EEVblog 1607 - PSU Switching Noise Reduction via Vibration

[@EEVblog]

Thanks to Rohde & Schwarz for finding the IEEE paper on PSU noise reduction via vibration that I demonstrated in this video. www.eevblog.com/forum/blog/eevblog-1607-psu-switching-noise-reduction-via-vibration/msg5424887/#msg5424887

[@egonotto4172]

It is fitting that the authors of the article are called First and April in Polish.

[@LawpickingLocksmith]

Who was it? Rohde or Schwarz?

[@Orbis92]

How long will it take Audiophile scammers to build and sell a shaker table for all their equipment to "get rid of the 50/60Hz noise"... ;)

[@billr3053]

Love it! My company is already on it: we added a green edge for even superior performance than our competition.

[@EEVblog]

Damped with cryogenically frozen springs.

[@alistairwilson5344]

@@EEVblog Finished in fairy dust and farted on by a unicorn.

[@MikeDawson1]

they've been selling them for decades

[@calholli]

Nah.. You sell them vibrating chairs.. It cancels the noise inside the ear drum... I would show you on the oscilloscope, but I can't probe your ears.. You'll just have to trust me . :)

[@zahlex]

Every year, reliably on the minute 👌 Thanks for keeping up this tradition 😀

[@IanScottJohnston]

This is the "Dynamic Acoustic Vibration Effect"........or DAVE for short.

[@EEVblog]

If no one can find the original reference, I'm claiming naming rights.

[@KeritechElectronics]

...and it's especially visible if using a manually controlled inertial kinetic impulse injection device, a.k.a. hammer - look at them parts go microphonic when you hit'em!

[@KeritechElectronics]

@@EEVblog you're go for launch!

[@calholli]

@@EEVblog AI gave me a word salad and I've reduced it to "anti-resonant destructive interference" --- When a wave is resonant but is 180 degrees out of phase, they cancel each other out. There's probably some sort of resonance ringing somewhere in the circuit layout.. and the.................... who am I kidding.. lol I don't know

[@Edisson.]

🤣

[@cheater00]

For a number of years now, work was proceeding on the _crudely_ conceived idea of an instrument that would not only supply switched mode regulated voltage for use in unilateral electronic circuits, but would also be capable of automatically synchronizing mechanical vibration for induced switching noise reduction. Such an instrument is the turbo encabulator.

[@peterlarkin762]

Ah yes, haven't heard about the turbo encabulator in years. Legend has it the military bought all the rights and it's now used to encabulate vibe and rations..

[@lightningdemolition1964]

Posted one day early.

[@Chriva]

Everyone, take a shot every time Dave says Alienware

[@demoncloud6147]

ALIENTEK upgraded to Alienware

[@LawpickingLocksmith]

So do the Bird Watchers?

[@zoenagy9458]

gun shot?

[@Chriva]

@@zoenagy9458 Obviously not

[@cheater00]

Dave what you should do is you should take this circuit, pull it apart into small PCBs connected with long flexible wires, and try vibrating one each while the others remain damped. See where that takes you. Is it just one part, or does the whole circuit need to be shook. Etc. I think this could be an iiiiiinteresting deep dive video.

[@joopterwijn]

Shake the electrons, do not stir…😂

[@EEVblog]

Bond, Valence Bond.

[@francomarianardini681]

@@EEVblog best answer ever! :) Happy Easter from Italy, Dave!

[@gblargg]

Is it okay if I spin them?

[@KeritechElectronics]

@@EEVblog don't break the Bond or the electrons gonna fall out!

[@DW7BPB]

This is one of the reasons why I still love this channel!

[@captainchicks]

Seeing "1 day ago" below the video caught me at first. But on second thought, it is April 2nd in Australia already, of course. Nice one. Love it!

[@MrMersh-ts7jl]

Imagine seeing your piece of equipment on a highly viewed channel and the excitement that would bring only to hear the words alienware over and over and over and over again?

[@Anthocyanina]

this is the best example out there of sinusoidal repleneration, amazing!

[@Gameboygenius]

It's a nice trick, but today we have modern materials to solve the problem. Just make the base from pre-famulated amulite. Takes the modial interaction of magneto-reluctance and capacitive diractance right out!

