Humbucker Test - No Cover vs. Cover (bode plots and audio)

[gpdb]

I've finally received my testing guitar from Reddick Guitars, and I'm starting to create audio examples to go along with bode plots to crystalize how accurate and valuable they are. Audio has been the missing link for most people to understand the curves. You can see in this example how the audio really matches the change in the curve. 
I performed this test while keeping the humbucker in the exact same place, just removing the cover for the uncovered take. This is comparing a copper plated nickel silver cover vs. no cover on the same pickup.

[antigua]

That's really neat. It's really lets a person hear the difference as it appears in a bode plot. 

If you ever want to cheat to save a little time, or demonstrate the difference artificially, you can take either recording and just use a DAW with a multiband EQ and increase the amplitude at the resonant peak but however many dB the bode plot says it should be raised, and the outcome should be close the practical demonstration, and valid. All you have to keep in mind is that the dummy load might put the resonant peak at 3kHz, but your actual rig might place the resonant peak of the recording somewhere above or below that. Even if you start with a humbucker with a higher peak, like 3kHz, and you use the EQ to create a fake resonant bump around 2kHz, and then bring down everything past that frequency, you will end up with audio that sound convincingly like a humbucker with a low resonant peak, like an Seymour Duncan JB. It's valid for much the same reason as it would be to have lowered the resonant peak with a capacitor; the timbre is fixed, and the pickup is merely acting as an attenuator. 

I few years ago I used the FFT visualizer in Adobe Audition to change all of the particular harmonic amplitudes of a single coil neck in order to match the harmonics of a humbucker neck, and it sounded convincing guitarnuts2.proboards.com/thread/7968/pickup-sound-another-modern-magic . This is going beyond attenuation to changing the harmonics as well, as the harmonics and their varying amplitudes are what define the timbre. I suppose it served to demonstrate once again that guitar pickups are more mathematical than magical. If you're comparing like for like pickups, or obviously a pickup to itself, then a simple EQ shift is enough. 

If the only difference between pickups is EQ, one might think that an EQ pedal, or on board active EQ like you see with piezo pickups should put the entire aftermarket pickup industry out of business, but guitarists seem to prefer the vintage correctness and simplicity (lack of knobs to fiddle with) of vintage style hi-z pickups. There have been a few different attempts to create Swiss Army Knife pickups that can sound like anything, and they've never been too successful, such as this www.cycfi.com/projects/xr-series/ and the Dialtone creativeguitarstudio.blogspot.com/2014/05/dialtone-pickups-infinitely-tone.html

[gpdb]

I haven't seen that cycfi system, seems like a more hands on approach to what the fluence product does. The part I'm still struggling with is finding a way to measure asymmetrically wound humbuckers. I've been testing the Jazz and 59, which on the bode plots look nearly identical. However, through my own audio tests and all the ones I've listened to on the internet, they are clearly different. Not massively, but more than the bode plot would show. The Jazz (asymmetrical) is absolutely adding in some harmonics that aren't present in the 59 (symmetrical). I recorded two audio clips and tried to use a spectrum analyzer, but I couldn't export the data so it wasn't helpful. Is there any way to create a bode plot that can account for asymmetry?

[antigua]

What about the Jazz is asymmetrical? The Jazz I have showed the two coils being alike, less the difference to be expected when one coil as screws and the other has slugs. I was surprised buy how alike the 59 and Jazz were given how people describe them sounding different, but I've never been convinced that they actually sound different. The Jazz is usually heard in the context of a JB bridge, because they're offered as a pair as OEM pickups, where as the '59 is usually paired with a bridge '59, and it might be that the contrast increases the perceived difference.

My theory is that the Jazz might have been a '59 neck that used all the same parts as a JB, four conductor wire, modern plastic and wire, and no cover. Seymour Duncan offers a four conductor '59 neck which is nearly the same, but that might have not been the case when they launched the product. 

[gpdb]

I'll post the audio from my test. The asymmetry is very slight, but it's definitely there. Seymour Duncan even clarifies it in their description when they say it has a "special coil wind." Any time they mention this, it means there's an offset. My Jazz for instance has coils @ 3.71k/3.6k. Compare this to my Sentient neck which is 3.93k/3.95k. That's a more reasonable error, but 0.1K is definitely extra turns. Regardless it's obvious in the audio. The only part that the bode plot shows that could be accounting for the sound is that the 59 has a tiny bit more bass. But that doesn't seem like it would change the sound in the way I'm hearing it. 

[antigua]

Dec 6, 2022 13:58:16 GMT -5 gpdb said:

I'll post the audio from my test. The asymmetry is very slight, but it's definitely there. Seymour Duncan even clarifies it in their description when they say it has a "special coil wind." Any time they mention this, it means there's an offset. My Jazz for instance has coils @ 3.71k/3.6k. 

