Inner stray capacitance of passive mag transducers. A few thoughts.

[freefrog]

More than two years ago, I was writing in another topic that if the stray capacitance of a passive magnetic pickup often ends drowned in overall capacitance (of pickup + cable used to plug the guitar), a low coil capacitance is not totally negligible IMHO/IME when volume/tone controls are lowered...

Better late than never. This topic is meant to explain my past statement. It’s a simple & humble attempt to share something checkable by anyone through personal experience or experiments. Please, keep in mind that English is still not my mother tongue and decipher my contribution accordingly.

Below are some simulations of how a Fender style single coil behaves when its volume control is lowered from 100% to 10% of its value.
The 5spice model used in this case is the simplest possible: (voltage source > inductor> resistance) // with a cap to ground to emulate stray capacitance. Specs selected: 6k, 2.96H - strange numbers, I know, but those were the values of a prototype whose actual Rz was checked with 5Spice. Its specs have simply been kept to do the following sims.
The pots modeled will be two 250k controls (a volume + a tone pot fitted with a 22nF cap). The input impedance, 1M.
I subjectively tend to model pickups as they would resonate without “integrator” circuit but I’ve opted here for a flat response in the name of clarity.

Here are the curves obtained for a virtual 120pF coil + pots plugged through a 380pF cable. Total stray capacitance: 500pF, therefore. The left vertical line is roughly where the resonant peak is placed with both pots full up. The right vertical line is approximatively where the corner / cutoff frequency is placed when the volume pot is @ 10%.



Now, here are the curves simulating a 180pF single coil plugged through a 320pF cable. The overall stray capacitance is still of 500pF and the resonant peak with both controls full up (in red) is exactly @ the same place than in the previous pic… but the repartition of frequencies according to the output level is starting to change: when the volume is @ 10%, the corner frequency is 500hz lower than with the 120pF coil and the Q factor is no more the same.



Below are now the frequencies of the same 180pF single coil through a 320pF cable but the tone control has been relocated after the volume pot, according to the recipe of “50s wiring”. The corner frequency shifts up for the lowest volume setting, reaching more or less 6300hz.



And here are the results when the pickup is seen again as a 120pF single coil through a 380pF cable, still with 50s wiring. This time the corner frequency @ the lowest volume setting is of almost 7500hz: more than twice the resonant frequency when the volume pot is full up.



Temporary conclusion: coil capacitance makes no difference when pots are full up. But when a volume pot is lowered, a coil with lower inner parasitic capacitance gives an extended high range + a lower Q factor.

For me, it potentially explains why scatter wound pickups are often perceived as different by players who fiddle a lot with their volume and tone controls (as I do myself). YMMV. :-)

[freefrog]

Now and even volume pots aside... Things get more complex but also more interesting with humbuckers when it comes to inner stray capacitance.

To illustrate this, I had the intention to use a more refined pickup model based on a transformer because it would have allowed to take in account magnetic interaction between coils. But in the name of consistency, I’ll start by simply doubling the previous 5spice schematic : two voltage sources, two inductors, two resistors, two parallel caps (one per coil). The specs chosen will be those of a typical P.A.F.: 4k and 2.2H per coil. Pots: 500k controls, in 50s wiring. 1M input impedance.

Let’s vary the stray capacitance from 100pF to 140pF for one coil and conversely for the other coil (I’ll comment these capacitive values later below).

With all 500k controls full up, here is how the response of virtual coil 1 changes when coil capacitances are modified.



And here is the response of virtual coil 2 as affected by the same capacitance variation.



Here is finally the response of both virtual coils in series when they are modeled as communicating vessels capacitively:



The two coils appear to correct each other in this last case of series operation but in my understanding & experience, it’s partly an experimental artefact: in a guitar, each coil “hears” the strings from another place than its counterpart (with the effect on harmonics modelized by Tillman) and one coil is often louder than the other for various reasons that I won’t dig here - magnetic circuit more efficient for the slugs coil of a Gibson style HB, height settings of the screw poles, angle of the pickup under the strings, possible presence of a cover with more or less effects due to Foucault currents and so on.

So, IME, the mutual connection & correction of coils rarely cancels their respective responses - even if they have an obvious influence on each others and if this influence is not only capacitive.

IOW, their actual sound never follows IME a unified curve like those obtained with an exciter vertically positioned in the middle of a humbucker. This statement is easy to check experimentally: the only thing to do is to reverse the slugs and screws coils of a Gibson style HB under the strings then to listen… https://youtu.be/SuM-a3RpEKM?t=100

That being said and to come back now to the capacitive values involved in the model above: in a humbucker more than with a single coil, there’s the stray capacitance exhibited by coil wire… AND the capacitance of the hook up cable, which is either coaxial either made of 4 conductors + bare ground wire in most cases.

