|
Post by guitarnerdswe on Apr 23, 2021 10:39:12 GMT -5
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. Thanks for your reply! I just wired up a 2PDT switch to go between these two options for the unused coil: 1. One wire to ground, the other disconnected 2. One wire to hot, the other disconnected I can honestly say that the difference is minimal at best. I could not tell which option was which in a blind test. I could tell there was some difference, but not which was which.
|
|
|
Post by JohnH on Apr 23, 2021 18:41:13 GMT -5
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. Thanks for your reply! I just wired up a 2PDT switch to go between these two options for the unused coil: 1. One wire to ground, the other disconnected 2. One wire to hot, the other disconnected I can honestly say that the difference is minimal at best. I could not tell which option was which in a blind test. I could tell there was some difference, but not which was which. Thanks for your testing. I thought id make this diagram of the two versions I'm talking about, since I'm not sure whether these are what you tested: Type A is the usual shunting version where all leads remain connected, while B is where the one lead on the unused coil is disconnected and the ground is moved to the central join between coils. If this is different to your tests, then I'd be very interested in your impression of any differences.
|
|
|
Post by guitarnerdswe on Apr 24, 2021 7:30:18 GMT -5
Thanks for your reply! I just wired up a 2PDT switch to go between these two options for the unused coil: 1. One wire to ground, the other disconnected 2. One wire to hot, the other disconnected I can honestly say that the difference is minimal at best. I could not tell which option was which in a blind test. I could tell there was some difference, but not which was which. Thanks for your testing. I thought id make this diagram of the two versions I'm talking about, since I'm not sure whether these are what you tested: Type A is the usual shunting version where all leads remain connected, while B is where the one lead on the unused coil is disconnected and the ground is moved to the central join between coils. If this is different to your tests, then I'd be very interested in your impression of any differences. That's not quite what I was testing. I had my hot of the unused coil going straight to hot, not through the other coil (I'm also running the humbucker in parallel on this guitar via a P/P, so that's why). So I had these two options: A. Unused coil hot straight to output of 5-way, other end disconnected. B. Unused coil hot disconnected, the other end to ground. It's important to note that while my guitar has passive pickups and a passive tone control, the output is buffered (so the cable capacitance is not there). It's a 2020 Music Man Luke III with the new Music Man high output humbucker and Cutlass single coils (60s style, around 6 k). I did my experiments with the bucker split to the screw coil together with the middle pickup, so the difference might have been lessened some. The humbucker is also quite dark, which might impact it all. I have say, I noticed more of a difference on the low strings. A bit more snap and volume with my option B. Could be placebo though.
|
|
|
Post by aquin43 on Apr 24, 2021 9:54:23 GMT -5
Consider two equal coils, inductance L, resistance R, mutual coupling M and one shorted.
The effect of short circuiting one of the coils is to reduce the inductance and increase the resistance of the other in a frequency dependent way. Given L and L coupled by M, the coupling coefficient is K = M/L
The new inductance becomes Ls = L * (1 - K^2 * U(f))
The new resistance becomes Rs = R * (1 + K^2 * U(f))
The function U(f) is f^2 / (f^2 + fc^2)
where fc = R/(2*pi*L)
U(f) is a smooth curve that goes from zero to one as f increases. It is equal to 1/2 when f = fc. For a humbucker where the overall resistance is 8k8, the inductance Lpu is 3.8H and the mutual inductance 312mH
(drive a current through one coil and measure the voltage induced in the other), we get. L = (Lpu - 2*M) / 2 = 1.59 (from the sum formula L=L1+L2+2M) R = 4k4 K = 0.197, K^2 = 38m8, R * K^2 = 171, 1 - K^2 = 0.961 fc = 4k4/(2*pi*L) = 441Hz
So R increases by 171 ohms and L falls by 4%, with half the change occurring by 441Hz.
Of course, the coil capacitance is there to complicate this still further but it is in parallel with the resistance which should dominate beyond the resonant peak of the unloaded coil.
Maybe to characterise a humbucker, the mutual inductance should also be measured.
|
|
|
Post by guitarnerdswe on Apr 24, 2021 13:52:31 GMT -5
Consider two equal coils, inductance L, resistance R, mutual coupling M and one shorted.
The effect of short circuiting one of the coils is to reduce the inductance and increase the resistance of the other in a frequency dependent way. Given L and L coupled by M, the coupling coefficient is K = M/L
The new inductance becomes Ls = L * (1 - K^2 * U(f))
The new resistance becomes Rs = R * (1 + K^2 * U(f))
The function U(f) is f^2 / (f^2 + fc^2)
where fc = R/(2*pi*L)
U(f) is a smooth curve that goes from zero to one as f increases. It is equal to 1/2 when f = fc. For a humbucker where the overall resistance is 8k8, the inductance Lpu is 3.8H and the mutual inductance 312mH
(drive a current through one coil and measure the voltage induced in the other), we get. L = (Lpu - 2*M) / 2 = 1.59 (from the sum formula L=L1+L2+2M) R = 4k4 K = 0.197, K^2 = 38m8, R * K^2 = 171, 1 - K^2 = 0.961 fc = 4k4/(2*pi*L) = 441Hz
So R increases by 171 ohms and L falls by 4%, with half the change occurring by 441Hz.