That degree of drift is typical in most humbuckers I measure, almost always at least 50 ohms apart, but occasionaly 100 ohms or more. I don't know for sure why it happens, but two possibilities are the manufacturing tolerances of the wire, as well as stretching of the wire that occurs as the coils are made, maybe a little of both. I'd recon that the turn counts are probably identical since a turn counters tend to be accurate, unless a human is operating the winder, and they fail to stop the winder right when it hits 4000 turns, or whatever.

In order to hear a difference between coils, there has to be a radical difference between the coils, such that it becomes half way towards turning into a single coil sound. I've wires splitting with a variable resister in order to variably load down one coil, and you can do that too, especially if you want to implicated it in a tonal difference. I don't think you really hear much difference until you get down to about half way, reducing the output of one of the coils by about half. It's a really useful mod for any humbucker that sounds too weak when one of the coils is ground out completely. 

Closely related, DiMarzio filed a patent for winding one coil with 43 AWG and the other with 42 AWG, or and different sized of wire, but the same number of turns, preserving the inductance per coil, and they claimed it produced a different sound, but the facts as I know them from all this testing and whatnot says the coils will perform the same if they have the same turn count. The substantial difference is the DC resistance, which doesn't have much impact, and the coil geometry, which is also not of much consequence. I think they patented something that doesn't make a difference, for whatever benefit they could derive from marketing a trivial difference, or maybe they really believe it sounded different. 

[gpdb]

Here is my own recording of the 59, Jazz, and Sentient. I think the chart does an excellent job showing how the Sentient is different from the others, but I still hear a slight "chime" with the Jazz that I don't hear with the 59. It could just be the slight increase in bass frequencies shown in the chart though - that's the only place where the Jazz and 59 are different. That changes the Q-factor slightly. Side note on my Q-factor number; I'm just subtracting the reasonant peak's db from the db at 100hz. I know that's not really what Q-factor is, but it's the easiest way to get a standard metric. 


This is a separate video on YouTube that I think really shows a stronger difference between the 59 and Jazz. I can't confirm how this person took their recordings, but again I can hear less bass in the Jazz, which I can only describe as "chime." You can definitely hear a massive difference when comparing the split coil samples in this video.

[gpdb]

Something strange I'm also seeing about the Jazz pickup is that the inductance and DCR aren't really matching up, at least compared to Sentient I have. 

Jazz
Series DCR: 7.4k
Series Inductance: 3.92H
Slug DCR: 3.65K
Slug Inductance: 1.74H
Screw DCR: 3.76K
Screw Inductance: 1.68H

Sentient
Series DCR: 7.83k

Series Inductance: 4.2H

Slug DCR: 3.93K

Slug Inductance: 1.79H

Screw DCR: 3.95K

Screw Inductance: 1.8H

What's strange is that the Jazz coils don't seem to correlate with DCR and inductance. The slug coil has a lower DCR but a higher inductance, while the screw has a higher DCR and a lower inductance. Visually, the screw coil appears to have less wire on it. That's my best guess as to what is different between the Jazz and 59, the screw coil is wound lower to remove some bass emphasis. My main issue is that my 59 isn't a 4 conductor, so I can't measure the coils individually without really taking it apart and I'd rather not do that. It's possible the slugs and screws are causing this change in inductance, but the Sentient is built the same and the DCR and inductance are equal on both. 

[antigua]

Dec 7, 2022 10:40:46 GMT -5 gpdb said:

Here is my own recording of the 59, Jazz, and Sentient. I think the chart does an excellent job showing how the Sentient is different from the others, but I still hear a slight "chime" with the Jazz that I don't hear with the 59. It could just be the slight increase in bass frequencies shown in the chart though - that's the only place where the Jazz and 59 are different. That changes the Q-factor slightly. Side note on my Q-factor number; I'm just subtracting the reasonant peak's db from the db at 100hz. I know that's not really what Q-factor is, but it's the easiest way to get a standard metric. 

Your plot shows a difference in the bass but it's well below 1dB so the science says you're not going to hear that difference.

When you put that they have peaks of 3.3 and 3.2 kHz, that's with the 470pF dummy load, but your recording have a real capacitance that is unknown, in the recording the resonant peak might be elsewhere. They sound a lot alike to me, though. The differences I hear sounds like it could be a combination of a slightly different resonant peak and plucking the strings with a slightly different manner.

When you test the humbuckers, are you hooking up the shielding to the test probe on the ground side? When you connect the shielding, it causes the screws and slugs to capacitively couple to the coils. You want to do that, since in real life you would connect the shielding and that capacitance exists. When you get it down to a headphone listening test, it can make a difference. Our ears are crude with respect to amplitude, but we can tell apart frequency very well. 