In 1ft of hook up 4 conductors cable, one pair of conductors can exhibit a stray capacitance of 100pF while another pairing will measure 60pF, for instance. It depends on how they’re connected as shown by Dr Lawing @ Zexcoil (with other capacitive measurements than the ones that I mention): https://www.thegearpage.net/board/in...uency.1162900/

IOW, a coil measuring individually 40pF can reach 100pF OR 140pF, depending on the added capacitance of the conductor connected to it. Hence the capacitive values that I’ve simulated.

This property of 4 conductors hook up cables, added to the stray capacitance of each coil contributes to generate Russian mountains responses like with the coils of this high gain blades HB (whose crude Rz are here shown on a linear scale; black line = one coil, pink line = the other coil, the pickup having a 4 conductors cable):



And such variations beyond the main Rz affect the tone IME…

Below is the measured response beyond main resonant frequency of the aforementioned humbucker once played (in orange) vs the response of EACH of its coils, measured in the same way than above - IOW, separately excited by an ultra low impedance air coil laid horizontally:



And here is the upper harmonic response of another HB whose peaks and dips caused by capacitance are located differently (and not far from my simulation of a 4.4H / 8k HB above in the first screenshots; it's never frustrating when actual measurements don't disagree with sims):



To obtain the orange lines, these pickups have been played direct to the board through a 1M input. Their responses would be much brighter through an amp…

In the first case, nevertheless, the scooped upper harmonics of the high gain HB contribute to make it sound congested and not open enough.

In the second case, the peak around 12khz causes a perceptible drone effect, potentially irritating through a cab with wide frequency response and/or with a bright switch enabled.


Nothing in common with the nice response below of a vintage PU selected as reference and whose coils/hook up cable capacitance hasn't the detrimental effects commented above... its harmonic spectrum is more open / even and more satisfying musically:



Fortunately, there’s various ways to correct peaks and dips due to parasitic capacitance when they happen...

More about this later if time permits, maybe. ;-)

[freefrog]

In the meantime, here are some last pics involving some stacked HB's...

Response of one of these pickups when played in mid position VS when excited by an air coil (without integrator, that I don't use personally since resonant frequencies affect only harmonics anyway):



The "up turned nose" in the upper harmonics noticed in this case was really unpleasing through any clean amp with a bit of sparkle.

Here is below how it looked initially with the same kind of pickups played in neck and bridge positions (the red arrows showing their annoying "drone" sounding harmonics) then once the undesirable effect of coils coupling corrected:



Oh, and here are below on a linear scale the electrically induced responses of the stacked HB mentioned above in its "not corrected" version (in black) then in one of its "corrected" versions (in red), compared to those of a Vintage Strat pickup (in pink) and, in green, of a "half-corrected" stacked HB (whose magnetic circuit hosts an added part meant to make the response flatter)... These tests have been done @ different moments with slightly different capacitive/resistive overall loads but at least, it gives an idea of what happens or not, in relation with capacitance in the pickup itself & its hook up wires.



FWIW and, once again: more later if time permits, maybe.

[Helmholtz]

What is/are
- Rz,
- your vertical scales
- corrected vs. not corrected plots,
- your PU models/sim circuits,
- the coupling factor between HB coils in your sims ?

[freefrog]

What is/are
- Rz,
- your vertical scales
- corrected vs. not corrected plots,
- your PU models/sim circuits,
- the coupling factor between HB coils in your sims ?

Hello,
-Rz = resonance (resonant frequency);
-the first screenshots used have been intentionally trimmed for contextual reasons but scales were the same for the pictures that I compare to each others;
-the PU model used in this case is the simplest possible, mentioned in the two first posts: (Voltage source > inductor > resistance) // capacitor to ground for each coil;
-the related 5spice sims don't take coil coupling in account and that's why I've stated in the 2d post my initial intention to use another model based on transformers (model with which I can vary the coupling factor and do it: it was in the "more" to come, maybe)... :-)

-and finally, "corrected" vs "not corrected" plots translate the recorded tone of a same stacked humbucker model in two versions: the first was affected by a not so subtle and unpleasant resonance beyond the main resonant peak. The second has been tweaked to have a flatter response cancelling this parasitic resonance and the correction applied had to do with the questions of capacitance evoked in this topic.

The whole thing is what it is: a rough sketch posted on a hurry a sunday morning, without knowing if i'll be able to keep posting and if my thread will make sense. Hence the "more later if time permits". :-))

[Helmholtz]

It is essential for quantitatve evaluation to know if the vertical scale is linear or log (dB) and how much change a division step means

Please show your HB model circuit. E.g. with a standard series wired HB the coil capacitances are not in parallel.

[freefrog]

Hi again,

Time to lunch and no archived data at disposal right now but I'll try to reply the most precisely possible in this situation, thx for your understanding.

