I need to get rid of some differential inverter noise in a small space.

That's called a common mode choke. :^)

If you truly need differential mode, use a much lower inductance and augment it with big fat capacitors on either side.

Tim

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

I am having a problem designing a differential mode choke for an inverter. I would like to put a high permeability core on the dc link leg. Can I push a dc current opposite to counteract the large dc flux created. I need about 100 amps. I was thinking to put 100 turns on the coil and put an amp through it.

I am interested in papers deal with this approach. If you know please forward info. I am having a problem with verbiage that Google will use for a good search.

I don't like blindly asking others to go out of their way to do my research so I will tell you that I saw a good movie last night with Ben Afflack called _The_Accountant_ Go watch it.

Thanks Bob N9NEO

It's fundamentally difficult to add a DC winding to cancel DC flux; where do you get the current from? I have seen it done with permanent magnets.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

No, or at least that is not how I understood it. He wants an inductor (diff mode) to get rid of noise, but wants to balance the large DC field (from 100 A) with 100 turns of one amp current going the other way. This seems sensible to me..(not a magnetics guy.) I mostly wonder what such a beast is called, I figured you'd be able to tell him.. or tell us both that it's impossible for some reason.

George H.

r. I would like to put a high permeability core on the dc link leg. Can I push a dc current opposite to counteract the large dc flux created. I ne ed about 100 amps. I was thinking to put 100 turns on the coil and put an amp through it.

why is that btw? The one issue I can think of is power dissipation, if supp lying unipolar pulses you'd need a lot of amp-turns to zero the dc componen t. I'm probably missing something.

NT

The DC bias adjustment with a secondary winding is common only in magnetic-amplifier circuits (which are not really common). It does make sense, but takes a second power supply and (if the current you're compensating shifts) some slow feedback.

In tube days, the filaments took a constant current that could be diverted for this purpose.

Oversize cores would be the solution, except this is 100A? Oversizing for that would get... costly.

Because true DC creates a standing flux field, not a moving one. An inline choke is gonna arrive a bit late for the show.

If there's AC signal in the core too, it will be coupled into the added current-bucking winding. The power supply will short that out. The fix is then to add a big inductor in series with that, which pretty much gets you back where you started.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

I know, I understood it that way, too. But where do you get the DC from? It has to be a high compliance current source. Otherwise, it runs out of compliance and stops filtering. It also has to track accurately.

The only current that's guaranteed to track is the return current, but as you note, that's wired the wrong way for diff action. :)

Note that it has to track, but not so quickly as to completely oppose the current (attempting to be) drawn. In other words, it has to be lowpass filtered first. In which case, you're really making a transformer-coupled gyrator, and you should use a capacitor multiplier or hum bucker instead. Which, indeed, has been used from time to time, so there you go? :)

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
Website: https://www.seventransistorlabs.com/

You could try wrapping a couple turns around a powdered iron toroid made to take the current and then add some capacitors like Tim mentioned. Powdered iron probably won't remove high frequencies (above ~ 1 MHz ?) a whole lot although you could try some newer alloys used for DC inductors. Mag-Inc maybe ?

I usually use high permeability high-loss (at frequencies of interest) common mode chokes for EMI reduction as well as differential and CM capacitors. Fair-Rite #31 is pretty wide band.

I've contemplated about adding multiple turns and low current to counteract volt-seconds to reduce saturation in an EMI reducint diff mode inductor but have never tried it. It might just work.

boB

rter. I would like to put a high permeability core on the dc link leg. C an I push a dc current opposite to counteract the large dc flux created. I need about 100 amps. I was thinking to put 100 turns on the coil and put an amp through it.

upplying unipolar pulses you'd need a lot of amp-turns to zero the dc compo nent. I'm probably missing something.

right

I'm not seeing how. Surely it gets you some injected dc to cancel the origi nal dc offset.

NT

Draw the equivalent circuit:

Power wot to be filtered --> primary of transformer Secondary of transformer --> filter inductor --> bias supply

The bias supply is CV so it reduces to zero (a supernode). This puts the inductor in parallel with the transformer (an ideal transformer has infinite magnetizing inductance). The inductor reflects through the transformer as equivalent primary inductance.

Ergo: Power wot to be filtered --> filter inductor

The filter inductor must handle _precisely 100% of the DC magnetization_ that a proper inductor in the traditional case would've needed. Except you've wasted a lot of runabout stuffing a fat extra transformer inbetween...

If you replace the inductor with a linear equivalent (usually a gyrator type circuit), you're just going about the swap-V-and-I equivalent of a CV shunt regulator circuit. Which you might as well do straightaway in the first place, because it doesn't require any transformers and shows up in all the textbooks. :-)

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
Website: https://www.seventransistorlabs.com/

I don't know what 'Power wot to be filtered' means or what 'wot' means. I can see that it's wasteful on cores, but it does work as far as I can see.

NT

Nah, no problem, it's just the inductive version of the Kanner Kap. A

100-turn winding can easily be driven from the 100-amp side using sense resistors with the same ratio as the transformer turns, plus a power amp connected to null out the voltage difference.

Of course this scheme has compliance limits, but so does a cap multiplier, and you don't have to figure out how to get rid of 80 watts worth of wasted power (0.8V x 100A).

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Post a schematic so it's clear what we're talking about.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

nverter. I would like to put a high permeability core on the dc link leg. Can I push a dc current opposite to counteract the large dc flux created. I need about 100 amps. I was thinking to put 100 turns on the coil and p ut an amp through it.

f supplying unipolar pulses you'd need a lot of amp-turns to zero the dc co mponent. I'm probably missing something.

iginal dc offset.

I guess it (what I'm thinking of) could be drawn as some gapped transformer . (Gapped so there is a well defined inductance.) One side has N turns and the other 100*N turns. The large 100 A current flows through the low turn side and 1 Amp through the other. (current flow such that DC B fields cancel.) I was thinking I needed a low noise 1A current. But after Phil's post I'm not sure that's true... even a current derived from dividing down the noisy 100A should be fine. (The high freq. noise doesn't matter.. as long as things are roughly linear.)

What do you need a schematic of? How I'd make the 1 A current? (It'd be a brute farce circuit anyway.. :^)

George H.

inverter. I would like to put a high permeability core on the dc link le g. Can I push a dc current opposite to counteract the large dc flux create d. I need about 100 amps. I was thinking to put 100 turns on the coil and put an amp through it.

x;

if supplying unipolar pulses you'd need a lot of amp-turns to zero the dc component. I'm probably missing something.

original dc offset.

er.

If a closed circuit LEM sensor is used to monitor the current it will provi de the correct balancing current. Put the secondary turns on EMI choke in s eries with the secondary on the LEM core.

an inverter. I would like to put a high permeability core on the dc link leg. Can I push a dc current opposite to counteract the large dc flux crea ted. I need about 100 amps. I was thinking to put 100 turns on the coil a nd put an amp through it.

lux;

n, if supplying unipolar pulses you'd need a lot of amp-turns to zero the d c component. I'm probably missing something.

t.

e original dc offset.

rmer.

d

vide the correct balancing current. Put the secondary turns on EMI choke in series with the secondary on the LEM core.

Not sure what LEM is (lunar escape module :^) But for a crazy idea you could put a B field sensor in the gap of the inductor and servo the current to give zero B field.

George H.

snip

Lunar Excursion Module.

[Snip!]

LEM is a Swiss brand of very nice current transformers. I suppose you transpondians would buy them from Pearson.

Jeroen Belleman