Use Multi-Section Windings to Reduce Self-Capacitance

[Joe Gwinn]

The classic way to greatly reduce the self-capacitance of a deep and wide multi-turn thick coil is to partition the winding into multiple adjacent deep but narrow coils connected in series. The adjacent narrow coils are magnetically coupled the same as when they were one wide coil, but the capacitance of the individual narrow coils is both reduced and in series. The reduction in overall capacitance can be quite dramatic. This method or approach has been around for about a century, as have banked windings.

For example, take a dual-winding ferrite pot core bobbin https://www.norwe.eu/products/en/006/E01.

The coil windings in each section of the bobbin have the same winding depth, but about one half the total winding width. The self-capacitance of each of the two section windings is maybe one half that of what we would have if we used a single-section bobbin https://www.norwe.eu/products/en/006/D01.

Because the two sections are adjacent, the magnetic flux threads the individual wire turns the same as before, so we get the full N^2 effect on self-inductance of the total winding turns N. But the electric fields are different, and each section has about half the self-capacitance of the one-section full-width coil, and because these individual self-capacitances are in series, the total overall self capacitance is about one quarter of that of the full width winding.

Because self-resonant frequency varies as the square root of the product of inductance and capacitance, the resonant frequency of the two-section coil will be double that of the one-section coil.

If one uses three sections, the resonant frequency will be tripled, in theory at least. The limiting factor is the loss of coil windings due to the physical bulk of the separators between the sections.

[Chuck H]

Sort of like the old P bass and G&L pickups?

For some types of pickups, perhaps a good idea. But don't we (guitar players) rather enjoy and take advantage of the resonant peak character of electric guitar pickups as they are? In this event it's a solution looking for a problem. But for acoustic instruments or maybe even bass guitar it could be the bees knees. At the very least it's a good theoretical topic.

[mozz]

What effect does this have on the "Q" of the coil?

[Antigua]

The classic way to greatly reduce the self-capacitance of a deep and wide multi-turn thick coil is to partition the winding into multiple adjacent deep but narrow coils connected in series. The adjacent narrow coils are magnetically coupled the same as when they were one wide coil, but the capacitance of the individual narrow coils is both reduced and in series. The reduction in overall capacitance can be quite dramatic. This method or approach has been around for about a century, as have banked windings.

I take it this works on a similar premise for why the Jazzmaster pickup had a lower capacitance, as seen in this other thread, that it puts more distance between portions of coil that have a higher voltage difference, reducing capacitance that way. So it seems that this is theoretically similar to pancaking Jazzmaster shaped coils in order to end up with a coil that is tall, but still has a low capacitance.

Sort of like the old P bass and G&L pickups?

For some types of pickups, perhaps a good idea. But don't we (guitar players) rather enjoy and take advantage of the resonant peak character of electric guitar pickups as they are? In this event it's a solution looking for a problem. But for acoustic instruments or maybe even bass guitar it could be the bees knees. At the very least it's a good theoretical topic.

I can think of a couple benefits, one would be to cram more turns of wire onto a pickup in order to get a higher voltage output, without also pushing the resonance peak to a low frequency. The other would be to start with a very high peak for a similar output voltage, and then dial it down to taste with a selection of capacitors. A counter point this might be that the winds that are further from the guitar string produce less voltage anyway, and so it might make more sense to avoid a "tall" pickup all together.

What effect does this have on the "Q" of the coil?

It shouldn't make much difference, since the core material and series resistance would be the same.

[David King]

This brings to mind those Fishman coils that are a stack of PCBs each with a spiral etched.

[bbsailor]

This brings to mind those Fishman coils that are a stack of PCBs each with a spiral etched.

David, Joe, and any MEF member

This also brings to my mind the possible advantage of making a CSE187L Triad style current transformer with three sections rather than just the top with the single turn primary and the lower section with the 500 turn secondary.
The three section transformer would have the primary in the center section with two secondary sections with 1000 turns each but tapped at the 700/300 turn point to allow a variety of output tones with a variety of series or parallel output connections. By using a square copper wire equal to the size of the center primary opening area, one could minimize the leakage inductance and have a variety of sounds that would target a typical 150 ohm XLR input with a bridging impedance of 2400 ohms.

Given that the upper and lower secondaries are stacked per Joe Gwinn's post #1, the capacitance would also be reduced, depending on how the secondaries are wired, and could result in some interesting acoustic effects if some current transformer manufacturer would or could make a three layer current transformer with the primary in the center layer.

I believe the guitar community has been locked into the old tube based high impedance input requirements of pickups that has resulted in the typical "electric guitar sound" with the resonance point being in the area where the human ear is most sensitive (about 3 to 5Khz). Anyone choosing to look beyond this "electric guitar sound" needs to consider how low noise, lower impedance inputs allow for a different variety of sound. The electronic theory still equally applies to all of these design variants.

