DIY Lab Power Supply
Mar 5, 2012 at 8:45 AM Thread Starter

#### Possede

##### 500+ Head-Fier
Hello all, been a while since I've last posted. I am currently in the process of building a variable power supply that is capable of supplying a positive and negative DC voltage (to be used as part of my University project class). However, I have always wanted to build one anyway, as it can be used for a variety of low current situations, e.g. powering opamps, tinkering with arduino projects, etc.
I have already built a prototype circuit (on a stripboard) based on this bipolar design:

I used a toroidal transformer (dual secondary), with the following specifications:

Secondary Voltages: 2 x 25Vrms
Secondary Current Nom: 1A
Power Rating: 50VA

What I have done is attached both the secondary 0V outputs to ground, and the 25Vrms outputs to each side of the diode bridge. After the diode bridge and smoothing capacitors the DC voltage level was around +/- 40V [i.e. 29Vrms (unloaded voltage from secondary output) * 1.414 = 41.0Vpeak), then minus the voltage drop across the diode bridge, gives around 40Vdc]. This allowed the LM317 and LM337 to give their full range of rectified output voltage, +/- 1.2V to +/- 37V.

This design worked as expected, but now I am in the process of either buying a PCB, or designing and getting one made through my University. I recently came across the PCB's made by Twisted Pear Audio, i.e. the LCBPS and the LCDPS. I understand that the LCBPS is practically the same as the design I have used above,but uses two diode bridges to achieve the same result, and the LCDPS utilises two LM317s to create a separate dual positive rail power supply (is that the correct terminology?).

My question is regarding the design of the LCDPS, shown below.

As I already have a dual secondary toroidal transformer, what design would be best to use? Is it possible to achieve a positive and negative output from the LCDPS, by attaching the two middle connections to ground? i.e. the top connection is at a positive potential (wrt ground), and the bottom connection is at a negative potential (wrt to ground)? What are the advantages of disadvantages of one design over the other?

I appreciate any feedback that is given, and hope that this made sense. Let me know if anything doesn't!

Mar 5, 2012 at 8:51 AM
Here is a dual rail supply I did.
You are welcome to use the layout.

Mar 5, 2012 at 12:43 PM

Quote:
Here is a dual rail supply I did.
You are welcome to use the layout.

Hello Avro_Arrow,

Thank you very much for your input and allowing me to use your layout design. I am trying to weigh up if designing a PCB and getting it made through the University is worth it, due to time scales. I'm thinking I might just buy either the LCBPS or LCDPS from Twisted Pear Audio, to save time (I have other parts of the project to do, i.e. programming and construction). If I do end up going down the designing and making a PCB, I'll definitely use your design as a reference.

Any chance if you know the differences between the LCBPS and the LCDPS?

Mar 5, 2012 at 1:18 PM
The LCDPS is a dual LM317 power supply. Basically two positive adjustable regulators.
The LCBPS is a dual rail LM317/LM337 power supply. Basically one positive adjustable regulator
and one negative adjustable regulator.

The example I gave you is like the LCBPS.

Mar 5, 2012 at 7:47 PM

Quote:
The LCDPS is a dual LM317 power supply. Basically two positive adjustable regulators.
The LCBPS is a dual rail LM317/LM337 power supply. Basically one positive adjustable regulator
and one negative adjustable regulator.

The example I gave you is like the LCBPS.

Okay doke. I have a few more questions regarding the build (regarding displaying the voltage and current), that I'll post in this thread in the next few days. Thanks again Avro_Arrow.

Mar 5, 2012 at 8:43 PM
Arrows build is simple and well designed. The only thing I want to add is perhaps instead of using two potentiometers to adjust voltage on both rails...you can use a circuit like this:

LT150A is EOL or discontinued I believe, but it can be replaced with another regulator like an LT1083 or maybe the standard LM317 (not sure on the LMs). It allows you to control both rails simulaneously and you don't need a dual tracking pot to do it, which can get out of calibration fairly easily, and you can't easily recal it.

Mar 6, 2012 at 6:56 AM
I've built that and it didn't track very well. It was OK at one end of the range
but way off at the other. Maybe you had better luck...

Mar 7, 2012 at 4:04 PM

Quote:
Arrows build is simple and well designed. The only thing I want to add is perhaps instead of using two potentiometers to adjust voltage on both rails...you can use a circuit like this:

LT150A is EOL or discontinued I believe, but it can be replaced with another regulator like an LT1083 or maybe the standard LM317 (not sure on the LMs). It allows you to control both rails simulaneously and you don't need a dual tracking pot to do it, which can get out of calibration fairly easily, and you can't easily recal it.

Hello TheLaw,
I have already made the design posted by Avro_Arrow on a strip board, and it works as expected - so I will stick with that. But thanks for your input!

Out of interest, does anyone know of a more sophisticated power supply design? Say, +/-50V, 3A/5A output, current limiters, etc. I have had a look myself, and I know there are many, but most seem to not be that well documented, or contain problems that people have encountered whilst building the design. I thought that if I'm building a power supply, I may as well go the whole hog (unless it proves too difficult). Any PCB layouts or even professional PCBs would be more than handy.

Cheers

Mar 7, 2012 at 4:34 PM
If you check through the data sheets you can sometimes find interesting circuits.
Read through the LM338 data sheet:
http://www.ti.com/lit/ds/symlink/lm338.pdf

Also AMB has a nice power supply here:
http://www.amb.org/audio/sigma22/

Mar 7, 2012 at 9:26 PM
Suggest use LDO voltage regulators from LT. They are great!!!