[@aneeshprasobhan]

​i didn't understand a single thing in your statement and i call myself an electrical engineer 😅😂​@@Gameboygenius

[@henrychan720]

@@aneeshprasobhan You might just need to adjust your sinusoidal dingle arm.

[@Gameboygenius]

@@aneeshprasobhan You should read up on Turboencabulators. Have a nice March 32nd. ;)

[@Valenorious]

@@Gameboygenius Capacitive diractance. You might be on to something here. Dave is shaking this thing with a continuous sinewave in search of amplitude changes. Maybe the secret in nulling out the noise is in capacitive diractance. As in the measure of compressed attenuation after dropping it onto the floor for only once in eternity on the condition there is no bounce.

[@cberge8]

Not this year Dave. You get me every year with these. Lol

[@roscozone8092]

He's having another suck of the sav....

[@trevorjansen4102]

Naim would be proud of you. This is their entire premise

[@ClausPriisholm]

I think the exact frequency must be 240.401 Hz for it to work

[@voxtelnismo]

420.69hz

[@lorenzol.8798]

More sophisticated every year!

[@DaRealBzzz]

Dave's once again living in the future.

[@mcconkeyb]

Early in my career I worked for a few years as a test technician performing temperature and vibration testing. It is a very interesting topic, at one point back in the day I was doing some research on the phenomenon of mechanical vibration heterodyning. But the management of the company changed and I left for greener pastures. The place I worked at had 2 electrodynamic shakers, the big one could do from dc to 2 kHz, below 5 hz the G's were limited by the stroke of the machine, which was +-5 inches, above 5 Hz we could do 100 G all the way up to 2 kHz. The amplifier was in 13 very large cabinets and had a total output of 68 KW It was water cooled and a pain the the butt to do maintenance. This is a topic that every EE should know about, as there is a few strange things that can happen if you ignore mechanical vibration (and shock).

[@cremvustila]

🤣one of the best April's videos, Dave! You're a true professional! Thanks for the laughs!

[@excitedbox5705]

Multilayer ceramic caps are piezoelectric (I have seen them used to steer probes on Atomic force microscopes by warping them with an applied voltage.). My guess is that you are hitting a frequency where the physical motion is producing energy at the right frequency to cancel out the noise. Since noise is caused by a fluctuation in the charge of the caps, adding more energy in sync with a drop in the system could even out the fluctuations. The only other alternative I can think of is that the layer spacing is changing through compression/expansion and changing the capacitance at a frequency that cancels out the noise.

[@ptamog]

This would make total sense if the vibration frequency was matching some component of the noise. But here he are fighting kHz with Hz

[@EEVblog]

In the seismic industry ceramic capacitor hydrophones are called "benders" for this very reason.

[@MrJosiahCochran]

Wonder if you start playing with phase if it matters weather you are in or out of phase with the ps oscillator

[@carpdog42]

@@ptamog otoh piezoelectric goes both ways, so the origin of a good amount of the noise very well could be the capacitors bending in response to charge/discharge. The induced vibrations. Piezo voltage and strain are not a linear relationship, maybe the added strain from the low frequency signal is causing clipping in the higher frequency one since it just can't drive the voltage higher/lower. Since the signals are multiples of each others frequency, they never get out of sync so the destructive interference doesn't drift? edit: I think this is testable. If I am right, it should also be possible to reduce noise by putting a fixed strain on the capacitors. If I am right, a capacitor under enough strain should produce less noise.

[@bzcup]

@MrJosiahCochran I was thinking the same, if the vibration is in phase the noise would increase. If it doesn't change with the phase then maybe the second phenomenon occurs.

[@qzorn4440]

Wow, a most interesting video. 🥳 Also, there are many space electronic articles. Cracking in ceramic capacitors is an old problem; it appeared in the 1970s when the first surface mount technology (SMT) chip capacitors were introduced to the market and began to be employed in NASA applications. Dave, great information. Thank you.

[@Edisson.]

