0
\$\begingroup\$

I just made a 32V to 16V 2 Amps buck converter with a 10mH inductor, I'm having an issue with the voltage ripple output. Too many spikes, and it's generally not consistent. Electrolytic and Film capacitors don't do the job. Does anyone have any suggestions?

The Blue waveform is the voltage output with a 470uf electrolytic capacitor. The red waveform is just Vgs so you can ignore that.

EDIT: I'm using N-Channel MOSFETs that are high side with a bootstrap. The diodes are DSEP30-12A. The reason I use such big diodes is that I plan on increasing the input voltage to a few levels of magnitude later on.

enter image description here

\$\endgroup\$
6
  • 4
    \$\begingroup\$ My suggestion is to post the schematic. \$\endgroup\$
    – Kartman
    Mar 8, 2022 at 2:51
  • \$\begingroup\$ What type of diodes and FETs did you use? \$\endgroup\$
    – Gil
    Mar 8, 2022 at 3:05
  • \$\begingroup\$ Long wire inductance is the main reason in 10:1 probe ground i.e. measurement error \$\endgroup\$ Mar 8, 2022 at 5:22
  • 1
    \$\begingroup\$ That's not ripple. Ripple happens at the switching frequency. That is ringing that happens at each edge of switching on and off. What type of electrolytic you are using and why? Which buck converter/controller chip you are using and why the cap is 470uF? \$\endgroup\$
    – Justme
    Mar 8, 2022 at 7:26
  • 1
    \$\begingroup\$ 10 mH? Did you mean uH? What switching frequency? kHz? Also, I'd be surprised if your oscilloscoppe didn't have a screenshot capability. \$\endgroup\$ Mar 8, 2022 at 9:43

1 Answer 1

3
\$\begingroup\$

As justme said, those big spikes you are seeing are ringing, not ripple. In fact, what you have there is really ringing imposed on your ripple.

For typical DCDC convertor output, you should expect ripple frequency on the order of your regulator switching frequency (Typically 1-5 MHz) with a peak to peak voltage on the order of 5-20 mV.

I cannot see your full test setup, but typical probing technique is not adequate for this application due to the low voltage and presence of high frequency noise.

You need an extremely short ground conductor connected to ground as close as possible to your test point to avoid coupling high frequency noise through your ground loop, which acts as an antenna. I recommend the paperclip method shown in the link below:

https://training.ti.com/power-tips-measuring-vout-ripple-dcdc-converters

Long probe conductors also increase ringing by adding inductance to the signal path (Together with stray capacitance you get resonance effects and ground lead corruption that cause the ringing to blow up your oscope output).

This will reduce the ringing quite a bit, but to see the ripple only you will want to either bandwidth limit you oscope to (Typically) 20 MHz, or else build an external filter if you don’t have this function.

Finally, using active probes instead of passive will improve things further as you reduce the capacitance effects added by the probe.

For full explanation of all these effects (Ground lead corruption, etc.), see:

https://www.electronicdesign.com/technologies/test-measurement/article/21799660/oscilloscope-probes-influence-measurements

There are some good app notes out there from TI, etc. for further info if you search for “measuring dcdc converter output ripple”

Cheers!

\$\endgroup\$
1
  • 1
    \$\begingroup\$ That was informative I knew it wasn't actually Ripple, but I didn't know how to describe the issue at the switching event I'm going to test what you said. \$\endgroup\$ Mar 9, 2022 at 14:22

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.