Damping caused by the unused coil when splitting a humbucker

[antigua]

I came upon another potentially useful finding while messing with the PAF Master regarding a split pickup situation where the adjacent disabled coil has, or does not have continuity. I'm starting a new thread for this particular effect because it doesn't seem likely that this is specific to the PAF Master, though it's the ony pickup I've tested with so.

In short, when the unused coil adjacent to the active split coil is a closed circuit, the unused coil is still moving a current and it causes damping on the primary coil. If the unused coil is open circuit, the the unused coil does not move a current (or at least not nearly as much), and does not cause any, or nearly as much, damping.

This potentially makes a real difference, because some coil splitting schemes result in the unused coil having continuity, for example, if you simply short the coil out with an SP/ST, but some more exotic wiring schemes split in a way that leaves the unused could open and without continuity. 

This plot shows high damping with the full assembled coil, with both coils split, one coil for each graph, and the other coil either is, or is not closed or open. The cover is on the pickup in these two plots:



The screw coil is in circuit with the slug coil open and closed:



The slug coil in circuit with the screw coil open and closed:




So this shows that the effect works with either coil in relation to the other.

There are still some questions to be asked, which I will try to answer later unless someone else wants to do this experiment first..

1) is this effect also present when there is no cover on the humbucker? By extension, is eddy current w/ cover activity in the unused coil responsible for the damping?

2) how much effect is seen under load, or, does this effect have any audible consequence in practice? The unloaded plots show damping in excess of 3dB, which could make the split pickup sound audibly darker if the same extent of damping holds under load. 


******** UPDATE #1 *********


To answer question #1, what effect did the cover have, it caused a lot of damping, but it was not solely responsible for the damping. Even with the cover removed, when the opposing coil is "closed circuit", the active circuit is still heavily damped by way of magnetic coupling between the two coils.

My theory at this point, and I think it would be hard to argue otherwise, is that we have a "transformer" situation, and when the unused coil is closed, the secondary winding is shorted, with a capacitor in parallel. When the coil is open, the secondary winding is in parallel with nothing but it's own parasitic capacitance.

I just happened upon another discovery which probably explains the asymmetry between the slug and screw coil, in these tests and in the PAF Master thread: the slug coil appears to have over 100pF more capacitance than the screw coil. With all the metal removed, both coils have about 900mH inductance, but the slug coil has a resonant peak of 10kHz while the screw coil has a resonant peak of 13kHz. I calculated that it would require about 110pF capacitance difference in order to put the resonant peaks that far apart for 900mH inductance. So when the secondary winding, or the unused coil is open, it's only in parallel with it's own capacitance, and since the slug coil has such a high capacitance, it has a much more dramatic effect upon the screw coil, than the screw coil's capacitance has upon the slug coil. 


The screw coil is in circuit with the slug coil open and closed, cover removed:


The slug coil in circuit with the screw coil open and closed, cover removed:



I pealed back the tape on the coils to see if there was an obvious difference that might explain the difference in capacitance, but I see no obvious differences. The insulation coating is an unusual brown color, not the usual dark red formvar.

Slug coil top
Screw coil bottom

[ashcatlt]

This actually touches on a debate we've had around here for a while and has implications in some multi-pickup schemes as well. If you get a chance and wouldn't mind adding a couple more questions to your list...

3) Does it make a difference if the shorted coil is on the "top" or "bottom" of the stack? Like, does it make a difference whether it's shorted to ground or to hot?

4) Does if make a difference if the two coils are physically further apart? If you've got two separate pickups in series, and short one, do we get the same effect?

[antigua]

Oct 14, 2016 11:57:00 GMT -5 ashcatlt said:

This actually touches on a debate we've had around here for a while and has implications in some multi-pickup schemes as well. If you get a chance and wouldn't mind adding a couple more questions to your list...

3) Does it make a difference if the shorted coil is on the "top" or "bottom" of the stack? Like, does it make a difference whether it's shorted to ground or to hot?

This is a side by side humbucker, so the issue of stacked humbuckers gets confusing, and it warrants investigation also. It wouldn't surprise me to learn that there is frequency dependent push back with a stack, given their reputation for having blunted high end.

