interesting inductor

On a sunny day (Sat, 17 Aug 2019 09:31:05 -0700) it happened snipped-for-privacy@highlandsniptechnology.com wrote in :

I did that once for tuning a 30 kHz to 150 kHz or there about inductive loop driver, Watch your voltages when the relays switch, I wound the coils on potcores myself.

Schaffner is a very old company, already used their stuff in the sixties.

But it would be saturated, surely?

Cheers

I will have a uP controlling everything, so I can kill the source when the relays switch. Cold switch.

Yuk. I hate to wind coils.

Sometimes I have to.

Toroids are worse. I'm thinking that the Schaffner part is a ferrite toroid.

Stock at Digikey.

We'll have to find out. In a common-mode filter, I guess that they assume that the line currents cancel in the core.

dot L1------/////////-------> load

L2------/////////-------< dot

At what saturation current? Q? Tolerance?

Good luck keeping 2.5mH at that current in anti-series.

The 2.5mH alone varies widely with ACV and manufacture. Buy a dozen from different distributors and see if they measure the same.

Ungapped ferrite is the Z5U of magnetics.

CMCs typically saturate in the 10s of mA.

I don't have one of these handy to measure. I have looked at the family before, and they have reasonable cost, and attenuation for EMC purposes, particularly at higher frequencies.

Yes, CMCs are actually transformers, not coupled inductors. Their leakage may not be great, but the isolation ratings are exemplary, making them tempting for such purposes as DC-DC converters. It's hard to get more than a watt or so through them, however.

Tim

I might want to output as much as 3 amps RMS. This will be a 3-phase AC source, actually a PM alternator simulator with programmable resistance and inductance. High inductances will be DSP simulated, but I want some real L to be there for high frequencies. Maybe just one real inductor would do. It's a messy simulation.

That's the hazard. We'll have to measure it. If this won't work, we'll have to buy a bunch of klunky single inductors.

It's only about 15 mohms per winding, so there can't be many turns.

This one should be good for lots of watts. Leakage L is only 17 uH!

On a sunny day (Sat, 17 Aug 2019 10:23:30 -0700) it happened snipped-for-privacy@highlandsniptechnology.com wrote in :

OK, nice,

No problem here, for fine wire I put it on the dremel, a thousand turns is no problem. Counting is.

That circuit, I was thinking about that, is basically the same as a TV horizontal output no? +12V | ____ ||| (_____ defection coill )|| ------ )|| ( HV coil ||| (____ |----------- c | |

-- b --- | e / \ === | --- | 'boost' cap | | | /// /// ///

----- -------- base drive, | | --

That gives a parabolic voltage pulse during flyback, the HV coil is tuned to IIRC (was a discussion about that here long time ago) the third harmonic. That pulse is the output you want, only you have a smaller coil, resulting in a higher frequency, The energy you store into the coil while the transistor is 'on' increases the longer it is on, and the higher then the pulse is. I have wound my own TV transformers both for tubes and transistors.

Never understood what the problem was with your coil. :-)

I dunno, I used their thyristor driver pulse transformers, also 4 kV isolation tested. Nice stuff.

Yes, that's the point.

Would be alright for some watts as a resonant transformer, but you still need secondary side regulation for each output channel; no improvement over a bunch of discrete DC-DC modules.

Tim

My Dremel won't wind toroids! There is one trick for hand winding toroids, where you make an H-shaped wire storage thing out of FR4 and keep passing it through the hole, doling out wire. That helps a little.

No, not the same, more of a non-resonant HV pulse generator.

I kept frying the inductors, two generations of Coilcraft parts. I was well within the RMS current ratings, but skin effect apparently dominated and the coils smoked.

The clever horizontal output thing was originally tubes of course, then bipolar transistors. I discovered that the c-b junction of one high voltage horizontal-output transistor accidentally had the ideal doping profile to be a kilovolt Grehkov drift step-recovery diode. Too bad they don't make it any more.

That's because the series connection is the line current, so you only have differential inductance, probably 100 times lower than the common mode inductance of 1.8mH

The common mode inductance of 1.8mH saturates in the mA region

Cheers

Klaus

A thing that size should store a few mJ.

A few uJ at best.

Look, I get it, you don't work with these things on a daily basis. But you are still aware that saturation is a thing, right..?

Tim

This is probably about the same size core as what's inside the CM choke:

That's 1 mH, rated 2.4 amps, stores about 1.7 mJ at rated current.

You can store a few uJ in a really tiny inductor.

So what, you think they put the same core in that CMC?

Tim

Toroid winding machines are cute. They depend on a split bobbin (two semicircular halves) which you put together so it goes through the hole in the torpid. You then wind the length of wire you need onto the bobbin, then wind it off again onto the toroid.

Elegant, but a pig to automate.

Ferrites store similar energy by volume.

Millijoules, not microjoules.

You snipped your "few microjoules" claim. We'll just forget you ever said that.

When gapped, yes.

A mu_eff about 1000 times higher than the powdered iron core is both very likely for the choke, and fully accounts for the 1000 times lower energy rating.

Capacitors go as epsilon E^2 / 2, but high-k dielectrics have considerably lower Emax so store about the same energy (give or take how much you want to saturate them in the process).

Inductors go as B^2 / (2 mu), so a mu=10k core stores f*ck-all energy at saturation.

Tim

The estimate of a few mA to saturate this core corresponds to ballpark

On a sunny day (Sat, 17 Aug 2019 21:17:26 -0700) it happened snipped-for-privacy@highlandsniptechnology.com wrote in :

John, Tim may have a good point, long ago I dissasembled (bought a load for 1 dollar or so on ebay) one of those coils,