Of course, the coil capacitance is there to complicate this still further but it is in parallel with the resistance which should dominate beyond the resonant peak of the unloaded coil.
Maybe to characterise a humbucker, the mutual inductance should also be measured.
So, can this be dumbed down for a guy like me that got a headache from all this?
|
|
|
Post by aquin43 on Apr 24, 2021 16:03:40 GMT -5
Consider two equal coils, inductance L, resistance R, mutual coupling M and one shorted.
...
Maybe to characterise a humbucker, the mutual inductance should also be measured.
So, can this be dumbed down for a guy like me that got a headache from all this? Try this:
1) You have two identical coils with inductance L and resistance R that share a fraction K of their magnetic fields with each other.
2) If you short circuit one coil, the inductance and resistance seen at the terminals of the other coil are altered.
3) At higher frequencies, the resistance increases by a fraction K squared and the inductance falls by the same fraction
4] The amount of change varies from zero in a smooth fashion according to the curve U which depends on frequency.
5) The reactance (impedance) of an inductance is proportional to frequency and the particular shape of U depends on the frequency, call it fc, at which the resistance of one coil is equal to the reactance of its inductance.
6) The shape of the curve of U(f) is such that half of the change has occurred by fc. In the example, fc is quite low, 441Hz.
This is a simple model of a humbucker with one coil shorted out.
|
|
|
Post by antigua on Apr 24, 2021 22:27:01 GMT -5
A. Unused coil hot straight to output of 5-way, other end disconnected. B. Unused coil hot disconnected, the other end to ground. I'm not exactly sure what your wiring diagram is, so there might be relevant information being left out, but I think the audible difference is small because there's probably not much electrical difference. The impedance is the same in both directions, so if you're essentially just turning the circuit around, it won't make any difference. But it might vary the amount of noise. If you have unused conductive wire on the hot side of the circuit, it will pick up static interference and create a noise voltage across the circuit. If it's hanging off the grounded side, it won't, because the static noise voltage is shunted to ground.
|
|
|
Post by guitarnerdswe on Apr 25, 2021 7:48:41 GMT -5
A. Unused coil hot straight to output of 5-way, other end disconnected. B. Unused coil hot disconnected, the other end to ground. I'm not exactly sure what your wiring diagram is, so there might be relevant information being left out, but I think the audible difference is small because there's probably not much electrical difference. The impedance is the same in both directions, so if you're essentially just turning the circuit around, it won't make any difference. But it might vary the amount of noise. If you have unused conductive wire on the hot side of the circuit, it will pick up static interference and create a noise voltage across the circuit. If it's hanging off the grounded side, it won't, because the static noise voltage is shunted to ground. I've done crude drawing of what I was doing, as simple as I can explain it.
|
|
|
Post by aquin43 on Apr 25, 2021 16:01:34 GMT -5
Humbucker driven by exciter coil modelled in LTspice. Readout is via an integrator. Both coils have eddy current damping, including the extra path from the exciter via the damping loop and so show the dip before the peak. Both coils have the same eddy current losses. In practice, the screw coil is usually more damped. The coils are also mutually coupled.
There is no inter-coil coupling via the exciter L3 because it is open circuit, being driven by a current source. The values of the damping inductors L4 and L5 are arbitrary. What matters are the coupling coefficients ked and the angular frequencies R10/L4 and R11/L5
Brown: the full pickup Red: one coil, centre tap grounded, the other coil ground removed.
Orange: one coil, the other shorted
|
|
|
Post by antigua on Apr 25, 2021 17:36:54 GMT -5
I've done crude drawing of what I was doing, as simple as I can explain it. I don't believe either of these will be significantly different, because it's symmetrical as far as the AC signal is concerned. B should cause more noise because the north coil is ungrounded in that case. This situation is not unlike a transformer, but because the coupling coefficient is so low between primary and secondary, nothing you can do will make much of an audible difference. With a tapped single coil or a stacked humbucker the consequences are more easily audible, and detrimental.
|
|
|
Post by guitarnerdswe on Apr 26, 2021 5:31:42 GMT -5
I've done crude drawing of what I was doing, as simple as I can explain it. I don't believe either of these will be significantly different, because it's symmetrical as far as the AC signal is concerned. B should cause more noise because the north coil is ungrounded in that case. This situation is not unlike a transformer, but because the coupling coefficient is so low between primary and secondary, nothing you can do will make much of an audible difference. With a tapped single coil or a stacked humbucker the consequences are more easily audible, and detrimental. Thanks! Good to know! You might have hit the nail on the head there. It might simply have been the noise floor that made the little difference I was hearing (ratio of guitar tone vs noise). I was running the wires to a P/P hanging outside the guitar, which was grounded, but it was still kinda noisy. Even with the fact that I only tested this with clean or edge of break up tones.
