I have an 18V 1.5A mains transformer (outputs around 30V peak with no load at 50Hz) followed by a full-bridge rectifier and a filter capacitor:

(there would be a schematic here if SE allowed me more than 2 images)

Now I need a negative supply (which will be regulated to -2V with LM337 and draw around 10mA, so it should provide at least -5V). For this purpose I added a negative charge pump:


At first it seems to work fine, the voltage at the output of the charge pump goes to around -27V after a few periods, however as the current is drawn it starts to sag and slowly approach zero volts. When I was trying to figure out what is wrong, I noticed that at the same time the voltage at the main filter capacitor rises from 30V to about 60V, which is undesirable. It seems I got double the input voltage, but not where I wanted it. To illustrate, this is a graph from LTspice:

Voltage graph

Is there any way to prevent this behavior? To keep the filter capacitor at 30V maximum while keeping full-wave rectification (i.e. the 18V 1.5A capability) and at the same time supplying a workable negative voltage for a small load via a charge pump? I don't want to use an inverting DC-DC switcher in order to avoid extra noise.

(Sorry if I misused a term or two, I am a hobbyist still learning this stuff.)

Last minute observation: I noticed this behavior only happens if I draw more current from the negative rail than from the positive rail, so I guess one way around the problem would be to design the rest of the circuit so that more current is always drawn from the positive rail. However, I don't like the idea of positive supply rail potentially shooting from 30V upwards damaging other components, and I would like to prevent it altogether if possible.

  • \$\begingroup\$ If you change the topology with a proper CP (powered between the two outputs of the AC source), it works. But this is acceptable if you can accept different grounds in your circuit. \$\endgroup\$ Jun 18, 2017 at 18:53
  • \$\begingroup\$ Unfortunately, I need a common ground for both outputs. \$\endgroup\$
    – fishgi
    Jun 18, 2017 at 19:02
  • \$\begingroup\$ What is Vout used for? I was thinking that with a simple bridge you can ge approximately +-18 V, and it might be enough for your needs. \$\endgroup\$ Jun 18, 2017 at 21:58
  • \$\begingroup\$ Thank you, I know what you are suggesting. However, Vout is the main supply rail that needs all the power the transformer can output, Vneg is just auxiliary and will draw little current in comparison. \$\endgroup\$
    – fishgi
    Jun 19, 2017 at 14:22
  • \$\begingroup\$ There also are integrated CPs that can do what you need, starting from the DC voltage. This might also be an option, I am sorry I can't really help with your design. \$\endgroup\$ Jun 19, 2017 at 14:46

1 Answer 1


I'm going to solve this by placing a 30V zener diode across the main filter capacitor, between Vgnd and Vout, to keep it at 30V maximum. In case more current is drawn from the auxiliary negative supply than from the main positive supply and the voltage starts to rise as described in the question, the zener diode will kick in and compensate by equalizing the difference in currents. I have just tested it on a breadboard and it seems to work well.

  • 1
    \$\begingroup\$ that's happens if you have a floating ground , i.e. no center tap and very low input frequency.. ( huge caps) with also a very hot zener \$\endgroup\$ Jun 23, 2017 at 18:41
  • \$\begingroup\$ tinyurl.com/yc8btkwf sim \$\endgroup\$ Jun 23, 2017 at 18:48
  • \$\begingroup\$ Your unregulated supply will have more noise than an inverting SMPS ( well designed) , far less efficiency and poor load regulation. Your problem is common.. Poor design specs (& false assumptions) resulting in poor design. \$\endgroup\$ Jun 23, 2017 at 18:54
  • \$\begingroup\$ Poor specs? False assumptions? Could you be more specific? I don't need anything fancy, I'm just trying to make do with what I have available (i.e. a transformer with no center tap) and maybe learn a thing or two in the process. Also, the zener gets barely warm (around 0.3W max) and there will be a regulator on both outputs. \$\endgroup\$
    – fishgi
    Jun 23, 2017 at 20:23
  • \$\begingroup\$ it is false to say SMPS are noisier than unregulated supplies. No specs include voltage tolerance no /full load and current rating as well as tolerance on input range. No specs on caps ESR . Have you tested worse case? \$\endgroup\$ Jun 23, 2017 at 21:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.