Are you using a guitar cable? If so, you might check the capacitance of the cable, the DE-5000 can measure that real easy. Your recordings sound fairly bright, if I were gambling I'd guess the real life capacitance is under 470pF.

If you are plugging in directly to an instrument input on a USB sound device, you can find the input impedance in the specs, that should be made note of also.  Typically they should be 1meg ohm, but it varies, sometimes higher sometimes lower. But this also highlights a problem with looking for precision, a sound demo will use a particular guitar cable, but some players are using wireless and others are using 20ft cables, so everyone is effectively hearing a slightly different pickup, while using the same pickup. 

Dec 7, 2022 10:40:46 GMT -5 gpdb said:

This is a separate video on YouTube that I think really shows a stronger difference between the 59 and Jazz. I can't confirm how this person took their recordings, but again I can hear less bass in the Jazz, which I can only describe as "chime." You can definitely hear a massive difference when comparing the split coil samples in this video.

I'm pretty sure the SH-1 he tests is the bridge model, where as the SH-2 is probably the neck model, so it's not really a functional comparison. Note that the SH-2 and the AlNiCo II Pro sound more or less identical, and they don't even use the same magnets.

[gpdb]

As far as the resonant peaks, both are pretty equal. Because I'm only measuring at 100hz intervals, it's rounding to whichever is the closest. So if there's a difference, its < 100hz. I agree that the mild bass difference should be negligible, but perhaps because it's raised over such a large frequency range, maybe it's more audible? The dummy load is 470pF/250K for each bode plot. The capacitance of my cable when it's connected to my amp is 500pF, so it's really close (my cable connected to my amp directly is 470pF exactly, but I connect it to my Fractal AxeFX 3 which then goes to my amp). Regardless, everything was recorded exactly the same. Any difference should be consistent. I agree both sound really similar, but when I'm going back and forth soloing them in my DAW, that "chime" is there on the Jazz. I even had my girlfriend listen to it and asked if she could hear a difference, and she came to the same conclusion. I recorded it with my Revv Generator 120 MKIII on the clean channel, everything pretty much at 12. To really be accurate I think I'll just need to get a 4 conductor version of the SH-1N.

When you say test, are you talking about the bode plot, in the guitar, or with the LCR meter? They were all connected to the guitar exactly the same. As far as the bode plots, I'm not sure. I wasn't aware there was a difference when connecting them in different ways as each plot looks correct when I use either side. I don't believe I was consistent on the LCR meter either. 

For the other differences you mentioned - I'm not sure how they're relevant if they were all recorded the same way. If someone was using a bad wireless unit, I would expect it to affect all the pickups equally.

It is possible he was using the bridge 59', but man that would be really dumb if he did. I agree with your point about the A2Pro sounding just as bright as the Jazz, so it's possible that's what's happening. But I've heard this difference in other examples on the internet too. 

[antigua]

I'm sure you can't hear it, it's being over a wide range would probably make it harder to notice since that detracts from frequency distinction. Another think to keep in mind if that if you're using a precise jig, and you're confident that the 59 has a higher output by ~0.3dB, not only is that difference really small, but even very slight differences in pickup height will cause a larger difference that 0.3dB, so I also wouldn't say the 59 is higher output in an practical sense. 

> Regardless, everything was recorded exactly the same. Any difference should be consistent
>For the other differences you mentioned - I'm not sure how they're relevant if they were all recorded the same way. If someone was using a bad wireless unit, I would expect it to affect all the pickups equally.

The larger the capacitance, the closer the resonant peaks become as they become lower, so if you had been using a wireless unit with very very low capacitance, the resonant peaks will be farther apart than they would be with the cable, and you'd likely hear that difference, so you might wonder why they sound different with no capacitance but look the same with a dummy load. But you're using a medium length cable, so that's probably OK. 

>  I agree both sound really similar, but when I'm going back and forth soloing them in my DAW, that "chime" is there on the Jazz. I even had my girlfriend listen to it and asked if she could hear a difference, and she came to the same conclusion. I recorded it with my Revv Generator 120 MKIII on the clean channel, everything pretty much at 12. To really be accurate I think I'll just need to get a 4 conductor version of the SH-1N.

I've tried making strumming machines a few times in order to take the human element out of it, and I tried dropping a weight and releasing rubber bands to actual the strum, and it seems no matter what I do, it's different every time, the transients end up looking obviously different in the DAW. That's to say I don't trust testing that involves strumming too much, not with a mechanical device but especially by hand. There's a lot of variables at play otherwise. Guitar pickups are fairly simple in that they don't move, their function is all electromagnetic, so it's basically all arithmetic, happening in space, as compared to speakers and microphones which are mechanical in nature, so the math is a lot more complicated. For those reasons, when you say the Jazz has more chime, my inclination is to look for errors in the testing rather than the conclude the Jazz innately has more chime. The extra chime I hear in the recording sounds like a resonant peak that is slid up to a higher frequency. You make a good point that maybe a difference within 100Hz is audible, but not presented in the graph.  