- The vertical scale is in dB in each of my previous posts, as most often by convention. With or without resonant frequencies measurements laid upon them, spectra of played tracks on dark screens follow vertical steps of 3dB. Isolated resonant frequencies on a white screen with linear frequency axis follow vertical steps of 1dB. 5Spice sims had auto-organized dB scales (with 5dB steps, typically) but of course, these dB scales have been aligned as soon as pics were compared. And all of these pics have indeed been trimmed intentionally for contextual reasons.

-I don’t post this answer from the computer hosting the necessary 5spice / LTspice programs (which is not connected to the Web, anyway) but… the pickup model that I’ve used here reminds the Thevenin equivalents developed separately by various persons these last years, after Helmuth Lemme or Malcolm Moore. Mine looks like an unpublished 5spice prequel of the LTspice sketch entitled in the page below “Modeling a humbucker wired in series...”

https://guitarnuts2.proboards.com/th...guitar-ltspice

The only difference is that for the tests shared here, I’ve connected the capacitor of the second coil to ground and not to the negative of its own voltage source. Why? Because for me, it allows to mimic the influence of parasitic capacitance as increased and distributed by 4 conductors cables. YMMV.

Now and to be clear: my goal here was not to promote a given theoretical model as faithful to such or such pickup - like the complex models with multiples inductors and resistors, in series and/or parallel, elaborated by your friend Manfred Zollner that I respectfully salute for his work.

The simpler 5spice sims used in my own first posts were simply judged functional enough because they reflect(ed) the measurements and visual captures of recordings done here these last decades.

Finally, the choice of trimmed screenshots was a way to insist on the principles involved rather than on my own experimental strategy and/or on numeral values, always discussible: people might now desire to do similar (or different) tests about stray capacitance and make their own conclusions. My intention in this topic didn’t go beyond that initial idea to possibly excite curiosity.

[freefrog]

NOTES before to continue this topic (or not).

Screenshots in my posts 2 & 3 show secondary peaks beyond the main resonant frequency and 4 to 9 dB lower than it.

It could logically be found negligible because probably not perceptible, especially when it’s located in a high frequency range that most guitar loudspeakers don’t reproduce.

But the idea to write the posts 2 & 3 came to me from an experience suggesting that things are not so clear - because guitar amps can be bright and can host bright switches, because IME the harmonic peaks and dips of their loudspeakers can align themselves with those of pickups, etc.

Ten or twelve years ago, a stacked HB based on a relatively unusual design was tested here. Its resonance was measured and the pickup model was played in various pickups positions / amps / musical contexts.

What happened is suggested in post 3: the design was initially generating a secondary resonance apparently negligible when the pickup was played direct to the board but making it really painful to hear through a Twin reverb with C12N RI’s or a JC120 with its aluminium dust caps. Through a few experiments around stray capacitance inherent to the PU and its hook up wiring, the design was corrected (by the humble freefrog) for a flatter response, musically more acceptable in any playing situation.

This experience drawn the attention on a second stack with the same issue, then on humbuckers… and it appeared that some 4 conductors HB’s were also producing secondary peaks, equally able to generate a kind of “drone effect” or unnatural harmonic “ping” through bright amps and cabs.

That’s the empirical motivation behind a part of this topic : I wouldn’t have wrote it if I hadn’t experienced several times the problem of annoying secondary resonance described here and worked to solve it.

More later if possible… and if it’s worth it, for readers as well as for me.

In the meantime, back to immediate reality. i'd like to be already retired but have still a few years of busy work and exigent family life before that... :-/

[Helmholtz]

The only difference is that for the tests shared here, I’ve connected the capacitor of the second coil to ground and not to the negative of its own voltage source.

That will give a different frequency response. You need extra caps to simulate the effect of the 4-conductor cable.

[freefrog]

That will give a different frequency response. You need extra caps to simulate the effect of the 4-conductor cable.

Agreed, extra caps would help to emulate more faithfully the multi-wires cable... and taking into account the coupling factor of coils would also help to simulate more realistically the whole system, rather than to simply mimic it through a simplifying projection as I 've done. This preliminary step based on a common spice model might be completed thx to the more refined transformer based model mentioned in my post 2 ASAP, if it's doable. Thx. :-)

[Antigua]

I've measured a lot of Strat pickups that are explicitly machine made, like the Tonerider Surfari set with around 80pF capacitance, and the BYO Storms around 90pF, both made in Hong Kong and retail for a rather low price, in contrast to Fender vintage correct single coils, that typically have a parasitic capacitance of 120pF (Pure Vintage series, Custom Shop '69), and then there's a winder going by the name Bootstrap, I measured their "54 Vintage" and Golden Ale", heavily implied that they're hand guided, but have a capacitance of around 200pF in all six pickups across the two sets. So I think it makes more sense to evaluate the capacitance of pickup sets on a model by model basis (or maybe even manufacturer to manufacturer), rather than to take the claim of scatter winding as an indication of capacitance. I suspect the difference might have more to do with the "build thickness" of the magnet wire, which is something I haven't been able to survey, but would help explain the why then trends often seem to follow with the manufacturer, rather than the model or production method.