Hi Dave, interesting experiment 🤔 everything around us vibrates 🤔 if we vibrate against it it will vibrate less 😁 Thanks to the quasi-quadral correlation of the Brundelivi equation, which evaluates the effect of gravity on an oscillating object and at the same time takes into account the inertia of the counter-oscillation depending on space and time, we can say with almost certainty that we know nothing at all even if the tongue is at the right angle. Nice day 😁 Tom

[@papalevies]

I came looking for this. It's a tradition of mine by now. 😊

[@edmaster3147]

You're down below, everything is upside down, you are where all the electrons wanna be. Thats wy. Great one Dave,

[@dave-d]

Next up: EEVblog demonstrates improved circuit linearity by percusive modulation, (hits things with hammers to shut them up!). You totally crack me up sometimes Dave. What the heck did I just watch? I wonder if the result would be the same with an analogue scope though?

[@leybraith3561]

I note that mrpete222 has discovered that inversion and vibration of a lathe clears debris and leads to better operation!... Great minds at work!

[@pasixty6510]

That’s how I like my PSU: shaken but not stirred.😅

[@Chriva]

Ps: You got me. You got me reeeeal good :D Just now figured out what date it is

[@SathnimBandara]

The only reason I can think of for this phenomenon is the modial interactions of magneto reluctance and capacitive directance of the turbo encabulator ;)

[@jerry1333]

Push it to the limit, 2400Hz and beyond!

[@leybraith3561]

Nice Discovery.... Maybe you should give the effect a name.... how about EEV-AF-2024

[@michaelmoore7975]

As long as you stay off of suspension bridges, you should be OK.

[@KeritechElectronics]

"What REALLY Happened At Francis Scott Key Bridge..."

[@Tom-90210]

Perfect timing

[@EEVblog]

God-like. You beat the porn bots, congrats.

[@berndrosgen1713]

​@@EEVblog😂

[@BritishEngineer]

And perfect timing could be the exact reason why the performance increases when subject to 250hz~. I don’t think the solder joints, traces, leads for the components or the “solid” conductors themselves would be affected by this mechanical vibration, so it leaves us with the inductors, caps and especially the transformer. I was thinking that this frequency was shaking the coils around at a harmonic frequency of the switching frequency, which maybe would’ve caused synchronisation at some point? This could be suppressing the existing inertia in the vibrations of the coils?

[@cheater00]

@@EEVblog the pom bots? what do they sell, potato wedges?

[@Tom-90210]

@EEVblog I was actually referring to seeing a new video just as I was opening KZbin. But..... 😆

[@feraloid]

You could vibrate at frequencies above the switching frequency using an ultrasonic transducer. The medical transducers go into the MHz range.

[@ppdan]

Was hoping you would try the next harmonic (500Hz), also when going up in and back down frequency you skipped 125Hz (went from 120 to 130) ... I guess a good old analog generator would be better to manually sweep and find those "hotspots"

[@universeisundernoobligatio3283]

In my design of minature RF telemetry transmiters installed in expermintal gas turbine engines. Found ceramic decoupling capacators acted as accelerometers, when the engine was running they were under 100,000gs of CF load and 1000gs of vibratory load. Had to replace them with tantalum caps to reduce the effect that was adding false signals to our data.

[@Damien.D]

There's probably a piezzo effect of some kind doing this. Wonder of how much it can be reliability repeated? I bet it's so precise that another unit would have another resonance frequency just because of how a ceramic cap is soldered. Or even the probes attached to it may distord the board a bit and change the frequency. Also tickling each component of the naked board, tightly strapped to a cushioning surface, with a vibrating probe could narrow down to the sweet spot that produce the beneficial effet.

[@CutoutClips]

I was thinking it was something to do with the coils of the inductor vibrating back and forth, slightly changing the impedance when pushing vs pulling current in a way that's beneficial. But that's just a guess. Kind of like the reverse of coil whine, where the changing current in an inductor causes the coils to vibrate and make noise. In this case we vibrate the coils, inducing a changing current at a frequency which opposes the current spikes caused at the switching points. But again, that's just a guess, could be totally wrong.

[@Damien.D]

@@CutoutClips not forgetting that the vibrating gizmo under it produces magnetic fields by design.

[@CutoutClips]

@@Damien.D very good point, that could definitely have an impact, probably even more than the coils vibrating now that I think about it.

[@TeraVoltLabs]

Probably the piezoelectric effect from the mechanical vibration at a *sub harmonic of the switching frequency causes destructive interference with the switching noise in the system? Very neat, wonder if you could set up a simple lc tank oscillator and vibrate each component independently at the resonant frequency and see if you can observe an amplitude change around the resonant freq. Testing different axes of the components could be interesting as well.

