GizmoWizard

I intend to build a bipolar, variable, lab power supply (non-switching), say plus and minus 3 to 25 VDC @ 1 to 1.5 amps. The National datasheets for the LM317 regulator series shows such an application, however the output voltage is set for both the positive and negative regulators using a ganged dual pot.

I would like to replace the dual pot with with a device(s) that uses a control voltage (from a microcontroller) to set the output voltage. Where things get fuzzy for me is the complimentary control element for the negative regulator. Any ideas?

Thanks

Roff

Do you want one power supply that will swing both positive and negative (through zero), or do you want dual tracking positive and negative supplies?

__________________
Ron

GizmoWizard

I need the dual tracking positive and negative supplies.

Roff

Here you go. I haven't built this, but I think it will work.
The references in the LM317 and LM337 aren't used to control the output voltages. These regulators are simply used as bullet-proof, high power voltage followers. Due to the voltage across the regulators, you will need massive heat sinks if you want to run high current at low voltage.

EDIT: Fixed polarity of C10.
Thanks to L. Chung.

EDIT: Added balance pot.

Attached Images

tracking_dual_supplies_bal1.png (21.1 KB, 374 views)

__________________
Ron

GizmoWizard

Wow Ron! I wasn't expecting you to go to this much trouble. I thank you very much.

I don't understand the LM317 and LM337 (with their Adjust pins tied to gnd) feeding the op-amps. How does this work please?

Roff

I was afraid you were going to ask that. First, notice that they are the "L" versions. in TO-92 packages (or maybe SO packages).
It's a long story, which I am going to make short. The op amp outputs are always 1.25V below the output voltages. I wanted to power one of the supply rails from the output, and the other as close to GND as possible, in order to minimize the voltage across the op amp, and to maximize the output range. These requirements called for a (relatively) high-voltage, rail-to-rail op amp. The LT1677 fits those requirements, but I was not comfortable running one rail at GND when the common mode input voltage is at GND, so I made the other rail 1.25V. Hence the LM317L and the LM337L. They put out 1.25V above and below GND, respectively.
See, I told you it was a long story. And that was the short version.

__________________
Ron

eblc1388

Polarity of C10 is shown reversed. This is an easy mistake to make when drawing +/- supply circuit. I have done the same many times.

__________________
L.Chung

Roff

Oops! I copied and pasted it. ops:
It's been fixed. Thanks, LC.

__________________
Ron

GizmoWizard

I'll order the op amps, breadboard the circuit, and let you know how it works out.

Thanks again,
Giz

Roff

Attached Images

crossed_fingers.jpg (2.7 KB, 383 views)

__________________
Ron

Mike, K8LH

I wonder if you couldn't simplify the design by using a pair of Maxim digital pots wired with the 'up' and 'down' push-button inputs in parallel and the variable-resistor outputs controlling the adjust pin on a pair of regulators?

Never mind... Ron's design is better... Forgive me, I posted before thinking (duh)...

Regards, Mike

Roff

To get the best possible match between Vpos and Vneg, you might want to put a 10k pot between R1 and R2, with the wiper going to the inverting input of U2. I've modified the schematic above.

__________________
Ron