[freefrog]

For me, it potentially explains why scatter wound pickups are often perceived as different by players who fiddle a lot with their volume and tone controls (as I do myself). YMMV. :-)

As a matter of fact, empirical reality is generally more complex than such generalizations. :-)

Maybe I should precise that "scatter winding" is not in my mind something asserted by winders but something that I've noticed on some transducers weakly capacitive. That said, I don't systematically equate scatter winding with hand guiding, nor any of both with low capacitance... A relatively well known winder, with whom I've had regular discussions, was/is using a winding machine set to mimic some randomness due to hand guiding but he did/does try to obtain the lowest possible capacitance in a very conscious way, that the notion of "scatter winding" doesn't reflect in my understanding. Conversely, I've already manipulated coils whose winding was loose and random to the point to get off the bobbins but still with a very high inner capacitance. And of course, I don't neglect materials with their variable dielectric constant, that aging can alter and that even ambient humidity can change (I remember Alex Keinis stating the equivalent of 147pF more parasitic capacitance in a coil exposed to steam of boiling water).
The statement above was just based on the fact that in my subjective experience, pickups with a low inner capacitance are most often not those with evenly machine wound coils (and I personally don't find that surprising when I think about how to minimize stray capacitance in general with electronic devices). .Other subjective experiences may differ.

Sentences subject to change without notice, if ever their apparent meaning appears to me as not reflecting my original thoughts in another language. My contributions to my own topic will also remain "scattered" since they depend on variable free time.

[freefrog]

I’ve dug a bit in my archives to see what I could share and toyed when I could with sims… here is another partial harvest.
Sorry by advance to keep trimming my screenshots: full pics of my humble work are for other contexts (those shared in this post belonging to the original designer of the pickup took as an example). Trimmed excerpts should suffice anyway to share some ideas.

Below is on a log scale the electrically induced response of the stacked HB model evoked in my posts 3 & 8, with its “up turned nose” 2d resonance found musically so annoying through bright amps… There’s actually three lines for 3 pickups built on the same design (neck, mid, bridge transducers). These measurements had been done with the PU’s wired in a guitar, with a pair of 250k controls and through a 590pF cable plugged to a 1M input.



Here is a first 5spice sim of this response, based on the same simple model than in my first posts here but tweaked to reflect the conditions of my physical tests (since I don’t use no integrator for Rz measurements).



Here is now a more refined 5spice sim, where the pickup is treated as a transformer with a defined coupling factor. The secondary resonance itself seems to be emulated with a bit more accuracy but it’s not the case for the frequencies beyond that. The “refined” model would apparently have to be refined more…



For the moment, none of the sims is exactly like the measurements, therefore, but both are close enough to confirm that the “up turned noses” in Rz was not an experimental artefact due to human error.

It had to be said before to underline that “secondary resonances” are stubborn: they don’t shift with the main resonant peak. Below is another measurement illustrating this idea, with the help of the aforementioned stacked HB’s, wired this time in series: albeit a bit flattened and dragged down in the spectrum by series wiring, the secondary Rz is still there and remains roughly at the same place (here on a “plateau” between 7 & 9 khz) even when cable capacitance or a tone control @ 0/10 shifts down the main Rz.



Below is now the response of each of the 3 pickups aforementioned once the original design corrected capacitively. As the goal was to emulate the response of standard single coils, resistive load has also been increased, thx to pots of higher resistance.



And finally, here is a comparative synthesis: upper pic = electrically induced response of original design with 250k pots VS corrected version with 500k controls. Bottom pic = the same things, modelized through 5spice.



These pics are meant to illustrate the following testimonial: if a pickup design produces musically unpleasing secondary resonances (and/or symmetrically missing harmonics), these parasitic peaks and dips won’t disappear easily. BUT they can be modelized, then corrected thx to a different distribution of stray capacitance in the pickup and/or its hook up wires. That’s what had been done above with the initially problematic design evoked. And it’s potentially transposable to transducers with 4 conductors cables.

Now, that’s still something depending on the randomness of physical reality. In the example above, the capacitive correction was of 50pF higher for the bridge pickup, for instance, albeit all three pickups were meant to be identical.

FWIW. As previously, more later if possible.

[Helmholtz]

Interesting, appreciate the effort.

I guess you're no going to reveal your "correction" measures?

[freefrog]

Years after the facts, I feel free to share the ideas at work but I can't be more precise about the product involved, its 5spice model and the corrections applied to it, since it was not my design.
I've simply said to myself that people knowing this kind of stuff could/would extrapolate on the principles involved and maybe make them work for their own benefit. :-)