0
\$\begingroup\$

I heard from somewhere that for some power supplies, it is important that there be no ripple and 'dips' in the output of a constant voltage output dc-dc converter. The specific cases are:

1.) When dynamic loading or hot swap testing, basically transient conditions it is not desirable for the output to have small negative dips or some oscillations even though they will eventually regulate at a fix value at steady-state.

2.) When the power supply is turned on, the output voltage must be monotonically increasing, if there is an upward spike which is small that is okay as long as there is no downward spike.

I heard that it depends on the load but is this really that important? Are there examples of load you could give which are really sensitive to these kinds of specs?

Also, I was wondering what if, if you apply a step increase in load, the output voltage of 5V dropped a little bit let's say 0.05V, becoming 4.95V and stayed there but it did not oscillate and it regulated there. Would that be better than if the step increase in load caused a downward dip of 0.05 and possible oscillation but in the end after some time to reduce the oscillation it regulated to 5V and not 4.95V. Which is better?

\$\endgroup\$
4
  • \$\begingroup\$ That is for the user of the power to decide. \$\endgroup\$ Commented Jul 26, 2013 at 7:19
  • 1
    \$\begingroup\$ Opening line: where did you hear that because in practise many, many regulators will have ripple and load induced dips and peaks. Do you have a link to what you heard? \$\endgroup\$
    – Andy aka
    Commented Jul 26, 2013 at 7:26
  • \$\begingroup\$ maximintegrated.com/datasheet/index.mvp/id/1196 \$\endgroup\$
    – Atom
    Commented Jul 26, 2013 at 9:09
  • \$\begingroup\$ Its a +3V and +5V µP Supervisory Circuits. \$\endgroup\$
    – Atom
    Commented Jul 26, 2013 at 9:10

1 Answer 1

3
\$\begingroup\$

It's not possible to make a power supply of any kind with no ripple. The ripple is made smaller until it's acceptable by improving the regulator or the filtering.

All of your questions depend on your load and application. If the load is a resistive heating element, you can ripple all you want: the resistor will still get hot. If your load is an extremely sensitive measurement instrument, an unstable or noisy power supply may couple noise into the measurements, which may be unacceptable. If your load is an ordinary microcontroller, a little ripple or slightly poor regulation may be fine; very bad ripple or very poor regulation may make the processor unstable. The microprocessor's datasheet will specify how good is good enough.

\$\endgroup\$

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.