Hantek 1833C LCR meter, great for pickups

[Helmholtz]

... inductance cancels capacitance

That only holds for series resonant circuits (and exactly only at resonance), where (opposite sign) inductive and capacitive reactances directly add.

Parallel resonant circuits are different and only capacitance is left at high enough frequencies.

[Antigua]

I do get good agreement between calculating the capacitance from the inductance and the resonant peak, and the values from the DE-5000 and the Hantek. In fact, the C measurement from the DE-5000 and the Hantek even correctly excluded the oscilloscope probe capacitance that resulted from the inductance by frequency calculation. Having measures maybe a hundred pickups the old way, and then a hundred the new way, I'd say the agreement is all around really strong.

As was discussed in another thread about three years ago, a coil that has slop in the winding pattern somehow forms coils within coils, and at certain frequencies the pickup become inductive again, and that was seen on impedance plots extending up to 100kHz. I assumed what happens is, for example, turn #3000 and turn #4000 will come very close together due to an uneven winding pattern, and there will be a capacitive short across that portion of coil at that point. Wholly machine made coils seem to be less likely to have those anomalies. One of those secondary inductances could occur near 100kHz, the highest test frequency of the DE-5000, how close of course depending on the Q factor, but roughly anywhere within 10kHz of the anomalous resonance is not "flat", and would throw it off.

The Hantek 1833C has four test frequencies well above the resonant peak of a typical pickup, all above 10kHz, so even if there is an anomalous resonance at 100kHz, there are still three other frequencies to sample at. When I test a pickup with it, the capacitance only varies by a small amount.

Here are some values for an SSL-1

100Hz --- -53.72nF
120Hz --- -52.48nF
400Hz --- -29.598nF
1kHz ---- -8.360nF
10kHz --- -507.7pF
40kHz --- 104.04pF
50kHz --- 108.54kHz
75kHz --- 106.05kHz
100kHz -- 109.50kHz

So you can see at 40kHz and beyond, the measurements are stable to within a few picofarads. There can be an anomaly in the coil and there will still be enough test points to get a good capacitance value.

[Helmholtz]

Negative apparent capacitance means inductive phase angle, i.e. it looks to the meter like an inductance.
That's typical for a parallel resonant circuit below resonance.
Of course, real capacitance is always positive.

[Antigua]

Negative apparent capacitance means inductive phase angle. That's typical for a parallel resonant circuit below resonance.
Of course, real capacitance is always positive

I just included that for completeness.

[Joe Gwinn]

I do get good agreement between calculating the capacitance from the inductance and the resonant peak, and the values from the DE-5000 and the Hantek. In fact, the C measurement from the DE-5000 and the Hantek even correctly excluded the oscilloscope probe capacitance that resulted from the inductance by frequency calculation. Having measures maybe a hundred pickups the old way, and then a hundred the new way, I'd say the agreement is all around really strong.

...

So you can see at 40kHz and beyond, the measurements are stable to within a few picofarads. There can be an anomaly in the coil and there will still be enough test points to get a good capacitance value.

So, we do have lab validation.

In the values provided by the various methods, do the errors of one compared to another method show any pattern to the eye?

[Antigua]

I should clarify, I didn't perform a hundred direct comparisons, but that I have come to be able to predict a pickup's capacitance based on the presence of shielded hookup wire, the coil geometry and the type of pickup, so when I started measuring pickups with the DE-5000 at 100kHz all of the values I was receiving look correct, but slightly lower, which is probably on account of the fact that probe capacitance isn't a factor when using an LCR meter. The capacitance measurement is so accurate in general that you can even seen it rise and fall if the DE-5000's leads get too close together, so I make a point of not allowing them to touch when I'm measuring the capacitance.

Since I recorded the resonant peak and the inductance and the capacitance separately for a lot of pickups, it would be possible to calculate C from the L and f, and from that it can be seen how much agreement there is between the two, but that would be quite a bit of work and I don't have cause to second guess the C values from the DE-5000.

It's possible that the aforementioned anomalies could result in bad readings at 100kHz, and I've come across a few pickups where the capacitance measurement was implausible, for example a reading of 500pF, and in that case I'd discard the measurement and have to calculate C from f and L, but thanks to the Hantek, I shouldn't even have that issue anymore. I'd say it's truly a one stop shop for documenting electrical values, because for whatever shortcomings it might have, it literally takes a tiny faction of the time it would take to gather the same values via oscilloscope. The only thing I'm less certain about is how useful the Q measurement is, say at 1kHz for determining how much eddy current damping a pickup causes. I've noticed that pickups with more steel parts and covers read a lower Q at 1kHz in general, but I'm not sure how representative it is of the sonic outcome.

[Joe Gwinn]

I should clarify, I didn't perform a hundred direct comparisons, but that I have come to be able to predict a pickup's capacitance based on the presence of shielded hookup wire, the coil geometry and the type of pickup, so when I started measuring pickups with the DE-5000 at 100kHz all of the values I was receiving look correct, but slightly lower, which is probably on account of the fact that probe capacitance isn't a factor when using an LCR meter. The capacitance measurement is so accurate in general that you can even seen it rise and fall if the DE-5000's leads get too close together, so I make a point of not allowing them to touch when I'm measuring the capacitance.

Since I recorded the resonant peak and the inductance and the capacitance separately for a lot of pickups, it would be possible to calculate C from the L and f, and from that it can be seen how much agreement there is between the two, but that would be quite a bit of work and I don't have cause to second guess the C values from the DE-5000.