The issue of grounding is yet another variable, because if the unused coil is grounded at only one end (grounded at both ends will permit coil continuity as the "closed" test above demonstrates) then you will have capacitive coupling between the unused coil and the active coil. If the unused coil is fully disconnected (as seen in the "open" test above), then you'll have essentially no capacitive coupling between the two coils. 

From what I've seen so far, I don't think there is dramatic capacitive coupling between the active and unused coil when one lead of the unused coil is groudned, because there is a about two millimeters of gap between the sides of the two coils. Contrast that with a tapped single coil, where the unused outer coil is literally wrapped around the active coil and in point blank proximity, and you see up to 600pF capacitance, which is like a 15ft guitar cable by itself. Check out my thread on the SSL-4 to see that particular test.

> 4) Does if make a difference ifthe two coils are physically further apart? If you've got two separate pickups in series, and short one, do we get the same effect?

In this case the cover is on top of the humbucker, so I can't move the coils farther apart, though I can do tests with one coil outside of the enclosure in order to determine if coupling is required. I'm nearly certain that magnetic coupling between the coils is required, because the shorted coil itself would be voltage neutral with respect to the active circuit. It would be as if you touched both ends of a battery to the same location within a circuit; it would add nothing to the circuit. In other words, I think the unused coil is only relevant insofar as it's able to interfere with the "in circuit" coil.

[ashcatlt]

Sorry. When I said "stack", I wasn't actually talking about physical properties. That's a different issue which is also interesting, but I was actually just talking about the wiring. Basically, does it make a difference which comes first in the series? Normally when we split an HB, we do it by shorting the series link to system "ground". If we want to split to the other coil, we can either just short that series link to the system "hot", or we can rewire the whole thing so that we can still short to ground. There are ongoing arguments as to whether one of these is better than the other. Some of that is about possible noise effects (kind of outside of your line of research), but there is some concern regarding possible effects on tone.

I think part of the argument also involves the idea of the shorted coil "stealing energy" from the string. IDK if that is something you have the capability to test for...

[antigua]

Oct 14, 2016 13:46:25 GMT -5 ashcatlt said:

Sorry. When I said "stack", I wasn't actually talking about physical properties. That's a different issue which is also interesting, but I was actually just talking about the wiring. Basically, does it make a difference which comes first in the series? Normally when we split an HB, we do it by shorting the series link to system "ground". If we want to split to the other coil, we can either just short that series link to the system "hot", or we can rewire the whole thing so that we can still short to ground. There are ongoing arguments as to whether one of these is better than the other. Some of that is about possible noise effects (kind of outside of your line of research), but there is some concern regarding possible effects on tone.

Oh, so you would alternatively split to the second coil by connecting the series joint between the two coils the positive lead, instead of connecting the series joint to ground?

In both cases, the inactive coil has no direct effect on the circuit, because the coil is shorted, and so there is no voltage difference between the two sides of the unused coil. Current still flows in the unused coil, and it resonates with it's own reactances, but the current and the voltage go nowhere. 

However, since that unused coil is still "reacting", or generating an alternating magnetic field around itself, it's magnetic field is able to interfere with the active coil and induce a current in it, and that is what appears to be happening here; magnemotive interference, as opposed to electromotive interference. 

Oct 14, 2016 13:46:25 GMT -5 ashcatlt said:

I think part of the argument also involves the idea of the shorted coil "stealing energy" from the string. IDK if that is something you have the capability to test for...

That's a good observation, and it directly applies here. When the unused coil is closed and current flows, the flowing current creates a new magnetic field that opposes the moving guitar string (Lenz's law). When the unused coil is open, current does not flow, and so no oppositional magnetic field is created. But not so fast; the magnetic coupling between the coil and guitar string is so low that the opposing magnetic field created by the pickup coils are too weak to have any measurable effect on the sustain of the guitar string. Only a small fraction of the magnemotive force of the guitar string is received by the guitar pickup, and a tinier fraction makes it back to the string via Lenz' law. It order for the pickup to damp the guitar string, it would have to be as close and as powerful as an EBow.

