Audible switching supply whine

Switchers are usually way supersonic, hundreds of KHz, so you may hearing a loop stability oscillation. That should be visible on an oscilloscope as a burst or AM sort of pattern. Probe 1:1 AC coupled at the input or output, or make a little loop pickup and snoop the magnetic field.

Great suggestion. Just connect the probe's ground wire to the tip to form the loop and have a good time snooping.

He does have a 'scope, right?

and while all of them test good, there is a difference between some of the m as far as audible whine. Most are pretty quiet but how concerned should y ou be about some creating more noise? Not bad whine but some you can hear f rom a foot or so away.

Are you a very light sleeper?

use to dampen the noise but to be safe, should I just scrap the noisier one s?

why?

is there a difference?

n expert on.

Or it might just be the regulation in action.

NT

Most of the output capacitors I've seen used in those eBay special switchers appear to be aluminum polymer electrolytics with relatively low ESR. The output cap on my LM2587 is rated at 220uF/35 volts, which seems way too small at any reasonable ripple for the 3 amp current the seller is claiming. The input cap on mine is rated at 220uF/50 volts; some of the ones I've seen have input caps rated at 220uF 35 volts which I definitely don't think is enough margin for a module specced at up to

If it's a loop stability oscillation my guess would be that they skimped on the output cap, it's not really what it claims to be, and combined with the low ESR there's not enough gain margin at 100kHz to keep it stable.

TLDR: I don't trust the capacitors

If it's not instability, it could be a response to load current waveforms with audible content. You're likely hearing magnetostriction in the choke core or electromotive ristling in the winding wires.

You can invoke manually by pulsing the load with a controlled rep rate and duty cycle, at varying input voltages - tuning for the sensitive spots (stability included).

Instability or otherwise, it's variable in audible amplitude, because of differences in winding and varnish application effectiveness between individual parts, and the differences in frequency of the disturbances present in a specific set-up.

RL

The whine comes from loop stability or pulse skipping (hysteretic) mode.

The sound source is most likely the inductor. Ferrite has modest magnetostriction, so it expands when magnetized. Magnetic field going up and down at audible frequency, there you go.

Type 2 ceramic caps do the same thing, if there's a little ripple around.

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website:

Does it point to a "potential" problem or just nothing to worry about?

I'm sure it has something to do with potting compounds or whatever they use to dampen the noise but to be safe, should I just scrap the noisier ones? Could it be a design flaw rather than a mechanical issue?

I'm an LM317 and big ole transformer dude so it's an area that I'm not an expert on.

But how would you know?

Poly caps are significantly more expensive. Or if they're Chinese, I wouldn't trust them to be filled with anything but liquid electrolyte, anyway...

AFAIK, they rate them by what the datasheet says, not what the layout is capable of. So maybe 1.5-2A, thermally limited. Maybe saturation and ripple limited too, given the small parts.

Tim

Acronym finder returns "Too Long; Didn't Read" for TLDR . What the hell?

Is everyone here too lazy to type out their message? Disgusting!

No, the acronym means it's a summary for those too lazy to _read_ the message. ;)

I've had pretty good luck with the eBay specials, as long as I derate their output current by a factor of about 3.

In real hardware I sometimes use a 100uF aluminum electro in parallel with a 10uF ceramic. You get a free lead-lag network that way, which can help when the usual phase lead capacitor on the feedback pin doesn't quite do it.

Cheers

Phil Hobbs

Or it could be "quenching", which might be what you're talking about.

The scope is the key, just get the waveform off that chopper and you'll know.

and while all of them test good, there is a difference between some of the m as far as audible whine. Most are pretty quiet but how concerned should y ou be about some creating more noise? Not bad whine but some you can hear f rom a foot or so away.

use to dampen the noise but to be safe, should I just scrap the noisier one s? Could it be a design flaw rather than a mechanical issue?

n expert on.

More likely "hiccuping" due to insufficient load current.

It's generally two things combining to make the noise. First, something is moving. Usually it's the inductor. Second, something is causing it to vibrate in the audio region. IME, that's normally a compensation issue. You might look for a bunch of audio frequency content in the output. The output filter may take care of it but there should be some evidence there.

Is this a dc/dc module? What are the input and output voltages? What are you feeding the input with?

It's an AC input module, one of those wide range 85 to 265vac types. Output is 5V.

I probably should use something heftier -- using it with a Raspberry PI, which needs up to 2.5a. This one says 2 - 4a -- it gets hot but mot too bad.

Like I say some are dead quiet but a few have audible whine.

I tested some of those $1 "3A" buck converters, they run hot at 2A out and overheat after a few minutes at 3A out, the "5A" ones cost about twice as much but run cool to 4A or more.

if the efficiency and regulation is acceptable, don't worry.

Raspberry Pi needs about 600mA. 2.5A is if you also fill all the USB slots with thirsty hardware.

You might quieten them by adding a 50-200mA (10% book rating) minimum load. Simple flybacks don't like to regulate on transient loads that approach zero.

RL