1
\$\begingroup\$

When it comes to validating a power supply (for example, a buck regulator), what parameters matter?

From the research I have done, I think that things such as ripple voltage, percentage overshoot during load step, and thermal performance all matter - what else should I be monitoring/recording?

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Have a look: oeg.co.jp/en/semicon/power-s.html \$\endgroup\$
    – Mike
    Commented May 16, 2022 at 5:40
  • 1
    \$\begingroup\$ Efficiency should almost always be on the list of things to check. \$\endgroup\$
    – Hearth
    Commented May 16, 2022 at 5:46
  • 1
    \$\begingroup\$ Temperature of the obvious bits under worst case conditions. \$\endgroup\$
    – user16324
    Commented May 16, 2022 at 11:39

1 Answer 1

1
\$\begingroup\$

First of all, you should have a design spec document which should definitely have:

  • Input voltage range
  • Output voltage (accuracy, tolerance and, if required, adjustment range)
  • Output power (continuous, peak and other ratings at different inputs)
  • Efficiency (may not be constant and may be given for different conditions)
  • Ripple and noise limits
  • Line regulation and load regulation
  • Environmental conditions
  • Protections with definitions of expected behaviour (e.g. OVP latch, OLP hiccup, OPP latch, etc.)
  • Applied standards and regulations if required (This includes electrical, mechanical and chemical stress)

NOTE: These all apply to a PSU regardless of that if it's an off-the-shelf or a designed-from-scratch.

Then, you can validate the PSU according to the specified requirements.

percentage overshoot during load step

This is defined under dynamic response and dynamic response should be defined well-detailed. The requirements may vary depending on the design but generally, allowed percentages of overshoot and undershoot during different load steps (i.e. 0 to 100%, 100% to 1%, 1% to 50% etc.) with different slew rates (e.g. 1 A/μs) are defined under dynamic response. Sometimes settling time is also a requirement. But what you absolutely don't want to see is oscillation. In some case you may want to measure the loop response with special equipment.

\$\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.