[@PhillipRhodes]

This is clearly caused by quantum transconductance of the Jacobian polyphase differential. Fermions bridging the void space around Planck points go into coherent superposition and phase lock the detractor nodes, vis-a-vis the Dingle Arm. It's the same principle we're exploiting in our (Patent Pending) new QUANTUM ENCABULATOR!!

[@toine512fr]

Audiophiles gonna love it! 😉

[@frobstube]

Best April Fools Joke Ever! Your acting skills are amazing, you really fooled me and I've been watching for years.

[@lewsdiod]

Interesting! Also strange that it didn't seem to have any regions where noise got worse, through constructive interference, only got better! 👌

[@LawpickingLocksmith]

When I do a morning walk near the airport I can tell the Airbusses from the Boeings. Airbus uses an active noise suppression system with the result of less thunder but some cluttery noises still slip thru. Senior technician would also remember the old audio amplifier that would pick up noise on their chassis.

[@veryboringname.]

At this stage Dell should just release a PSU and cash in on the free advertising! That was pretty interesting, quite a big difference! I wonder if any components inside have come loose.

[@EEVblog]

I'll take Dell's sponsorship money post-video.

[@McTroyd]

Thanks Dave. Now every fleabay marketing guru is going to do this to goose the numbers. 🙄😉👍

[@Tech2C]

What about the next multiple up, 2.5khz? Would be good to see if it reacts similar.

[@JumpingSpiderDesign]

The chocolate malaise this year didn't help - love you Dave!

[@DanBowkley]

There's a new one now: I read an article recently about how they're using what sounds an awful lot like an MLCC as a novel gamma ray sensor. Really neat stuff. I don't know if it's the same type of ceramic but structurally that's pretty much exactly what they're doing. So now you're gonna have to zap it with gamma rays and see what happens.

[@supercompooper]

This is going to be huge in the hi-fi world😊

[@electrifyingvids3545]

My guess would be the higher amplitude of the sub harmonic vibration overides the low amplitude switching noise some what? Or maybe the sub harmonic creates a beat frequency with the switching noise, and you'd have to slow the time base to see it? I'm not sure. I'm just spit balling here. But this is an interesting phenomenon. I wonder how long it would take until components with built-in vibrators come into the market 😂.

[@massimilianocacciamani7736]

Have you ever wondered why when you put something that vibrates on some bubble wrap, it makes less noise and the sound sounds deeper? Did you also hear about the second law of thermodynamics where energy cannot be created or destroyed? Did you know that f = 1/T is the formula for frequency. If you were to install a piezo speaker on the pcb and rest it on a sheet of bubble wrap, the bubble wrap will act as a "time capacitor" which will initially reduce the frequency until you pop the bubble, and release the trapped chronotons, to give you a "boost" in frequency. The tricky part is to pop the bubbles in such a way that they are in sync with the existing waveform to create a reinforcing frequency effect. Happy april the first. Proud to be a kidney.

[@analoghardwaretops3976]

Well you got the observation of that phenomenon ..on ONE axis of vibration....similarly it should be done for the other two axis ..all three will throw up three different res.freq.results....and also their amplitudes will tell which is the best/worst axis..for final mounting..i.e which to use/avoid.....

[@chewyboy]

Very interesting video. I'm curious what would happen to the noise if you doubled the frequency from 250 to 500 just out of curiosity sake

[@StickySli]

I was learning in my MEE subject about Instrumentation and Metrics about dithering last week, and for some reason, this vibration variable you have introduced seems to dithering duty cycle of the switching converter, reducing the noise; otherwise, increasing the resolution. I wonder if there would be an appreciable difference in a class D amplifier, where a "ADC-to-DAC" topology is used (that's how I like to envision a ΔΣ modulator), and if the ENOB (effective number of bits) of the ADC.

[@blg53]

A purely mechanical equivalent of this phenomenon would be you jumping on some kind of a platform. If then this platform also jumps up and down with the same frequency as your jumps (or a multiple of that frequency) the amount of relative movement between you and the platform (the noise) is reduced. I am not being totally serious about the equivalence :)

[@fedimakni1200]

Dave, Could you please make a video about what kind of changes we need to do to our design if the operating environment has vibration and we want to make our product working ok under those conditions? Thanks.

