mutual inductance

I have a question about mutual inductance of two choke coils.

Say I have two 15H coils stacked right on top of one another.  Does the effective inductance of each increase or decrease?

They should remain unchanged.

Cheers

Ian

If the two coils were in two consecutive stages of LC filtering in a power supply, would there be any disadvantage to orienting them in such close proximity?  I know the Neumann NG power supplies had a dual choke wound around one core. Just considering layout for a power supply project, and determining the size of the enclosure I'd need.  If I could stack these two chokes one on top of the other, it would save a lot of space, but the mutual inductance (a concept I don't have a deep understanding of) gives me pause.

But, do your inductors have a common core, or rather a common magnetic circuit?

If they are two regular chokes then their magnetic circuits are almost completely separate and there should be no problem stacking them one on top of the other.

Cheers

Ian

cool. thanks

ruffrecords said:

But, do your inductors have a common core, or rather a common magnetic circuit?

If they are two regular chokes then their magnetic circuits are almost completely separate and there should be no problem stacking them one on top of the other.

Cheers

Ian

Click to expand...

It depends very much on the core arrangement. Single-I cored inductances can produce large mutual inductance. Standard E-I cores exhibit much less, but still, the effects are measurable.
And it can be easily shown that having mutual inductance between inductances in a PSU filter arrangement is generally detrimential. See attached figure, which shows serious degradation of hum rejection. Both are 40uf/8H/40uF/8h/40uF in double-pi arrangement (as per Neumann NG), green without mutual inductance, red with.
In that respect, I don't understand the Neumann arrangement.
In this recent post, you can see a particular arrangement where mutual inductance is used profitably.
http://www.groupdiy.com/index.php?topic=42454.msg527809#msg527809

abbey road d enfer said:

It depends very much on the core arrangement. Single-I cored inductances can produce large mutual inductance. Standard E-I cores exhibit much less, but still, the effects are measurable.

Click to expand...

That's why I said 'if they are regular chokes' which in the vast majority will be EI cores. Since the flux path is almost entirely contained within such cores then mutual inductance between a pair of them must be very small indeed. With I cores I agree 100% with you.

And it can be easily shown that having mutual inductance between inductances in a PSU filter arrangement is generally detrimential. See attached figure, which shows serious degradation of hum rejection. Both are 40uf/8H/40uF/8h/40uF in double-pi arrangement (as per Neumann NG), green without mutual inductance, red with.
In that respect, I don't understand the Neumann arrangement.

Click to expand...

Have you tried. the simulation with the polarity of one choke reversed? What was the assumed coupling factor?

Cheers

Ian

Yes I have tried with polarity reversed; it's marginally less bad (I can't write "better").
I used 100% coupling as a start.
Indeed, lowering coupling factor ultimately turns the system in a standard arrangement (with two independant coils), so as coupling factor decreases, performance gets better.
The problem is the performance is better with one single non-tapped inductor.
See attached:
Red: double-Pi no mutual
Green: with mutual (k=0.9) in-phase
Cyan: with mutual (k=0.9) anti-phase
Magenta: single Pi (40uF 32H coil 80uF for identical BOM)

So the bottom line is, no surprise, it all depends on the coupling factor. So, the big question is just what value would that be for a pair of physically separate stacked EI cored inductors.

Put another way, at what value of K does the coupling not make much difference. Could you check this out on your sim?

Cheers

ian

Pls see attached. The red graph is with k = 0.1, which is what you have with EI cores put too close together.

Great, and helpful discussion. Did you omit the attachment on your last post?

Yes I did...  :-[

Thanks!!


by the way, this would be two EI chokes, 15H apiece.

Just so I understand right-- in your professional opinion, should I avoid mounting them vertically, one on top of another?  Would that be worse than mounting them side-by-side, fairly close together but not touching?

> choke coils

Covers a LOT of types, from iron shell-core to 4 turns around a soda-straw.

> this would be two EI chokes

Jam them in how ever fits.

Except right at a gap, the field is 99.9+ in the iron, does not go through air.

EI *chokes* are usually gapped; don't butt the gaps against each other (this would be difficult to do). A gap-thick space improves things a lot.

At the school we had drummers and we had harpsichordists. They needed practice rooms. Imagine a room 99.9% sound-tight, 0.1% through the door-crack. If two doors are spaced along a hallway, further than a crack-width, leakage is teeny. If we ram the doors together (there are modular practice rooms which can be moved) crack-to-crack, leakage will be higher.

JAMMING an E against another EI's gap is also bad, bridges the gap. It is nearly impossible to get enough contact/proximity to matter, you'd need a hammer.

The EI is critical. When you move to loudspeaker chokes, they are often air-core or stick-core. These will mutual very significantly. You see aircore crossover chokes mounted well apart and at right angles.

> Does the effective inductance of each increase or decrease?

Of "each" is often not as important as what the coupled signal does. In a L-C-L-C power filter, any coupling from first to second throws large ripple through to the output. In a stereo tone-filter, the bleed from one channel to the other may show as "poor" stereo separation.

Can't you breadboard? Mock it up spread-out. Test. Then flip the one choke over the other, test again. Actual trial and success/disappointment may tell and teach you more, and more memorably, than innernet chatrooms.

thanks. I'm kind of mocking up a layout to determine which enclosure I'll want/need, before even breadboarding.