> When you say test, are you talking about the bode plot, in the guitar, or with the LCR meter? They were all connected to the guitar exactly the same. As far as the bode plots, I'm not sure. I wasn't aware there was a difference when connecting them in different ways as each plot looks correct when I use either side. I don't believe I was consistent on the LCR meter either.

Sorry, in the case of the 4 conductor wires, you have four wires and a shield, so you tie together the red and white wire, put the positive test probe to the black, and the ground probe to the green wire, but you also want to connect the ground probe to the shield, because that will include in the capacitance of the shielding, impacting capacitance measurements and bode plots, but it wont affect resistance or inductance. You can measure with and without the shield, but because pickups are almost always connected with shielding, I think it's best to include it, to have it connected along with the green wire to the ground probe. 

> It is possible he was using the bridge 59', but man that would be really dumb if he did. I agree with your point about the A2Pro sounding just as bright as the Jazz, so it's possible that's what's happening. But I've heard this difference in other examples on the internet too.

I'm pretty much certain, the 59 neck is called the SH-1n, but he lists the SH-1, and the bridge has an inductance of 5 henries, the lower resonant peak it fairly easy to discern. 

[heardsoundcircuits]

In a “perfect” humbucker the cancellation pushes all noise well out of the audio spectrum. As we introduce a mismatch the noise creeps back into the higher frequencies and then mid frequencies as the mismatch becomes greater, up to a point. I have no proof of this idea it is just a layman’s theory I thought could be plausible. What I don’t know is how wrong I am.

I would love to understand the physics and math involved in determining which harmonics are being attenuated (and subsequently which ones are being accentuated) over a range of coil mismatch values with all other components being hypothetically equal.

[newey]

Dec 16, 2022 18:30:04 GMT -5 heardsoundcircuits said:

I have no proof of this idea it is just a layman’s theory I thought could be plausible. What I don’t know is how wrong I am.

No, you're right. The noise cancellation happens becasue the noise signal from one coil is 180° out of phase with the other (but the pickup signal is nonetheless in-phase, that's what happens with one coil RWRP with respect to the other.) Imagine the signal visually, on an oscilloscope- where the peaks from one coil match perfectly to the troughs of the other, the noise is cancelled. In an idealized situation, if two coils were exactly identical (meaning you could lay the two frequency plots on top of one another and they would coincide exactly), then by putting one coil OOP with the other, you would in theory have no signal at all. But if we also reverse the magnet polarity-RWRP- we have then put our idealized coils back in phase- but the noise signal (because the noise has nothing to do with the magnets, only with the wire) is now out of phase. And, in our idealized situation, any noise would be cancelled entirely.

But IRL, no two coils are ever identical, and so cancellation is imperfect. As the coils become "less identical", fewer frequencies are common to both signals, so fewer get cancelled.

Whether that happens, as you say, first in the higher frequencies? I don't know. We hear higher frequncies better than lower, so that may just relate to where in the spectrum the noise is coming versus our hearing of it.

[antigua]

Dec 16, 2022 18:30:04 GMT -5 heardsoundcircuits said:

In a “perfect” humbucker the cancellation pushes all noise well out of the audio spectrum. As we introduce a mismatch the noise creeps back into the higher frequencies and then mid frequencies as the mismatch becomes greater, up to a point. I have no proof of this idea it is just a layman’s theory I thought could be plausible. What I don’t know is how wrong I am.

I would love to understand the physics and math involved in determining which harmonics are being attenuated (and subsequently which ones are being accentuated) over a range of coil mismatch values with all other components being hypothetically equal.

With respect to string harmonic cancellations, there's something pretty close to demonstrating the harmonic cancellations of a humbucker www.till.com/articles/PickupResponseDemo/ , you create two pickups with the demo, then put the pickups side by side. 

As for errant EMI noise, if the coils are identical, then the only cause of noise is the fact that the pickups are in different locations in space, so they receive different proportions of the same EMI, because one of the two coils is probably a little closer to the noise source than the other at any given moment. The classic PAF style has one coil with slugs and another with screws, so they're rarely identical.

As for whether the higher frequency of EMI gets through more so with a mismatch, that's an interesting question, because one of the two coils will have a higher self resonance than the other, so I suppose that difference represents a band of leakage, but I'm not really sure if it works out that way or not. 

One interesting thing is that the coil with the higher inductance will contribute more voltage (more sound from its portion of the guitar string) to the output than the coil of lower inductance when wired in series, but when wired in parallel, the coil with the lower inductance will produce more voltage, because of how the two wiring schemes impact the impedance of the coils.