# Dual variable power supply with LM337 and LM317

I am making this circuit:

C1, C2 2200 μF 50 V Electrolytic Capacitor

C3, C4, C5, C7 2.2 μF 50 V Electrolytic Capacitor

C6, C8 100 μF 50 V Electrolytic Capacitor

R1, R4 5K Potentiometer

R2, R3 220 $\Omega$ 1/4 W Resistor

D1 to D4 IN 4007 Diodes

U1 LM317 U2 LM337 T1 24 0 24 Center Tapped 2 Ampere Transformer

the goal is to create a power supply that can give a variable supply of -15V -to-0 and 0-to-15V

I have already placed all the components in the breadboard and tested the output. The problem is the readings on the positive side is from 1.6 V to 30 V when I turn the potentiometer from he left most to the right most same with the negative side( but the sign is negative)

What can I do to get the specific range I need of of -15V -to-0 and 0-to-15V

• I was going to comment that you can't get zero volts out of this, but then I realized that you actually can, because you have a negative supply (and vice versa) – pipe Jul 31 '16 at 18:39
• Are the midpoints of C1/C2 etc. connected to ground? I don't see a dot at the connection. – markrages Aug 2 '16 at 0:58

The simple solution is to install a 4.7K in series with each of the 5K pots leads going to ground. That should cut the 30vdc to about 15 vdc for both positive and negative supplies, at the 'maximum' setting. That is two 4.7K resistors total.

Since the LM317/LM337 series have an internal 1.25vdc reference, the easy fix is to install series diodes in the outputs, after the capacitors, causing a 1.25 volt drop. Two each 1N5408 3 amp diodes plus using 2.2K 1 Watt resistors across the outputs as minimum loads will bring the outputs close to zero volts at the 'zero' setting. That is 4 1N5408 diodes total and 2 2.2K 1Watt resistors total.

Be sure to heatsink the IC's if drawing more than 200mA from them. Also, a room temperature change or regulator change of 20 degrees F will cause at least a 100mV change in the output voltage. These are general purpose voltage regulators, not precision regulators.

There are more accurate solutions that use zener diode and op-amps and trim-pots, but I was not sure what level of priority that had.

EDIT: Your probably using 5% resistors, so don't expect precise tracking or that maximum volts is = +15.00, etc. Or that zero volts is equal to 0.00. Or that a substantial change in temperature of U1 or U2 causes a change in voltage. If you want better accuracy replace the 4.7K resistors with 10K trim pots. Put their wiper positions in the center to start with, then trim for max voltage = +/-15.00vdc.

• You're missing a trick here. See my answer to Making LM317 output voltage adjustable down to 0 V and Russell's suggestion for further improvement. – Transistor Aug 1 '16 at 22:53
• @transistor. I had suggested as much to the OP about more accurate solutions, including tracking regulators. I have built all of them over the decades, but the OP seems content with a simple if-not-accurate fix. You could post an answer to get the OP's attention. Worst case is a null or +1 vote. – user105652 Aug 1 '16 at 23:22
• @transistor. OK, I'll give you a +1 for posting that more accurate option as an answer. The OP should see the options, if he is still paying attention to this page. – user105652 Aug 1 '16 at 23:41
• I like to ask is it possible to remove some of the capacitors? and replace it with a different value? what benefit does having that kind of set up with capacitor have? – Edward Montilla Aug 7 '16 at 6:34
• You could remove some of the low value capacitors, often placed close to the LM317/LM337 for stability. The large value capacitors keep the voltage study under changes in load. You have an ideal capacitor layout already. Remove them at the risk of loosing voltage stability. – user105652 Aug 7 '16 at 18:58