[JohnH]

Yes, we have been aware of this shunting issue for years here on GN2, and the way that it sucks energy if the unused coil is forming a closed loop. But we were never able to really nail it in a quantitative way. Most coil cuts everywhere are done by shunting a coil, rather than moving the ground up to the connection point.

I think a key question is what differences do you see in a loaded test, with the other coil either shunted on open?. If its more than a db, we should revive our campaign to avoid shunting coils in wiring designs!

[stratotarts]

Yes, the wiring for a short is more convenient on on multi switch, than for an open. The loaded difference on an SD SH-1N measures about 1.5 dB:

[antigua]

Oct 14, 2016 15:59:17 GMT -5 JohnH said:

Yes, we have been aware of this shunting issue for years here on GN2, and the way that it sucks energy if the unused coil is forming a closed loop. But we were never able to really nail it in a quantitative way. Most coil cuts everywhere are done by shunting a coil, rather than moving the ground up to the connection point.

I think a key question is what differences do you see in a loaded test, with the other coil either shunted on open?. If its more than a db, we should revive our campaign to avoid shunting coils in wiring designs!

I think the schematic difference looks like this:

Unused coil open





Unused coil closed / shunted




Correct me if I'm wrong, but it appears to me that shunting the secondary coil prevents that coil's capacitance from adding with the primary coil. I think that's what is seen in the practical plots above. Agree or disagree?

If we want to take it a step further, we can put a capacitor across the unused coil and see if the effect is even more dramatic. 

[stratotarts]

The disabled coil still has resistance.

[antigua]

Oct 14, 2016 16:20:21 GMT -5 stratotarts said:

Yes, the wiring for a short is more convenient on on multi switch, than for an open. The loaded difference on an SD SH-1N measures about 1.5 dB:

If you have the time and the setup handy, it would be interesting to see if the curves are similar if you reverse the coils. I think I have some  major capacitance differences between my coils causing a large asymmetry.  

[antigua]

Oct 14, 2016 16:23:21 GMT -5 stratotarts said:

The disabled coil still has resistance.

Is that coil resistance only in play when the secondary coil is a closed loop, or is it always in play? 

Is his how it would look:

Closed secondary:


Open secondary:

[stratotarts]

Yes, the resistance should be included. There is little difference in behaviour between the screw and slug coil, in the effect of the other coil being open or shorted:

SH-1N (stock no cover)

[JohnH]

Oct 14, 2016 16:20:21 GMT -5 stratotarts said:

Yes, the wiring for a short is more convenient on on multi switch, than for an open. The loaded difference on an SD SH-1N measures about 1.5 dB:
View Attachment

Good, that's enough to make it worth avoiding where possible IMO. It can usually happen with the same switches as usually used, unless we are also doing clever tricks with phase switches too to maintain humcancelling

[perfboardpatcher]

Very interesting tests, Antigua!
It reminds me of the tests you performed with a single coil sized pickup with stacked coils on Strat-Talk.
I stick to the explanation I'd given over there. Not damping but a superimposed signal originating from the shorted coil.

I presume you stacked the excitation coil on top of coil #1 (illustration below) and therefore voltage source #2 except for being smaller in amplitude is out of phase with voltage source #1. My little theory could be falsified by placing the excitation coil in the usual way for measuring (rotated by 90 degrees in between the humbucker coils), because when correct, voltage sources #1 and #2 would be in phase now and test results should look entirely different.

[antigua]

Here are three plots, each one has the secondary coil open and closed. 

This plot has the driver coil laying on it's side over both coils:



With the coil sideways and between the primary and secondary, the shorted secondary has a substantial effect on the output voltage. There is high end attenuation and a much sharper resonant peak. 


---------


This plot has the driver coil upright over the primary coil:



With the coil only above the primary coil, it matters much less whether the secondary is open or closed. Also notice that this curve closely resembles the first plot, with the driver coil over both pickup sideways, having the secondary coils shorted.  

This arrangement causes the secondary coil to be in the weaker return path of the driver coil.