[@stevenwilliams6258]

Take it apart! Is it all surface mount? Can glue be added to hold parts in place better? Does glue help? We all know or should know that a gentle tap on compnents results in an output bump, at least is audible in audio gear, and particularly tube gear. I add non-acidic silicone adhesive to wriggly parts in my DIY audio gear as do many mfg's. I have never done a before and after test, so in that way, it is fairy dust or poop. How about pot the entire board to dampen internal vibration and test again? Interesting video. Keep 'em coming!

[@Paxmax]

Idea so great it Hertz. Would be cool to know if the same effect will work tomorrow... reason is ofcoz the moon has moved a few degrees 👍😂

[@ernestb.2377]

Interesting to know ceramic capacitors respons to vibrations. Another source could be the tribo effect. That comes from moving cables. For low noise design there are special coax cables with extra carbon layer to minimize this effect.

[@mc_cpu]

Is it that time of year already?

[@michaelwilkes0]

Wow, if this effect could be characterized, a table could be built that would recommend PSU components or circuits given a known vibration environment. Could be a big deal in industry where you often have a known vibration in a machine. I don't expect anyone will be adding vibration to reduce noise, but choosing a PSU design for an already vibrating environment would be simple.

[@earx23]

Interesting. My guess would be cancellation of vibration of the elcos and inductors. But why at a subharmonic?

[@MarianKeller]

This is seriously a thing for fiber coupled laser illumination, when you want to even out the speckle for imaging purposes. You just physically vibrate the fiber in a contraption called a fiber shaker to slightly modulate the phase, creating a varying speckle that evens out during the integration time.

[@aneeshprasobhan]

could it be that the whole output power capability is also affected ? We could test that with a power supply on load and see if something drops at the resonant frequency.

[@waveinversion]

So if you have a power supply that is making a lot of noise, would that be an indication of some ceramic caps getting ready to fail?

[@erikhovdahl]

The probe got a capacitor, and may be swinging to... Pluss the scope is on the same bench as the transducer... and the innput stages there will act upon the low frequency to.

[@TriodeTetrode]

Very interesting video Dave ;)

[@cliveradvan3414]

Do you think this has any relationship with the idea that sound vibrations from speakers can affect (apparently) the sound quality of audio amplifiers . Vibration isolators are supposed to help ( apparently)?

[@EEVblog]

Capacitors are microphonic, see my other linked videos on this.

[@gunderd]

Is it because you're moving the coil through the magnetic field that's created as part of the switching process, and the physical movement causes the field to interfere with itself and acts as a kind of a low-pass filter for the noise? It seems unlikely given the speed at which the field should be propagating relative to the slow-motion movement of the table, however, maybe at the higher frequencies involved in the noise here it doesn't take too much physical movement for the wavefronts to destructively interfere?

[@qutube100]

Could be recto-cranial inversion? Probably due to antipodean time differentials!

[@randomlyselecteduser]

This is why audiophiles try so hard to isolate their equipment. Whether they can actually hear the microphonics from individual components (capacitors, resistors, etc.) I won't even err to debate, but Mr. Carlson's Lab has a video showing some piece of equipment (I'm sorry I can't remember what it was) that made said microphonics audible. Anyway... not surprising.

[@rdson1621]

The vibration probably modulates the value(s) of some component(s) thus changing filtering characteristic (whether an explicit filter or even an implicit one?). Maybe the modulation of that characteristic is even going in pair with the signals getting filtered?!

[@Foetusmachine]

Can you display the vibration signal on your scope on one of the other channels to see where the vibration signal aligns with the noise on the power supply output?

[@gamerpaddy]

wondering where it works too. like using piezos to vibrate a image sensor to affect its noise or something same with intense modulated light hitting a high absorbing surface causing it to produce sound (strong flashlight with pwm dimming onto candle sood deposited on a tin tray can reproduce this)

[@rusle]

Since we have past 1 of Apr. I think we can have some more serious responses now. 🙂 Since it is known that crystals do give a voltage out when hit (lighter) and I suspect that some ceramic capacitors also can affect the signal due to vibration. What about a video regarding those effects and what to avoid when making a designing for use in a area with high vibration.

[@Audio_Simon]

Absolutely fascinating! Not great for product lifetime!

[@thedanyesful]

Great video!

