Aargh! Inductor Coupling Question...

Aargh! Inductor Coupling Question...

Some may remember my post about coupling between adjacent inductors some time back.

In that case the inductor cores were face-to-face and I found a coupling equation that matched lab tests, and the customer's product worked.

Now the customer has changed to surface mount inductors facing upwards, so the coupling problem is side-by-side, and the product quit working :-(

Can anyone point me to a coupling equation for side-by-side inductors?

Thanks! ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson
Loading thread data ...

What kind of inductors?

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

Without knowing the inductor construction I don't think this can be done. Maybe you can come close if they're unshielded bobbin cores, but otherwise I think you need to measure the coupling.

Come to think of it, once you do all the fun math, you'll need to measure the coupling ANYWAY, to make sure your assumptions were all correct. So why not just cut to the chase? Get a few boards and a handful of inductors, hand-solder them on one board as close as they'll go, centered on their pads on the second board, and as far away as they'll go on the third. Then measure the coupling to get min, typical, and max.

And for a second come to think of it -- I could be wrong, but if there's cores involved I suspect that you would not be able to get a nice tidy symbolic solution anyway -- it's probably a job for a 3D FEM program, which would just boil down to experimentation with virtual boards rather than the real thing. Moreover, if there are ground planes involved then predicting coupling is going to get even hairier, particularly if you care about any frequency dependencies.

--
www.wescottdesign.com
Reply to
Tim Wescott

I think its worse than that. Unlike Jim's initial end-to-end mounting of inductors, the field outside a solenoid coil, and hence coupling, is going to be low. It would then depend very much on the proximity of magnetic and diamagnetic materials just as much as the accuracy in the inductor's mounting.

I also have no idea of the constructions of these inductors. If off the self there's a fighting chance they will be designed to minimise external fields.

--
Mike Perkins 
Video Solutions Ltd 
 Click to see the full signature
Reply to
Mike Perkins

Yes, that's a good approach

Very true. So, can you convince the customer to spring for two C-cores mounted side-by-side?

Seriously, the internal construction of a surface-mount inductor is going to have significant impact (even the mounting pad design counts) There's little likelihood that a FEM can be constructed without doing experiments on the actual parts. By the way, you DO realize that vertical inductors are likely to have unknown polarity? Swap the pin 1 and pin 2, and the polarity of the coupling changes.

Reply to
whit3rd

Oh that last is a good point. Picture the fields of two dipoles, head to tail the fields are in phase, side by side the fields are 180 out of phase.

JT I don't suppose they can glue a piece of ferrite across the coils?

George H.

Reply to
George Herold

Modulation is amplitude, not phase. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

I thought of that but was vetoed by the customer who is in Hong Kong. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

The coupling equation for side-to-side inductors is exactly the same as for face-to-face inductors, though the mutual coupling coefficient is a whole lot less.

The flux of interest forms a toroid around each coil, and only that bit of the toroid going through the second inductor creates coupling - the off-axis flux has a lot more places to go that by-pass the second toroid than the on- axis flux ever did.

The equations are still the transformer equation

V1 = L1. dI1/dt + M. dI2/dt

V2 = M. dI/dt + L2. dI2/dt

where M is some fraction of the square root of L1.L2 - a smaller fraction for side by side coils than coaxial coils.

Nobody ever seems to teach the transformer equation - even Win Hill couldn't squeeze it into AOE-3 - but I've been pushing it here for years.

Bill Sloman, Sydney

Reply to
bill.sloman

Are they isolated from one another? If they're both referenced to the same ground you can couple them easily enough by running their cold ends to a small inductor, making a 'T'. Presto -- instant coupling.

--
www.wescottdesign.com
Reply to
Tim Wescott

You've never heard of phase modulation?

Reply to
John S

Bill, I must have missed your pushing. Will you please present the transformer equation again?

Thanks

Reply to
John S

There's a whole X-Y array daisy-chained, so that won't work.