|
|
|
Post by aquin43 on Apr 27, 2021 10:45:42 GMT -5
Humbucker driven by exciter coil modelled in LTspice. Readout is via an integrator. Both coils have eddy current damping, including the extra path from the exciter via the damping loop and so show the dip before the peak. Both coils have the same eddy current losses. In practice, the screw coil is usually more damped. The coils are also mutually coupled. There is no inter-coil coupling via the exciter L3 because it is open circuit, being driven by a current source. The values of the damping inductors L4 and L5 are arbitrary. What matters are the coupling coefficients ked and the angular frequencies R10/L4 and R11/L5
Brown: the full pickup Red: one coil, centre tap grounded, the other coil ground removed.
Orange: one coil, the other shorted The same arrangement with the K coupling between the coils removed. The orange curve overlays the red one. This confirms that it is the magnetic coupling that causes the frequency response change when the unused coil is shorted as has already been suggested.
|
|
|
Post by JohnH on Apr 27, 2021 16:01:29 GMT -5
^^^ a really excellent insightful analysis, thanks!
If its possible, Id be interested to see the same set of runs with a 470k/500pF load and also a 250k/500pF load, then it would represent what we might hear with typical pots and cable connected.
|
|
|
Post by aquin43 on Apr 28, 2021 5:47:35 GMT -5
Repeated with
470k/500p: 250k/500p:
|
|
|
Post by JohnH on Apr 28, 2021 7:05:58 GMT -5
Thanks for running those. They seem to show about 3db more high treble with the coil shorting avoided. It's a small amount, but should be noticeable and worth having.
|
|
|
Post by guitarnerdswe on Apr 28, 2021 10:04:50 GMT -5
Thanks for running those. They seem to show about 3db more high treble with the coil shorting avoided. It's a small amount, but should be noticeable and worth having. Isn't the red graph in this case exactly like A on my drawing above?
|
|
|
Post by JohnH on Apr 28, 2021 17:03:26 GMT -5
Thanks for running those. They seem to show about 3db more high treble with the coil shorting avoided. It's a small amount, but should be noticeable and worth having. Isn't the red graph in this case exactly like A on my drawing above? Yes I think it is, and your system avoids this damping. I was interested in the analysis in general, since the shorting type of single-coil cut is more common, and I have both types on various diagrams and in my own guitars.
|
|
|
Post by ms on Apr 28, 2021 18:34:01 GMT -5
So we have the unused coil (uuc) and used coil (uc). How does the uuc cause the changes in the frequency response of the uc? I see two possibilities, damping of the response of the uc from coupling to the uuc, and signal transfer from the uuc to the uc from the coupling. Let's define the meaning of these two possibilities by showing how to measure the effects. Use a tiny exciter coil that fits on top of a pole piece. This allows exciting one coil with almost nothing picked up in the other coil. So to measure damping we excite the uc and make measurements with the uuc open and shorted. (The uc has the usual 500pf cap to move the resonance down to something useful.) Significant damping would be indicated by appropriate differences between the measurements in the two cases. For the signal transfer case, we move the exciter coil over the the uuc, but continue to look at the output of the uc. With the uuc shorted, but excited, some current will flow in the uuc that will induce voltage in the uc. We look for a response, which, when added to the normal response of the uc, could cause the appropriate spectral changes. So the first two images are the damping test, uuc open and shorted. There is no significant difference, and so there does not seem to be any damping. The next image shows the signal transfer from coupling. This looks like what we need.
|
|
|
Post by aquin43 on Apr 29, 2021 2:44:17 GMT -5
So, with this last measurement by ms it all seems to come together. The shorted coil behaves in the same way as an eddy current loop. From the string, there is a direct path into the active coil and an indirect one via the current in the shorted coil and the mutual coupling of the coils. The second path has a low pass characteristic with an expected -3dB at omega = Rcoil/Lcoil, leading to a low pass step in the overall response. In addition, part of Rcoil and minus Lcoil are reflected into the used coil in a frequency dependent manner.
|
|
|
Post by guitarnerdswe on Apr 29, 2021 4:59:31 GMT -5
Isn't the red graph in this case exactly like A on my drawing above? Yes I think it is, and your system avoids this damping. I was interested in the analysis in general, since the shorting type of single-coil cut is more common, and I have both types on various diagrams and in my own guitars. Yeah I know you were more interested in the technical aspects of it. I'm just a noob at these more advanced things and I really appreciate the clarification 🙂
|
|