High Current regulated PSU using pass transistors

[pinnocchio]

I'm looking at some help designing a high current regulated PSU using pass transistors. So I'm asking help of PSU wizards like Salas and others

Ideally 15A-20A min. with positive rail only (maybe bipolar in the future). I'm looking in the range of +45Vdc as output

I have found this design but it will need to be adapted and this is where I will require your wisdom. It needs to be really clean supply since it will be used for powering an amplifier. Impedance and speed is also a factor...

Here's what I have found

LM7812 High Current Power Supply by TIP2955 Pass Transistors | EEWeb Community

An externally hosted image should be here but it was not working when we last tested it.

Thanks for any help!

Do

 

[abraxalito]

To achieve a really clean supply I'd suggest LCLC filtering prior to any active regulator and potentially after the reg too. Cleanliness of supply really matters only right at the amp's power terminals - any intervening impedance in wiring will add noise (assuming this is a classB amp you're powering) so it'll be best to optimize the amp and power supply together as a complete system.

 

[gmphadte]

The fuse is not needed at the output of 7812 as it is short circuit protected. You need resistor at each collector.

Gajanan Phadte

 

[geraldfryjr]

I just designed a similar circuit but using NPN power transistors.

See this thread,

http://www.diyaudio.com/forums/chip-amps/261125-lm3886-problem.html#post4064510

I will also eventually be adapting this design for voltages up to 60Vdc input as described in the posts, I have already described these notes in the discussion.

I have made another similar power supply many years ago with a 15V output and it was able to do about 30 amps at 15v and about 50amps or so with a 9v output, I used 4 NPN power transistors in it.

I used a transformer from burned out 450watt computer UPS unit to power it.

Sorry, I don't have the schematic that I designed for that one as it was a bit different as I used an FET to drive the transistors and I used a LM317 to make it variable as well.

But it worked very well as it was a very heavy duty power supply.

For fun I used to use it to make same very powerful electromagnet's out of a few simple spools of wire that I had laying around. 😉

I have not simulated this configuration using PNP's yet, But the version that is similar to this that uses a PNP to drive the NPN's shows poor load regulation and the resistor values need to be tweeked in order to get good performance out of it.

I have already discussed this in the above thread.

It is coincidental that this type of supply design as come up again in these threads at the very same time I have been working on such designs.

I don't see anything wrong with using PNP's types if you already have them but there are many NPN types that are more available and cheaper to use as well.

As a note it is best to have the gain of all of the transistors the same with matching hfe's, or else you will have current hogging issues between the devices.

FWIW

jer 🙂

 

[pinnocchio]

Thanks all for the replies!

I was looking into this Cree high current Mosfet and don't know if they have a P part and if it would be a good idea to use 1-2? Would be easier against current hogging?

View attachment C2M0080120D.pdf

How easy is it to adapt this design to N type Mosfets instead of BJT?

Thanks
Do

 

[geraldfryjr]

Mosfet's work differently than BJT's.

I have a few high current FET's from TI that I have messed with before, and I have been thinking about making a regulator out of them as well.

In fact I am thinking about a Class A Aleph or Aleph-X amp using them as well or something.

The just came out with a 80v and 100v series of these FET's too!

N-Channel MOSFET Transistor | Products for N-Channel MOSFET Transistor | 44 mosfet battery charger

I have the CSD18533KCS 60v 135amp and the CSD18534KSC 60v 108amp FET's.

All it takes is a simple opamp to drive them as they are Voltage input devices instead of requiring a current type source to drive them.

I was quite impressed with there high power rating and current handling when I was testing one with my 32Vdc supply!!

I burned up one or two of them but that was an extreme uncontrolled test with just a pot to control the gate voltage.

jer 🙂

 

[gmphadte]

If you are thinking high current from mosfet, go for 2N3773, NPN bipolar.

Gajanan Phadte

 

[Elvee]

Here's what I have found

It is a typical cliché of ~high current supplies; but as most clichés, it leaves a lot to be desired.
With an opamp combined with a voltage reference, it is possible to do much better.
A TL431 more or less mixes both and can provide a very simple and effective solution: here is an example (if you don't need the current limitation and indication, just discard them).
Use one 2N3055/3A load current with balancing resistors.

Stability, regulation and dropout will be better than for your example.

 

[pinnocchio]

It is a typical cliché of ~high current supplies; but as most clichés, it leaves a lot to be desired.
With an opamp combined with a voltage reference, it is possible to do much better.
A TL431 more or less mixes both and can provide a very simple and effective solution: here is an example (if you don't need the current limitation and indication, just discard them).
Use one 2N3055/3A load current with balancing resistors.

Stability, regulation and dropout will be better than for your example.

Oh, very nice!

I will most likely start with only one output BJT for testing then scale from there.

Thanks
Do