# Parallel power supplies?

Discussion in 'General Electronics Discussion' started by rdoty, Aug 17, 2015.

1. ### rdoty

2
0
Mar 6, 2010
I have a requirement for 36VDC at approx. 30A. Is there any problem with running two 15A power supplies in parallel? I presume I would have to isolate with diodes.

2. ### AnalogKid

2,428
692
Jun 10, 2015
Depends on the supplies, but the diodes help a lot. A problem with them is the negative temperature coefficient of Vf. Whichever diode starts out conducting more current heats up more, it's Vf decreases, and it becomes even harder for the other supply's diode to conduct. Some of this effect is offset by the way Vf increases with current. Still, you need to be able to adjust the two supply outputs fairly precisely to get anything close to 50/50 sharing. If the load and ambient temperature are constant, then maybe. Otherwise, one supply at 100% rated current and the other at only 80% might be all you can sustain.

ak

3. ### Colin Mitchell

1,417
313
Aug 31, 2014
If you put 3 x 10 amp diodes in parallel from each supply you will get very good current sharing, providing the voltage from each supply is absolutely identical.

4. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,449
2,809
Jan 21, 2010
if you can tolerate the small voltage drop that a diode gives you (could be in excess of a volt) plus a bit more, you can add a series resistor to each diode that will do about 1V at full current from that power supply. This will give you much better current sharing. Having the output voltages as close as possible is also a requirement. Depending on the power supplies you might be able to dispense with the does but its safer if you don't

5. ### Colin Mitchell

1,417
313
Aug 31, 2014
Feel the diodes and that will let you know if they are sharing.

6. ### BobK

7,682
1,686
Jan 5, 2010
Diodes are not good for current sharing. They will work fine if the voltage is equal, but a small voltage difference will cause a large current difference due to the exponential I/V curve. Resistors work better. I would go with Steve's diode (to prevent reverse current) + resistors combination.

Bob