# Dual power supply circuit dropping voltage on applying load

I'm not an Electonics professional, and hope someone can explain why the circuit below, I've used to build a dual rail +15v,0v, -15v Power Supply drops voltage down to about -7.7 volts when connecting -15v and 0v across a 4.7k resistor; the voltage should remain stable.

What I would say is that on removing C5, C9 and C10 out of the circuit it then stablises as would be expected. This circuit, with the capacitors is quite common; could anyone tell me what's happening in this case?

If a load is places across -15V and 0v, and more current is drawn, does this decrease the voltage drop across C9 and therefore adjust the output voltage?

• whats with D10 and D6? – BeB00 Oct 19 '17 at 21:58
• the same with D14 – BeB00 Oct 19 '17 at 22:00
• @BeB00 'Diode D2 protects against capacitor CAdj discharging through the IC during an output short circuit.' Here D2 is D6, D10 and D14, and CAdj are the OP's capacitors in question. – calcium3000 Oct 19 '17 at 22:04
• oh right, it looks like he shouldn't need them though, as his outputs are less than 25V and his CAdg's are not more than 10uF – BeB00 Oct 19 '17 at 22:11
• @Alex, what are you using for your 18v inputs? – BeB00 Oct 19 '17 at 22:27

The old-fashion LM317/LM337 family requires minimum load, 3.5-10 mA. The feedback/reference circuit is usually made of fairly low-Ohm resistors, pin1-pin2 should be 200 - 240 Ohms, to make it work. This circuit has 1k+ between Pin1 and Pin2. This is likely the problem with this PSU.

• I was trying to find an actual spec on the required resistor in the datasheet, but this is what I was thinking – BeB00 Oct 20 '17 at 0:24

These capacitors -- C5, C9 and C10 -- are there to stabilize the output voltage feedback. From the LM317T datasheet:

The adjustment terminal may be bypassed to ground to improve ripple rejection. This capacitor (CAdj) prevents ripple from being amplified as the output voltage is increased. A 10uF capacitor should improve ripple rejection about 15 dB at 120 Hz in a 10 V application.

The datasheet also recommends a >1uF capacitor on the output (though it is 'not needed for stability').

Have you measured the input voltage to the LM337T? Does the IC feel warm? Are you loading down the other outputs (+15V and +5V) while testing the -15V output?

• Thanks, I fixed the issue, on placing a 10uf electrolytic capacitor across the outputs; namely across D8, D12 and D16, the PSU was stabilised and on adding a load to the PSU there was no drop in output voltage as would be expected. Any idea why this would be the case @calcium3000. Thanks again – Alex2134 Oct 21 '17 at 20:48
• @Alex2134 Ah glad to hear! I've found the output caps you added in fact do help with stability, and most other linear regulators specify >1uF output caps are required. Bizarre that this one doesn't. – calcium3000 Oct 21 '17 at 23:35