filter capacitors. It is important to note that if you get a transformer and you measure it to be say 34 volts AC you cannot then say well the regulators can handle 35 volts so I'll be OK. You won't be.  You're going to get something more like 45 volts DC out of that after it is filtered and rectified.

How is this possible you may ask. Surely the diodes will drop a little voltage and how can the capacitors boost the voltage they get? I think the Keebler elves are partly responsible or this phenomenon but I'm not 100% sure. All I know is it is what it is. Something to do with measured AC being a fraction of its peak to peak power may play some small role in this mystery too.

Whatever the cause it is something you should be aware of and watch for as well. It may be a good idea after you assemble this section of the circuit to test and measure its output before you connect it to the regulation board. Better safe than smokey is my motto!

This is something I didn't do on this particular power supply but I use it on many others I have built. The utility of connecting a pilot lamp to a power supply this way is twofold. One It gives a visual indication of equipment being on. Something all pilot lights do. Secondly connecting it in the manner I describe bleeds a potentially dangerous charge off the filter capacitors as well, and gives a visual indication while it is doing so too.

Then there is the added benefit of the lamp fading after the power is switched off which I find rather appealing personally. But that is just me everyone's tastes differ. In any event consider adding this modification to your supply too.

My preferred method for determining the correct value for the Current Limiting Resistor is to simply jumper in a potentiometer temporarily, adjust it from high resistance to where the LED glow looks correct, disconnect the potentiometer from the circuit and measure it's resistance. I'm aware that there are other more "correct" ways of calculating the proper resistance value but hey I'm trying to have fun. Of course it never hurts to have a meter in the circuit in series to monitor the current while doing this. At least then if the LED burns out I might just learn what not to do the next time!

Maybe it isn't such a bad idea to measure the voltage and along with the current measurement to calculate the dissipated wattage just to make sure what sort of resistor is needed. A quarter Watt one might not be enough. Ohm says that Amps times Volts equals Watts. If you're a tough guy like me you'll just grab the resistor to see if it is hot or not. Ouch! Usually I prefer my electronics on the cool side if possible.

One technique I discovered while I was doing this project was an easy way to cut copper clad printed circuit board. I'd used saws in the past with varying degrees of success to cut PCB board. Cuts never came out too straight for me, and the task was laborious, and detrimental to saw blades as well, especially when cutting fiberglass boards.

This time I tried something completely different. I used a hobby knife and a steel straight edge to cut a piece of PCB. The technique I discovered that worked the best was to angle the knife sideways differently in perpendicular relation to the steel straight edge with each pass I made. A bit difficult to describe with words so I made a graphic of what I am trying to impart.

Done correctly thin threads of the board material should eventually be freed from within the cut while making a V shaped groove into the board material. Scoring from both sides of the board will ensure a clean break when the board is cut enough so it snaps apart easily.

Cut in this manner the board's severed edge comes out very straight, although not perfectly perpendicular to the flat faces of the board. Putting a piece of sandpaper onto a table then sanding the board edge a few times should remedy that if it is important to you.

I like doing it this way better than sawing so I figured I'd offer it up as a suggestion to others who may find it works for them as well.