I'm toying with, instead of sending 1:0, send 1:0.5, leaving carrier always present, allowing coherent AGC and coherent detection. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

consider PSK or FSK

use a limiter in the receiver, no AGC needed.

m
Reply to
makolber

Why? The two lines of algebra say it all, which is most of the charm.

The voltages have to be alternating, since you can't increase current indef initely without blowing up the coils (or generating a magnetic field big en ough to destroy super-conductivity, which comes to the same thing).

L is just common-or-garden inductance in volts per rate of change of curren t, and M is mutual inductance - the volts in one coil induced by changing t he current in the other coil. The fact that it is the same for any two pair ed coils ought to be obvious, but takes thinking about.

You seem to have to fiddle with inductors for a bit before you can properly take it on board - I must have seen the equation long before it made sense to me, and I had to spell it out to the guy who was my boss at the time. H e had trained as a physicist, and was one of the better electronic engineer s I know, at least in terms of understanding what a circuit did, though I n ever had much trouble changing his circuits to get better performance, whic h did miff him from time to time.

--
Bill Sloman, sydney
Reply to
bill.sloman

While I might agree with you, the real question here is how would you calculate the mutual inductance? What is your magic fraction?

--
Mike Perkins 
Video Solutions Ltd 
 Click to see the full signature
Reply to
Mike Perkins

Oh. I see that it comes from a freshman physics textbook. I though you had something more informative for us.

Well, that is sort of correct in a perfect universe.

Yes, you should do that.

From 1980 to about 1995 I designed magnetics for inverters ranging from

40Hp to 500Hp to drive 3-phase motors. The magnetics included DC filter bus chokes, inductors, and tapped transformers for those inverters. Also, I designed current transformers for use for current sampling of those motors. At one point, my career was on the line. I had to find a solution for a tightly coupled center tapped inductor but not so tight that di/dt would destroy the thyristors. I specified the inductor lead length such that it would add enough series inductance to protect the devices from di/dt. It was the 500Hp inverter which IIRC, had to commutate about 2000A. It worked!

I'm not trying to impress you or anyone else with my 30 year-old magnetics experience. I'm just saying that you could offer more information rather than a freshman physics textbook example. We could benefit from your detailed knowledge. In other words, how would you handle a design for a 10w, 120V to 12V output transformer using EI lams? Please show the M in your so-called equation.

Thanks for your indulgence.

Reply to
John S

Huh... air coil? Then you are losing by the area described by Bill S. You (or some minion at the company) can look up the field from a magnetic dipole and integrate... you'll be getting fields from all the nearby coils too.

George H.

Reply to
George Herold

Grin... I'm a physicist so admittedly biased, but for me, "freshman physics"* is my basis for understanding the world. Of course there is lots more.

George H.

*Fenyman lectures and Berkeley series count.

I though you

Reply to
George Herold

Since Jim hasn't posted any actual dimension - if fact it sounds a bit as i f his customer hasn't told him about them - there's not a lot to calculate with.

My understanding is that once you have the dimensional information you do s ome sort of complicated integration to get the right answer. There are elec tromagnetic field-plotting software packages that will do it for you, and I 've got a copy (reprint) of Frederick W. Grover's "Inductance Calculations" ISBN-13:978-0-486-47440-3 which tubulates a lot of hand-cranked calculation s - it was first published in 1946.

Chapter 18 is on the mutal inductance of circular elements with parallel ax es, which is probably what Jim needs.

Chapter 19 goes on to circular elements whose axes are inclined to one anot her, which may be what he will end up needing, if his customer was vague en ough to get Jim to work things out for coaxial inductors when the board was laid out with inductors side-by side ...

The book does go on a bit, and my experience is that winding and measuring is just as quick as plowing through the tables, and the results are easier to sell to management.

--
Bill Sloman, Sydney
Reply to
bill.sloman

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.