This is just a little audio food for thought.

Joseph J. Rogowski

[Antigua]

This brings to mind those Fishman coils that are a stack of PCBs each with a spiral etched.

That's definitely a low impedance coil that uses active tone shaping to mimic the response curve of a high impedance pickup, so while the pancaked circuit board might result in a lower capacitance, it's probably irrelevant anyway. The capacitance in that case might also depend on the size of the printed traces, which I'd suspect are larger than 42AWG.

[Joe Gwinn]

What effect does this have on the "Q" of the coil?

As Antigua says, probably not a lot of difference.

[Mike Sulzer]

You might not want to do anything that reduces the number of turns you can fit in the available space because you lose output, that is, unless there is something else to gain, and decreasing capacitance is not an important issue because the cable capacitance is much larger than the coil capacitance. Thus the contribution of the coil capacitance to the actual resonant frequency in use is not large.

However, the inductance with the split bobbin is interesting. Since the cores are short, all the flux from one coil does not pass through the other even if the permeability is high. It is not obvious to me whether the inductance changes with a split bobbin, or which way it changes if it does. Suppose it decreased with the split bobbin. Then you could add some more turns and get higher output for the same inductance.

[Joe Gwinn]

You might not want to do anything that reduces the number of turns you can fit in the available space because you lose output, that is, unless there is something else to gain, and decreasing capacitance is not an important issue because the cable capacitance is much larger than the coil capacitance. Thus the contribution of the coil capacitance to the actual resonant frequency in use is not large.

However, the inductance with the split bobbin is interesting. Since the cores are short, all the flux from one coil does not pass through the other even if the permeability is high. It is not obvious to me whether the inductance changes with a split bobbin, or which way it changes if it does. Suppose it decreased with the split bobbin. Then you could add some more turns and get higher output for the same inductance.

One can compensate for reduced winding space by using the next finer wire size. Unless the separators are fairly thick, inductance won't be much affected, so long as the same total number of turns is used. Cable capacitance is of course unaffected, but reducing the self-capacitance of the pickup increases our wiggle room.

[rjb]

What is the base of your assumptions - other than paranoia?

"Paranoia?" Hunh?
Oh my god, they're conspiring to change my inductance!
Quick! Don the tinfoil hats and run to the fallout shelter!

“Just because you're paranoid doesn't mean they aren't after you.”
- Joseph Heller, Catch-22

-rb

[David King]

One could wind the narrow coils sequentially between movable winder plates with just a thin paper or mylar separator. Each layer could even be fused with wax, lacquer or epoxy before winding the next, adjacent layer. Not the fastest way to wind but it would minimize the lost volume/turns. This would allow one to maximize the self-capacitance reduction by experimenting with different coil layer thicknesses and intra-winding dielectric barriers and thicknesses.
For mass production the coil layers could be wound simultaneously using fusible insulation magnet wire on a mandrel with rigid separators, the coils then heat fused, the mandrel slipped out of the assembly and the rigid separators removed and replaced with thinner dielectric material. Finally inserting a magnet or iron core and wrapping the whole thing in tape would tidy it up.

Alternatively one could sell traditional pickups with a low self-capacitance cable, about $20 from Kimber cable...

[ken]

The classic way to greatly reduce the self-capacitance of a deep and wide multi-turn thick coil is to partition the winding into multiple adjacent deep but narrow coils connected in series. The adjacent narrow coils are magnetically coupled the same as when they were one wide coil, but the capacitance of the individual narrow coils is both reduced and in series. The reduction in overall capacitance can be quite dramatic. This method or approach has been around for about a century, as have banked windings.

If you make a Jazzmaster type guitar pickup with six coils to reduce the pickup's capacitance, wouldn't this pickup be impossible to use 'normally' aka in a passive instrument's wiring environment? I think you would have to use a preamp with some strong toneshaping to sound anything like its passive counterpart.

You might not want to do anything that reduces the number of turns you can fit in the available space because you lose output, that is, unless there is something else to gain, and decreasing capacitance is not an important issue because the cable capacitance is much larger than the coil capacitance. Thus the contribution of the coil capacitance to the actual resonant frequency in use is not large.

This is interesting. Would the cable capacitance of whatever the guitar cord is made of make a bigger difference to the sound than the pickup capacitance? I remember some people using coily guitar cords for this reason.

Ken

[David King]

The coily guitar cords were actually among the worst offenders IIRC. If they were popular it's because they killed off some treble and added crackly noises to the mix.

[rjb]

Sorry to be so clear to you, but since we are not talking about Neumann u47 Transformers, but a shitty e-guitar pickup,...

Thanks for finally making your point clear.

There is a difference between saying "you are making much ado about nothing" or even "you are picking fly shit out of the pepper" than saying "you suffer from a mental illness".

-rb