Use Multi-Section Windings to Reduce Self-Capacitance

[Joe Gwinn]

It sounds like you're saying that it's because of the core that in increase in inductance per turn is not linear, but it's not linear even if the core happens to be air.

In pickups, the core has a very low effective permeability, because most of the magnetic path is in air (or non-magnetic materials).

So, imagine a pickup coil in air. Visualize what happens to the arrangement of single turns in space if I split the pickup coil into two parallel sections. if the separation between the two parallel half coils is small, so will the effect be small.

[Joe Gwinn]

Edit: I see what you are saying. Sorry to be so slow. First consider this (theoretical, not practical): wind the turns on the coil individually, leaving each one unconnected, that is, just with the two ends of each loop sitting close to each other. Now consider different was of connecting them in series, presumably using twisted pairs that have zero resistance and occupy no space. Then no matter what order we connect them in all the flux generated by each turn passes through each other turn in an identical way. I am not quite sure this represents the actual case closely enough, but it probably does.

Yes. My instinct is that the effect on inductance, while real, is small, and easily compensated for should it be necessary.

[Helmholtz]

My instinct is that the effect on inductance, while real, is small, and easily compensated for should it be necessary.

To me a reduced inductance with the same output means a benefit that goes together well with the reduced capacitance. It would allow for a greater degree of freedom to design the PUs frequency response without sacrificing output. And IIUC, Mike Sulzer thinks the same.

[Joe Gwinn]

To me a reduced inductance with the same output means a benefit that goes together well with the reduced capacitance. It would allow for a greater degree of freedom to design the PUs frequency response without sacrificing output. And IIUC, Mike Sulzer thinks the same.

Could be. I'm guessing someone is making a two-section singlecoil as a test.

[Antigua]

I don't know why I didn't think of this sooner, but the Fender Noiseless pickup also demonstrates the lower capacitance that is to be had with two coaxial coils. Lo and behold, around 35pF, just like the Jazzmaster pickups. The parenthesis represents measured capacitance minus probe capacitance.

Fender Noiseless Bridge
DC Resistance (series): 10.27K
top: 5.14K
bot: 5.18K
Inductance (series): 2.592 H
top: 1.445H
bot: 1.591H
Resonant Peak: 13.7kHz kHz
Calculated C: 32pF (52-20)
Coil width: top: 0.6090" bottom: 0.5870"

Fender Noiseless Middle
DC Resistance: 10.00K
top: 5.05K
bot: 4.95K
Inductance: 2.623 H
top: 1.470H
bot: 1.612H
Resonant Peak: 13.3 kHz
Calculated C: 35pF (55-20)
Coil width: top: 0.5985" bottom: 0.5850"

Fender Noiseless Neck
DC Resistance: 10.19K
top: 5.10K
bot: 5.09K
Inductance: 2.555 H
top: 1.398H
bot: 1.588H
Resonant Peak: 13.7 kHz
Calculated C: 33pF (53-20)
Coil width: top: 0.5954" bottom: .5880"



The inductances of 2.6H result in a loaded peak f of 4.1kHz, with a test load of 470pF, where as it would usually take a Strat pickup of 2.2H to get a loaded peak f that high, so knocking about 100pF off an equivalent single coil allows a few more turns of wire to be placed on the coil(s) before the peak f drops into the range of a "hot" pickup that is less suitable for clean and clear guitar tones. On the other hand 2.2H and 2.6H Strat pickups are barely any different in terms of voltage output, and I think the only real solution there is to get more turns of wire closer to the guitar strings, regardless of how many turns of wire you ultimately have. That would best be accomplished with a coil that is shorter and wider, rather than thin and tall. The Bill Lawrence Microcoil represented a short wide pickup that would still manage to fit in a Strat pick guard, but I've never had one to analyze.

I had already fashioned a two sectioned Strat pickup to make a test pickup for measurement, but now I can make Bill Lawrence Microcoil style pickup out of it instead, and see what happens with one of those.

[Helmholtz]

I am not familiar with the construction details of Fender noiseless PUs. But I guess that series wiring means that the coils are wired electrically out-of-phase, right? I am asking, because the series inductances are lower than the sum of the individual values. The difference must be caused by (weak) magnetic coupling. Is it possible to invert one coil for series L measurement?

[rjb]

I guess that series wiring means that the coils are wired electrically out-of-phase, right? I am asking, because the series inductances are lower than the sum of the individual values. The difference must be caused by (weak) magnetic coupling.

True but possibly irrelevant:
The series inductance of a "standard" humbucker pickup is greater than the sum of the individual coil values.
The series inductance of a "sidewinder" pickup is less than the sum of the individual coil values.
Mumble, mumble, something about geometry and magnetic coupling.

-rb

[Antigua]

I am not familiar with the construction details of Fender noiseless PUs. But I guess that series wiring means that the coils are wired electrically out-of-phase, right? I am asking, because the series inductances are lower than the sum of the individual values. The difference must be caused by (weak) magnetic coupling. Is it possible to invert one coil for series L measurement?