---------

This plot has the driver coil upright, over the seconary

NOTE: this plot is 5x magnified, 5dB per vertical segment. This arrangement caused dramatic swings in output voltage by frequency.




The most dramatic attenuation occurs this way, with alternating magnetism applied most directly to the secondary. This arrangement causes the primary coil to be in the weaker return path of the driver coil.



-----

Finally, here is the driver coil beside the screw coil:




The fact that this plot shows so little difference pretty much proves that the interference caused by the secondary coil is induced by the driver coil, not the primary. 

Here is a pic of the driver beside the screw coil:




In all of the plots, you see two distinct resonant peaks, with one peak prevailing over another depending on the geometrical circumstances. One of the peaks is at 10kHz, the other is at 12kHz. 

[perfboardpatcher]

Oct 15, 2016 17:19:13 GMT -5 antigua said:

This plot has the driver coil upright, over the seconary

NOTE: this plot is 5x magnified, 5dB per vertical segment. This arrangement caused dramatic swings in output voltage by frequency.




The most dramatic attenuation occurs this way, with alternating magnetism applied most directly to the secondary. This arrangement causes the primary coil to be in the weaker return path of the driver coil.

 

I'm trying to grasp the test results. If my schematic were correct than "slug coil shorted" should produce more signal than "slug coil open". And what is it that caused the increased peak for "slug coil open", some sort of capacitive coupling?

[antigua]

Oct 17, 2016 14:15:36 GMT -5 perfboardpatcher said:

Oct 15, 2016 17:19:13 GMT -5 antigua said:

This plot has the driver coil upright, over the seconary

NOTE: this plot is 5x magnified, 5dB per vertical segment. This arrangement caused dramatic swings in output voltage by frequency.




The most dramatic attenuation occurs this way, with alternating magnetism applied most directly to the secondary. This arrangement causes the primary coil to be in the weaker return path of the driver coil.

 

I'm trying to grasp the test results. If my schematic were correct than "slug coil shorted" should produce more signal than "slug coil open". And what is it that caused the increased peak for "slug coil open", some sort of capacitive coupling?

In that particular graph the driver coil was over the secondary coil, meaning the primary was in the driver coil's return path, so the two humbucker coils are producing opposing currents and fields, so I believe the drop in voltage is a result of flux cancelling. Since the driver was directly over the secondary, it's opposing field was especially strong. 

Based on what I've learned from this experiment, if I ever want to measure just one coil by itself without taking the whole humbucker apart, I would place the driver beside the coil being tested, rather than anywhere on top of the pickup, in order to make sure that the secondary coil sees as little alternating flux as possible. 

[perfboardpatcher]

Okay, so I treated my Ibanez Paf Pro pickup as if it were a transformer. Signal generator's output went into the first coil and the oscilloscope measured the voltage across the second coil. No other components were used. The sigal generator's output level was 5.0 volts peak.
The signal (peak) levels across the second coil were:
100 Herz - 240 mV
200 Herz - 420 mV
500 Herz - 700 mV
1.0 kHerz - 800 mV
2.0 kHerz - 850 mV
7.9 kHerz - 1.3 V
max 12.7 kHerz - 2.6 V

Putting a plate with a copper layer on top of the pickup caused the maximum to drop to 2.3 V. The other frequencies didn't seem to change much.

Perhaps its' interesting to mention that my Paf Pro has the coil screws sawed off, the screws barely protrude from underneath the baseplate.
When measured there wasn't a scoop like in Antigua's case.
Is there some 1:1 relationship between scoop and screw coils protruding, also Antigua's test results of SD custom 5 seems to point in that direction?

[antigua]

Even when you cast the humbucker as a transformer, there's still a large inductive reactance that increases with frequency, because the two coils are so poorly coupled. The reactance causes a +6dB/octave voltage to frequency relationship, so you only see a scoop if you integrate to remove that +6dB / octave slope, using a circuit such as the one stratotarts designed. The scoop is still there, it's just harder to recognize it for what it is, because it's within the slope.  

