I am trying to make a low noise voltage regulator.

My concept is to use a symmetrical regulator using LM338 followed by capacitance multiplier (ESP version).

According to Rod Elliot from ESP, the capacitance multiplier is able to source around 2.5 amps whereas the LM338 can source around 5 amps. My question is, if I follow the scheme like this, (Rectifier--->Capacitance Multiplier--->Voltage Regulator) and if my load draws more than 3 amps, will it put too much pressure on the capacitance multiplier?

If so, is there any other way to achieve more current handling capacity from the capacitance multiplier circuit by paralleling output devices?

  • 3
    \$\begingroup\$ let's have a look at a schematic for this capacitance multiplier \$\endgroup\$ – Neil_UK May 7 '20 at 8:21

You're trying to solve something backwards:

Your approach:

  1. Need low noise voltage supply
  2. I know that I can reduce noise using a large capacitor
  3. Therefore, I need a large capacitor
  4. Therefore, I need a capacitance multiplier to give me a low-noise supply.

3 (and, in some ways, 4) are logical fallacies.

What you need is a low noise voltage regulator. The idea that this would be realized with a capacitance multiplier is from the 1950s to 1960s, when no good voltage regulators existed, but people were getting used to transistors and later opamps and were excited about the possibilities¹!

A large capacitor, either a real one ore one simulated through a capacitance multiplier, is just that: a kind of energy storage, a cap. It can help smooth out things, but it's just a smoother, not actually a regulator.

You have all the energy you need to keep things smooth (on the input side of your regulator), so you can do better than just having a large capacitor.

If you look into the data sheets of modern linear regulators, even low-drop regulators (LDOs), you'll find that they can achieve awesome things like Power-Supply Noise Rejections (PSNR) of > 60 dB. No way you can do that with a capacitor, and especially not with one controlled by a less-than-complex feed-forward "regulator":

Full capacitance multiplier

Seriously, all the transistors involved are ancient, and the results will be nothing but a voltage that's worse regulated than one supplied by a sufficiently designed actual voltage regulator based on higher-speed, lower forward drop semiconductors. 1.5 A output current really isn't that much. Use something less ancient than the LM338, and a fraction of the actual capacitors used here, and you'll get better regulation.

more current handling capacity from the capacitance multiplier circuit by paralleling output devices?

Two options:

  1. either that capacitance multiplier works well in emulating a low-ESR large capacitor, then: probably not, because you might end up building an oscillator, could be solved by adding dampening, but that would counteract the purpose.
  2. or that capacitance multiplier has a high output impedance, in which case it'd help, but that would just prove that the idea of using this capacitance multiplier is a bad one.

¹ The fact that this is from an audiophile forum does reinforce that presumption, as they tend to glorify the "sound of old-timey tech".

  • \$\begingroup\$ So the point is; trying to achieve higher current from the capacitance multiplier will defeat it's purpose. Is it? Can you please suggest me a way of achieving regulated +-30V @ 3 amps with the lowest possible ripple? Once again, thank you for your suggestion. \$\endgroup\$ – Nakib Hassan May 7 '20 at 9:17
  • \$\begingroup\$ no, it's not. The capacity multiplier itself doesn't really fulfill it's purpose, and paralleling them only goes to highlight that. Also, you're now posting new requirements: the original multiplier was 20V and 1.5 A, you're doubling the power. Also, "lowest possible ripple" doesn't exist, there's always a way to be even better. You need to specify which amount of ripple you can accept. Also, already answered the question of how to approach that problem: Get a modern linear voltage regulator. Alternatively, design a voltage control loop yourself. \$\endgroup\$ – Marcus Müller May 7 '20 at 9:20
  • \$\begingroup\$ My requirement is +-30V @ 3 amps, with around 10mv (P-P) ripple. Can LM317 and LM337 used with output power transistor provide the requirement? \$\endgroup\$ – Nakib Hassan May 7 '20 at 9:29
  • \$\begingroup\$ Can you ask that as a new question? It's really hard to answer such things in comments. You will also have to tell us which ripple your power supply has – I'm presuming that in 2020, you wouldn't use a transformer with a bridge rectifier. \$\endgroup\$ – Marcus Müller May 7 '20 at 9:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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