That's a good observation. One half is 1.4H, the other is 1.6H, and is series it measures 2.6H, which is less than half, so I think it's fairly reasonable to assume that with the coils in series the inductance would be somewhat in excess of 3.0H. Suppose the combined inductance is 3.1H, the resonant peak of 13.7kHz then calculates out a capacitance of 43.5pF.

[Helmholtz]

True but possibly irrelevant:
The series inductance of a "standard" humbucker pickup is greater than the sum of the individual coil values.
The series inductance of a "sidewinder" pickup is less than the sum of the individual coil values.
Mumble, mumble, something about geometry and magnetic coupling.

-rb

Yes, Gibson type humbuckers have an inductance about 20% greater than the sum of the individual coil values. The reason for the deviation is the mutual inductance.
Mutual inductance is another way to descibe magnetic coupling and it is proportional to the square root of the product of the individual inductances. A square root always has two solutions, a positive and a negative one. The correct sign depends on the relative orientation of the self-induced fields of the individual coils in series configuration. I.e. the serial inductance Ls can be either Ls=L1+L2+2M or Ls=L1+L2-2M (M meaning the absolute value of the mutual inductance), depending on the relative electrical phase of the two coils.
No coupling means M=0. 100% coupling of two coils both having the same inductance L results in either Ls=4L or Ls=0.

[Chuck H]

Yes, Gibson type humbuckers have an inductance about 20% greater than the sum of the individual coil values. The reason for the deviation is the mutual inductance.
Mutual inductance is another way to descibe magnetic coupling and it is proportional to the square root of the product of the individual inductances. A square root always has two solutions, a positive and a negative one. The correct sign depends on the relative orientation of the self-induced fields of the individual coils in series configuration. I.e. the serial inductance Ls can be either Ls=L1+L2+2M or Ls=L1+L2-2M (M meaning the absolute value of the mutual inductance), depending on the relative electrical phase of the two coils.
No coupling means M=0. 100% coupling of two coils both having the same inductance L results in either Ls=4L or Ls=0.

But isn't mutual inductance somewhat relative to proximity? I've experimented with series coils separated by a shield with significant alterations to the effect (audibly, not tested for finite figures). So the particulars of configuration, spacing or additional inter device shielding can also play a role. I think.?.

[Antigua]

It makes sense that a PAF's coils mutually increase the inductance, because the coils are wired out of phase, so each is in the magnetic return path of the other, much like the cross section of a toroid inductor. I suspect that if the two coils were in phase, the overall inductance would drop.

[Helmholtz]

But isn't mutual inductance somewhat relative to proximity? I've experimented with series coils separated by a shield with significant alterations to the effect (audibly, not tested for finite figures). So the particulars of configuration, spacing or additional inter device shielding can also play a role. I think.?.

Yes, of course mutual inductance M12 also depends on distances, arrangement of coils, relative directions of current flows and materials that change flux distribution (cores, shielding), in other words on coupling. The complete formula is M12= M21= +/- k* (L1*L2)^0.5, where k is the coupling factor ranging between 0 and 1.
The ab initio computing of k can get rather complex, but M12 is easily calculated from L measurements and the formulae given above.

But I just wanted to explain the increase/decrease thing in series configuration.

[Joe Gwinn]

Abstraction to essentials of series-coil inductance

Let us reduce this problem to the barest of essentials, a two-turn coil, where the distance between the two turns is varies. The turns are connected together using twisted-pair wire, so that only the two turns have significant magnetic fields. The turns are parallel to one another, for maximum magnetic coupling.

If the two turns are adjacent to one another, then the magnetic field of each turn is fully felt by the other turn, and their mutual inductance is the same as the self-inductance of either turn alone.

If the two turns are connected in series-aiding order, the total inductance is L1 + L2 + M12 + M21, or four times L1, because L1=L2=M12=M21.

If the two turns are connected in series-opposing order, the total inductance is L1 + L2 - M12 - M21, or zero, because L1=L2=M12=M21.

As the distance between the two turns increases, M12 and M21 (while still equal) both decline, to negligible at say ten coil diameters, and to zero at infinity.

If one repeats this exercise, but with three one-turn coils, the factor of four will become nine, and so on.

[Helmholtz]

Yes, I already gave these extreme examples in post #129. Perfect coupling between two turns, meaning k=1, probably requires them to be wound on a high µ toroidal core as two turns cannot be in exactly the same place. Otherwise the inductance in "constructive" arrangement will rise somewhat lower than N^2.

[Joe Gwinn]

Yes, I already gave these extreme examples in post #129. Perfect coupling between two turns, meaning k=1, probably requires them to be wound on a high µ toroidal core as two turns cannot be in exactly the same place. Otherwise the inductance in "constructive" arrangement will rise somewhat lower than N^2.

What was missing was the visual picture and qualitative physics.