The copper sheet causes the most losses at the max frequency because eddy currents are a resistive and inductively reactive impedance, so they become more prominent with an increase in frequency. It's like you have a transformer where the coils are poorly coupled, based on the orientation of the copper sheet with respect to the coil's magnetic field, and the secondary coil, that is the copper sheet, is in series with a resistor, the value of which corresponds to the resistivity of the copper. Since the impedance is complex, the shape of the curve you'd see caused by eddy current losses would vary depending on where the copper sheet is in relation to the coil, altering the extent to which it inductively couples with the coil. 

Shorter screws would just have the effect of slightly lowering the inductance, probably by no more than a few millihenries, and it would cause a very slight decrease in flux density in the strings. 

[perfboardpatcher]

Sorry for the confusion! I meant that I measured my Paf Pro with driver coil/integrator circuit some time ago. Like I said, not so much scoop in my case.

I thought my measurement as a way to determine the coupling factor of the "transformer".
(Leakage inductance)

Now with the circuit I previously posted in mind (& unlike the drawing says, out is to be measured across the "C" of coil 1)
Say, k= 850/5000, then the inductances of coil 1 are known, coils are more or less identical, therefore same k and inductance ratio for coil 2, the "transformer" has a 1:1 ratio. A simulation should now be possible since all the component values are known.

[straylight]

Oct 14, 2016 11:28:12 GMT -5 antigua said:

I pealed back the tape on the coils to see if there was an obvious difference that might explain the difference in capacitance, but I see no obvious differences. The insulation coating is an unusual brown color, not the usual dark red formvar.

I'm not sure if I'm stating the obvious, but most of the DiMarzio pickups are wound asymmetrically. It's most obvious in things like the Tone Zone which has a double peaked frequency response.

[antigua]

May 22, 2018 5:48:14 GMT -5 straylight said:

Oct 14, 2016 11:28:12 GMT -5 antigua said:

I pealed back the tape on the coils to see if there was an obvious difference that might explain the difference in capacitance, but I see no obvious differences. The insulation coating is an unusual brown color, not the usual dark red formvar.

I'm not sure if I'm stating the obvious, but most of the DiMarzio pickups are wound asymmetrically. It's most obvious in things like the Tone Zone which has a double peaked frequency response.

There's not really an audible double peak, it's just a marketing gimmick. 

[straylight]

EDIT: DISREGARD! CANNOT REPRODUCE

Have you seen the the frequency response curves? They're not on echoes of mars and I can't find anything on here.

I'm not a fan of this pickup, it's not working for me in the mahogany guitar it came in, it might go into something bright later on.

This was done with a 200kOhm  and 470pF load in place. I was waiting to post data until I had an integrator in place and maybe better hardware, but it's relevant here.

[blademaster2]

I am interested in the apparently profound impact that you saw with the cover removal.

I heard the difference myself recently when adding/removing a chrome plated nickel cover from a single coil pup, but I have not until now seen any measurements to show what impact the cover introduces to the pickup response. I think that those plots did not include the loading effect of the cable and amplifier (please correct me if I am wrong).

Are there any results that anyone has to show the frequency response of a loaded pickup (single coil or humbucker, connected via a volume control circuit and loaded by a length of cable and an amplifier), with and without a cover to show what difference the cover makes when all else is equal?

[antigua]

May 24, 2018 8:40:38 GMT -5 straylight said:

Have you seen the the frequency response curves? They're not on echoes of mars and I can't find anything on here.

I'm not a fan of this pickup, it's not working for me in the mahogany guitar it came in, it might go into something bright later on.

This was done with a 200kOhm  and 470pF load in place. I was waiting to post data until I had an integrator in place and maybe better hardware, but it's relevant here.

I don't have a Tone Zone, but I have several DiMarzios with coils features different gauge wires, including the Norton and the Humbucker from Hell

guitarnuts2.proboards.com/thread/7757/dimarzio-air-norton-analysis-review

guitarnuts2.proboards.com/thread/7754/dimarzio-humbucker-analysis-review

And saw no 'audible' double peak. There is a second peak past the primary LC resonance, but that's typical of all humbuckers, and it disappears with load, and comes about because there are two separate coils, not because of a deliberate coil imbalance. In fact, according to the patent, both coils are supposed to contain the same number of turns, which means they will have very close inductance values, since inductance follows turns. The degree to which your RB and WG plots lines overlap demonstrates the extent to which the two coils are otherwise similar. We know from wiring mods that if you want to create a band stop which produces a strong double peak:



And this is similar to why humbuckers have a slight double peak, especially if there is four conductor wiring, because the junction between the two coils capacitively couples with ground. The simulated plot shows 4nF to ground in between the coils, which is a much higher value than you would get through parasitic means alone. 

