Author Options:

Could i use pwm to vary voltage for a benchtop power supply? Answered

I wanted to create a variable benchtop supply from a computer atx supply. Could i use a MOSFET with an arduino to vary the voltage coming from a 12v power supply?

Initially i wanted to use a LM317 but since that would be inefficient.
The mosfet in question is IRF540N
Here is the circuit diagram



Best Answer 7 years ago

Absolutely Not !!#**.
The idea of a bench top power supply is to deliver a selected voltage to other electrical items and not to blow them out of the circuit.
PWM delivers full voltage to the load during ON time and zero voltage during Off time.
A power supply must have a steady smooth output.


What if i add a smoothing capacitor on the output (4800uF) ?
I accidently omitted it

Here is a Block Diag of what you and orksecurity are  thinking of.
Trust me when I tell you both a group of  EEs would not be up to making one.
Which is why I didn't bring it up.

Of course the experience would make a career for you.
I would help answer more questions if you really try.

A  4800 uF smoothing filter would only damage your MOSFET,
Read and understand the pointer first.



Uh these?

have you tried to get free samples for those?
digikey is a ***** in my country charges $40 for shipping

Try telling NS that you are a design engineer for the Qwerty company and
would appreciate 2 samples of the IC and if it is liked by the management,
you think it might go into production (about 20K to 30K per year ).
While the detail application is secret, it is an impressive toy for the
international market.

I have received parts like these all my life, for free.
So can you !  Saving you  ***  $40.
It helps if you have a publication or two to alude to.
I have a construction article in Electronics magazine to reference.


I live in india and i need a suitable switcher ic for this project (12v input, 2-12v adjustable output) or close.

Good one, you always have a neat IC solution.
Layout looks almost critical ?

Not as simple as a 78xx but definitely no black magic. Used them sometimes and they even worked reliably in a hand wired board, even better on a PCB. Tight design, short leads, components as in the assembly notes. Wont win a prize for creativity but will work.

As you've alluded, switchers are buggers to get right - a VERY tight layout helps enormously.


Yeah i dont think i am up to it, (the pointer) seems quite complicated.
I think i will have to resort to a lm317 based system :/

A benchtop supply should ideally be a linear supply to minimize radiated noise associated with the switching waveform into the variety of circuits you may use it to power (ie, from rough circuits like motors to precision like signal conditioning...) Some will be virtually immune, some will be rather displeased by that noise.

To use a PWM method for producing a a variable DC voltage, rather than a pulse-width modulated fixed dc voltage which uses the natural RLC characteristics of its load to produce the equivalent of a DC voltage, you need a more complex design, often employing transformers or RLC networks to minimize  output noise and potentially other downstream  regulation components.

National Semiconductor has a bunch of useful documents on the subject for its "Simple Switcher" line...

Example part with links to possibly useful docs :

Homepage for simple switcher here

What you're trying to create is, essentially, a "switching-mode power supply." Standard designs for that do not need a microprocessor in them.

Websearch the quoted phrase for more info, including at least one design tutorial.