I haven't hooked it up yet but I made a little series regulator with an LM317 chip and a few pass transistors on a heat sink. The 5v & 12v seem to fulfill all of my needs so far.
That looks pretty good Potseal but I would do a couple things a little differently. 1.) Substituting an LM350 for the LM317, will allow > 4 amps with proper heat sinking. Since your +12V output is probably pretty beefy, you might as well be able to use more of it. 2.) Your supply is really about 1.25-9+ volts as-built. Should you really want closer to 0 output, you can use 1 or 2 diodes, in parallel with a cap, and then in series with a resistor to the -5V supply. Referencing the low side of the pot to the bypassed cathode's slightly negative voltage, offsets some or all of the 1.25 volt adjustment voltage. 3.) D2 is usually placed across the 240 ohm resistor when using a filter cap across the adjustment pot. The datasheet shows that 10uF gives about a 15dB improvement in ripple rejection. D2 protects from that cap discharging into the adjustment pin. BUT, if D2 is shown as protecting a battery load from causing damage if connected backwards, there should be a fuse before the + output terminal. 4.) While you don't have to use 100uF as an output cap, definitely use at least 1uF, or you risk oscillations with low capacitance loads. The 1uF or more swamps out that possibility. Joe
I have the transistors, already on the heat sinks (from IBM) and the LM317s are pennies each in quantity. I bought 20 a while ago. I got a bunch of 270 ohm resistors too at the same time so it makes a pretty cheap regulator. Most of the ones I have made are 3vDC, straight off the 317, to eliminate all of the batteries in stuff I have around here. If you run them off of a 5v wall wart the 317 never even gets warm. I seem to have collected a box of those wall warts over the years. USB A cords are handy too since that is the output of a lot of these wall warts, (lighter plug adapters etc)
That looks pretty good Potseal but I would do a couple things a little differently. 1.) Substituting an LM350 for the LM317, will allow > 4 amps with proper heat sinking. Since your +12V output is probably pretty beefy, you might as well be able to use more of it. 2.) Your supply is really about 1.25-9+ volts as-built. Should you really want closer to 0 output, you can use 1 or 2 diodes, in parallel with a cap, and then in series with a resistor to the -5V supply. Referencing the low side of the pot to the bypassed cathode's slightly negative voltage, offsets some or all of the 1.25 volt adjustment voltage. 3.) D2 is usually placed across the 240 ohm resistor when using a filter cap across the adjustment pot. The datasheet shows that 10uF gives about a 15dB improvement in ripple rejection. D2 protects from that cap discharging into the adjustment pin. BUT, if D2 is shown as protecting a battery load from causing damage if connected backwards, there should be a fuse before the + output terminal. 4.) While you don't have to use 100uF as an output cap, definitely use at least 1uF, or you risk oscillations with low capacitance loads. The 1uF or more swamps out that possibility. Joe
Thanks Joe
I build circuits as often as I date supermodels... Which means I appreciate the input.
You are correct about the volts for my variable voltage. To be exact and using my DMM I measured 0.25 V at the low end and 8.99 V at the high end.
I will draw out the entire circuit with everything I used including switches and fuses and let you pick it apart (if you have a moment). I would like to get the most "bang for my buck" so if a few changes can improve the final product that would be great.
A malfunction at the junction -------------------------------------- Dwayne
![[Linked Image from gfretwell.com]](http://gfretwell.com/electrical/5-12v%20PS.jpg)
