Mutual Inductance

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

ruffrecords

Well-known member
Joined
Nov 10, 2006
Messages
16,155
Location
Norfolk - UK
As part of the work on the Mastering EQ project I have been testing a board that contains a number of separate inductors wired in series.

DSCF1417.jpg


You would therefore anticipate that the inductance between any two point would be equal to the sum of the individual inductors between them. The string starts at the top left with a 6.8mH inductor. I measured the inductor before I put it in and it was close to 7mH (i.e within 5%). I calculated the capacitor for the required frequency (8KHz) fitted it and tested. I found the frequency was lower than it should be. So, I measured the value of the 6.8mH inductor in circuit and the reading I got was 7.8mH which is about 15% out. I recalculated the capacitor for this value, fitted it and the measured frequency was very close to 8KHz.

This was unexpected and my suspicion is that it is caused by coupling (mutual inductance) between the 6.8mH inductor and its neighbours. They are fairly close to each other but I would have expected the coupling between them to be so small as to make no measurable difference. What do you think?

Cheers

Ian
 
Yeah I would have thaught that the seperate inductors should add up linearly ,
maybe offsetting the axis's 90 degrees between each sucessive inductor might make things opperate more as expected ,its one of those situations where the laws of physics seem to bend ,the theory doesnt cover all aspects of reality.Excellent stuff though and thanks for pointing this out , Im working on a chain inductor type arrangement myself so its good to know that predicted results maybe be out of whack from reality by a good margin .
 
ruffrecords said:
This was unexpected and my suspicion is that it is caused by coupling (mutual inductance) between the 6.8mH inductor and its neighbours. They are fairly close to each other but I would have expected the coupling between them to be so small as to make no measurable difference. What do you think?
Indeed, with these inductors where the magnetic circuit is open, as opposed to pot cores, mutual inductance is a factor. I have been alerted to this issue a long time ago when the QC guys told me the passive loudspeaker crossovers did not pass the test. Someone had decided that putting inductors on two layers would make a more compact package. I had to explain him that two inductors constitute a transformer. Later, when I made inductor-based graphic EQ's, I had considered it, but I found it of 2nd-order.
The coupling coefficient in your case seems to be about 0.05, which is enough to alter the values by about 10%.
 
Hi Ian Just wait until you get some ferrous (steel) screws or chassis materials close and the inductance and linearity changes and you may be disappointed. You can insert some Alum or copper foil between or over the inductors to reduce the effects.
Duke
 
Ohhh I see those little brackety do-hickies you were on about now too ,handy stuff.
 
So is what your saying Abbey that the pot cores tend to contain better the inductance and allow less interaction than 'open cores'?
 
Tubetec said:
So is what your saying Abbey that the pot cores tend to contain better the inductance and allow less interaction than 'open cores'?
Absolutely; in a pot core, the flux lines go through the external part of the core. In a I-core, the flux lines return via air.
 
Looks exactly like this image. Except the top-hats make yours couple better.
ind37.gif


Suppose you fill the space between tophats with iron. Closed loop. You now expect 2X the turns to give 4X the inductance.

Winding your own radio coils is good experience. For loading a transmitter tank, or coupling two tanks to get a double-tuned response, you push coils together.
 
A battery of inductors!  I wondered if they would interact, not serially but by being adjacent to unselected ones...

As an aside RR,  I was looking at options for my HPF using some inductors, and then read that it's not good to put them next to Output TX's.  Since I stuck my T1 near the HPF I decided to forgo the inductor method.
 
abbey road d enfer said:
Indeed, with these inductors where the magnetic circuit is open, as opposed to pot cores, mutual inductance is
The coupling coefficient in your case seems to be about 0.05, which is enough to alter the values by about 10%.

I am curious to know how you came p with that figure. CAn you please explain?

Cheers

Ian
 
Tubetec said:
Ohhh I see those little brackety do-hickies you were on about now too ,handy stuff.

Yes they are pretty neat and a standard item for fixing front panels to Eurocards. They area tad expensive being about £1 each but they do hold the PCB firmly.

Cheers

Ian
 
Thank you everyone for the replies. Looks like this will not work as is.  I also notice some small interference spikes in the frequency response plot which I do not get with a mu-metal screened inductor elsewhere in the same design. Looks like I eed another custom Sowter inductor. For a mastering EQ there are no short cuts.

Cheers

Ian
 
those tuned circuits are not exact science, ran into the same problem with the Pultec, which is not a precision instrument to begin with anyway,

caps have a tolerance, so if they are off a bit...

inductance will vary with frequency and level, you can do things to try and get it linear, like use certain core materials, or play with the gap if using pot cores,

playing with turns count might be easier than finding the right cap,

what kind or cores are those?

i bet if you put them on a table ans spread them out that you would get the same results,
 
Hi Ian

would it help if you tried a thin sliver of copper- covered aluminium between each of the inductors to act as a screen - you will probably have to earth it.  Alternatively Mu-metal, but this is pricey 

You can get the self-adhesive copper sheet on Evilbay, but if you can't get any, PM me and I'll drop some in the post.  I use it a lot to screen the inside of equipment cases and it works well.

Worth a play before approaching Mr. Sowter.

Mike
 
from a fair-rite pot core datasheet:

The core configuration provides a high degree of self- shielding.
 
What a timely thread. I’m building two MEQ-5 and there are three inductors in the can, relatively close to one another, and I have concerns about them interacting. I’m using the Carnhill toroidal inductors. Thoughts?
 
AusTex64 said:
What a timely thread. I’m building two MEQ-5 and there are three inductors in the can, relatively close to one another, and I have concerns about them interacting. I’m using the Carnhill toroidal inductors. Thoughts?
Toroids are the ones less prone to interact, since the flux lines are almost totally constrained within the core (typically about 99.97%). I've found that pot cores of the RM series (which are not the most constrained flux-wise) have negligible interaction when spaced only 0.1" apart. Only EI and open cores need to be taken into consideration in that regard.
 

Latest posts

Back
Top