[@pa4tim]

I do not thinks so but just to be complete, Is it the scope or the powersupply ? Because the table will vibrate too. No clue why it reacts to vibrations. Very interesting. Is it date driven ?

[@KeritechElectronics]

Feel the Positive Vibrations!

[@ptamog]

Looks really weird. Are you sure that it is not an issue with the adquisition or the way that it is triggering? Oscilloscope rms calulation may mess up things when working with coherent captures vs incoherent. Screen captures turn much more coherent when you hit the right frequency. The vibration may messing with the PSU control loop and the oscilloscope trigger may be focusing on a particular time.

[@EEVblog]

Pretty sure.

[@ptamog]

@@EEVblog Maybe you can add a stick to that vibrator, open the PSU case and start poking things to see who is responsible of this madness

[@ptamog]

@@EEVblog Maybe you can open the PSU case and try to locate who is responsible. You can try to transfer the vibration locally with a rod.

[@retrobrw919]

Hmmm, I think Dave has a bridge to sell us.

[@izimsi]

can you test it with something like a piezo element taped to the enclosure and maybe the whole thing hanging from a string? Im curious if it would work with much lower power vibration and you can probably push the piezo in the khz range, maybe it gets more efficient the closer it is to the actual switching frequency.

[@edgar9651]

Somehow I thought this sounds like an April 1st joke. And then I realized that Dave and Australia is a couple of hours ahead...

[@mitchell6you]

The European Hyperloop Center (EHC) has opened its first test track. The Dutch developers want to carry out initial tests with the capsules at 420 meters, which will be sent through steel pipes at high speed.

[@collectorguy3919]

So, around once every 1000 cycles is enough for the vibrator to interrupt the component's physical rhythm. Any one of us would get frustrated at a far lower sub-multiple.

[@SuperHaptics]

Is there any chance that mechanical vibration changes your scope probe connection point electrical properties like attenuation? Maybe same thing could be tried with soldered twisted wires instead of a probe? Just in case...😊

[@Pidroe]

Does it do it if you lift it just above the vibrator? (magnetic Interference)

[@EEVblog]

Nope.

[@costarich8029]

Any chance it is correlated to the position of the earth with respect to the sun? It might not work so well in other astronomical orientations.

[@dosgos]

"Silent Running Audio" makes audio equipment stands in NY. Previously, the founder worked on submarine systems. He probably knows what is going on ...but can't tell you.

[@marcusuussalu3016]

I would guess that the piezoelectric effect of ceramic capacitors play a role here inducing currents in the capacitors around the switching frequency. This could then increase the currents in the capacitor if the phase of the vibrations is in line with the rectifying currents of the output capacitors.

[@ka9dgx]

I suspect you're mechanically adding energy to the capacitors (or coils) in the power supply, and thus lowering their loss. You should be able to do the same thing with a tuned circuit, watching insertion loss and/or "Q". You could also hook the power supply up to a VNA and see how it reacts at that frequency when you shake it at that same mechanical frequency, with the supply both on and OFF.

[@ef7408]

Nobel Prize for sure !!

[@peterlarkin762]

This also works when you shake noisy humans but according to my lawyer it's not permitted by IEEE standards.

[@SciFiFactory]

Could the magnetic field of the shaker have any influence as well?

[@patqu738]

Sounds legit... Be in sync with the vibrating electrons

[@whatcouldgowrong7914]

This is fascinating as Japanese audio manufacturers went through a phase in the 90s where bitumen was added to chassis / wrapped around d capacitors, relays etc. Along with copper chassis plates and I thought it was BS. Maybe they were on to something…

[@calholli]

I wish you could compare this with the FNB58 .. that is a USB tester. It has a similar display and can act as a power supply also. I wouldn't be surprised if it were the same electronics on the inside.. It even has a built in oscilloscope and even some sort of software app used on PC to go with it. It would be a great comparison to this thing.. The features and setup are very similar.

[@copernicofelinis]

I found a strange effect in my toroidal transformer when the frequency reaches nearly 90 kHz, but only for a certain type of reactance. If I can I will be uploading the video of the effect in a few hours...

[@MrAwyork]

What do you see when you run the FFT on the scope when DUT is not shaking? Does it match up to this experiment?

[@galen__]

Subwoofer design implications? 😅