I'm not sure why your series plot has a double peak, I might try to see if I can recreate that with my DiMarzios. 

[antigua]

May 24, 2018 11:12:22 GMT -5 blademaster2 said:

I am interested in the apparently profound impact that you saw with the cover removal.

I heard the difference myself recently when adding/removing a chrome plated nickel cover from a single coil pup, but I have not until now seen any measurements to show what impact the cover introduces to the pickup response. I think that those plots did not include the loading effect of the cable and amplifier (please correct me if I am wrong).

Are there any results that anyone has to show the frequency response of a loaded pickup (single coil or humbucker, connected via a volume control circuit and loaded by a length of cable and an amplifier), with and without a cover to show what difference the cover makes when all else is equal?

If the cover is brass, there is a strong attenuation, regardless of whether the pickup is loaded or not, but if the cover is nickel silver, the unloaded attenuation is not as extreme, loaded attenuation is very small. Using a nickel silver cover is almost like having no cover at all. What particular pickup do you have?

brass:




nickel silver:

[blademaster2]

Thanks,

I used a StewMac nickel humbucker cover over a Seymour Duncan single coil SSL-1. I cut the sides lower on the cover, and drilled holes for the SSL-1 poles to protrude through them (made a mess of the hole locations, but that is beside the point  :-)  ).

With that I was able to quickly slip the cover over the SSL-1 and remove it just as easily, and that permitted me to make a reasonably good comparison. I did notice a slight loss in crispness in the tone (very subtle), and I gather that is the loss that your plot shows for "loaded/covered" versus "loaded/uncovered". Very interesting to see, thanks very much. Even with the difference between humbucker and single coil, it at least explains that there is indeed a slight high-frequency loss in the response when a nickel cover is added, however subtle.


My ears were not wrong.  :-)

[straylight]

May 24, 2018 12:02:29 GMT -5 antigua said:

I'm not sure why your series plot has a double peak, I might try to see if I can recreate that with my DiMarzios. 

I'm gonna do a proper analysis of the tone zone and my methods. It could be that I screwed something up. Shall we continue this conversation when I get that done later this evening?

EDIT: Disregard the double peak, I can't reproduce, investigating.

[guitarnerdswe]

I'm just wondering, what classifies as an open coil in these discussions? Do both leads of the coil have to be free floating (not connected to anything)?

Or is it enough if only one wire is floating, with the other going to hot or ground?

Also, is connecting both leads to ground the best solution for minimal dampening on the active coil?

[JohnH]

Apr 21, 2021 8:11:09 GMT -5 guitarnerdswe said:

I'm just wondering, what classifies as an open coil in these discussions? Do both leads of the coil have to be free floating (not connected to anything)?

Or is it enough if only one wire is floating, with the other going to hot or ground?

Also, is connecting both leads to ground the best solution for minimal dampening on the active coil?

My view would be that one wire to ground and one disconnected is fine to reduce the damping of tbe other active coil, and that a full disconnection is no further benefit.

These effects are mainly caused by magnetic coupling between the two coils, due to their proximity and usually a single magnet. So currents in the 'off' coil can draw energy from the system affecting the 'on' coil.

On my version of JP wiring, the coil split is done by moving the ground connection to the join between coils. Now one coil is open and has only a ground, and the other is between hot and ground.

But, I also have humbuckers with normal shunting coil cuts, and they sound fine for what they do.

Another implication, if I have the right understanding, is that these effects will apply to coils within a humbucker, and not really to schemes with series wiring of singles where one pickup gets shunted. In this case, energy is still being drawn into that shunted pickup, but it comes from the string (I think, with little effect) without interacting with other coils.