The Shop > Electronics & IC Programing

How to measure rippel of small DC/DC converter.

(1/2) > >>

PekkaNF:
Planning to use a small DC/DC convertter to power up an instrumentation amplifier and maybe later an isolated amplifier. Power supply rippel (and RF-componenet) could screw up an amplifier signal.

Don't remember much how to measure rippel of the power supply. Luckily I rememeber some termonology.

Searched some  "how to measure power supply ripple" videos


And this is the first victim (because I happen to have it):
https://www.tme.eu/en/Document/a29dd13f72ddce9b78226ed02d82913d/qdc2wsil5_15-15.pdf

I could conform the switching frequency and it does seem to have rippel - which is actually harder to measure than I though. I can separate swiching spikes (and see swiching frequency and see if it has any effect on amplifier).

Found out easily that voltage and rippel depends a lot on how you load amplifier.

Two scrneenshots, neither of them even close to trutfull. First shows the effect of gounding clip. Second is measured with a "spring clip" (shortest probe length/grounding loop), 70% of nominal loading - purely resistance and keeping all wires as short as possible.

The scope is RIGOL DS1052E and it is 50 MHZ 2GS/s model. It has very little control over samplig rate/filtering, set it 20MHz and can't verify DC/DC converter ripple one way or another.

I should see characteristic ripple, should I?

PK:
Scope settings look right, ripple period should be either 8uS or 4uS depending on what rectification is used. Maybe it's a good power supply?


PK

PekkaNF:
It does looks good enough. But it was not a simple aim&shoot probe hitting.

Turned out that ground clip picks up all signals there is in the universe. First picture: probe grounding with springgy thinggy.

Second picture: feeder PSU, DC/DC converter not connected.

Third picture: same spot than #1 (primary  of  DC/DC) converter connected and loaded 70%. Looks like could use capacitor on input side.

Fourth picture: Output of DC/DC converter. Looks ok.

Need to wait op amps to arrive and then hook them together and see if the power supply ripple created problems.

Is there any easy rules on ferrite beads and miniature chokes on power line?

eskoilola:
Rules with interference tolerant circuits:

* Keep the impedances suitable - not too high and not too low either
* The Op-Amp inputs should be grounded at he same physical point. This is to avoid Your ground-clip effect  :zap:
* The supply traces should be as short as possible and be shorted wit a low ESR capacitor. The capacitor does net need to be large - just low ESR. Maybe a 0u1 ceramic surface mount
* If you can filter out high frequency components out of the ripple then the Op-Amp can take care of the rest if there is no differential interference in the inputs. The common mode rejection usually takes care of common mode interference.

PekkaNF:
Yup. Thank you, sounds vaguely familiar. There is ofcourse a small problem....

Planning to use this DC/DC converter to feed:

1: isolation amplifier, low voltage(immediate need is <60V, really would not mind to build final version to survive three phase 400 VAC), low frequency (isolation amplifiers are not fast) AMC1200 + some pre amplifier - which I haven't fiqured out yet.

2: small signal (TTL and such) differential amplifier. Nothing too fancy, my scope is 50 MHz. Video amplifier should do. There are many designs, some very ambitious and very involved.

There are some hurdles to overcome...some are very trivial to those who are in the know, but I'm fighting to understand.

Right now I am in PSU rippel measurement/definition phase.

Next thing is to figure out how to transfer that signal to scope. For low frequency righ now shielded and twisted pair cable, maybe even ethernet cat5e cable looks pretty good. Traditional 50 ohm transmission line feels safe, but my scope is that basic - it does not even have internal 50 ohm termination.

If I need passive probe, I'll buy it. I wish there were cheap 1:5 or 1:10 ready made compact scope probe for 0,1" square test pins - maybe I need to buy cheap one and bucher one to fit a pinheader.

Pekka

Navigation

[0] Message Index

[#] Next page