Ahh. OK.

It's possible that the aforementioned anomalies could result in bad readings at 100kHz, and I've come across a few pickups where the capacitance measurement was implausible, for example a reading of 500pF, and in that case I'd discard the measurement and have to calculate C from f and L, but thanks to the Hantek, I shouldn't even have that issue anymore. I'd say it's truly a one stop shop for documenting electrical values, because for whatever shortcomings it might have, it literally takes a tiny faction of the time it would take to gather the same values via oscilloscope. The only thing I'm less certain about is how useful the Q measurement is, say at 1kHz for determining how much eddy current damping a pickup causes. I've noticed that pickups with more steel parts and covers read a lower Q at 1kHz in general, but I'm not sure how representative it is of the sonic outcome.

I would think that it would be worthwhile to dig into a unit showing such an implausible capacitance.

The Q has two practical uses. First, the width of the resonance curve is set by the Q. If the Q is too high, the pickup will be shrill at resonance, if this is in the audible band. Second, coils with very low Q values (compared to the run of production) probably have a shorted turn or two. The more direct measure is the AC eddy-loss resistance (measured AC resistance minus DC resistance).

[Helmholtz]

Some info on the relation between inductor Q and resonance Q:
https://www.gitec-forum-eng.de/wp-co...E_Q-Factor.pdf

[Antigua]

[SIZE=16px]
I would think that it would be worthwhile to dig into a unit showing such an implausible capacitance.

This is what it looks like in a bode plot, this pickup has two additional resonances, one peaks at 86kHz and the other at 153hHz. It would be hard to look at the coil and find the cause, because it just looks like a coil inside and out, but what I suspect happens is that suppose you wind the coil so that its "clumped" on the left, and then you move the wire to the other side of the bobbin and clump is to the left, then what happens is that the capacitance isn't evenly distributed by layer, and there's a portion of coil, that "clump" that has an particularly high degree of capacitance across is.

The Q has two practical uses. First, the width of the resonance curve is set by the Q. If the Q is too high, the pickup will be shrill at resonance, if this is in the audible band. Second, coils with very low Q values (compared to the run of production) probably have a shorted turn or two. The more direct measure is the AC eddy-loss resistance (measured AC resistance minus DC resistance).

AC resistance - DC resistance, that seems like a very good idea. I suppose I have to measure it at 1kHz, I don't think there is a more suitable frequency available on the common LCR meters. I'll grab an assortment of pickups and see how the reactance varies in ohms.

[Joe Gwinn]

This is what it looks like in a bode plot, this pickup has two additional resonances, one peaks at 86kHz and the other at 153hHz. It would be hard to look at the coil and find the cause, because it just looks like a coil inside and out, but what I suspect happens is that suppose you wind the coil so that its "clumped" on the left, and then you move the wire to the other side of the bobbin and clump is to the left, then what happens is that the capacitance isn't evenly distributed by layer, and there's a portion of coil, that "clump" that has an particularly high degree of capacitance across is.

Main peak is at 10 KHz or so. I bet that Terman's method will get the correct self-capacitance, because it will never see all that business above its highest test frequency, never mind 10 KHz.

I have no idea if the clumped-winding theory is correct, but I'd expect that a clumped winding effect would be far more subtle that the bode plot shows.

AC resistance - DC resistance, that seems like a very good idea. I suppose I have to measure it at 1kHz, I don't think there is a more suitable frequency available on the common LCR meters. I'll grab an assortment of pickups and see how the reactance varies in ohms.

If there is a hard short between turns, the effect won't be subtle. Testing at 1 KHz ought to work.

[Antigua]

Main peak is at 10 KHz or so. I bet that Terman's method will get the correct self-capacitance, because it will never see all that business above its highest test frequency, never mind 10 KHz.

I think I understand what you're saying, but I think that idea is contradicted by the fact that it measures nearly the same capacitance at multiple frequencies:

40kHz --- 104.04pF
50kHz --- 108.54pF
75kHz --- 106.05pF
100kHz -- 109.50pF

the reading is not distorting in any particular direction as frequency rises, except to increase slightly overall.

I have no idea if the clumped-winding theory is correct, but I'd expect that a clumped winding effect would be far more subtle that the bode plot shows.

I'm open to suggestion on the cause of the higher capacitances, but the only similar effect I know of is where a humbucker is measured, and you see a secondary resonance of the two coils resonating individually, which occurs at a much lower frequency. It could also be that the wire that is wrapped around the eyelets are forming tiny series inductors. Maybe that's more likely.

[Joe Gwinn]

I think I understand what you're saying, but I think that idea is contradicted by the fact that it measures nearly the same capacitance at multiple frequencies:

40kHz --- 104.04pF
50kHz --- 108.54pF
75kHz --- 106.05pF
100kHz -- 109.50pF

the reading is not distorting in any particular direction as frequency rises, except to increase slightly overall.

I'm confused. I thought we were talking about those pickups that yielded about 500 pF of capacitance well above 10 KHz.

I'm open to suggestion on the cause of the higher capacitances, but the only similar effect I know of is where a humbucker is measured, and you see a secondary resonance of the two coils resonating individually, which occurs at a much lower frequency. It could also be that the wire that is wrapped around the eyelets are forming tiny series inductors. Maybe that's more likely.

I would not think that a few loops in an eyelet could possibly produce sufficient inductance to do any such thing.

Inductance of a solenoid is proportional to the product of core cross sectional area times the square of turns count, divided by the length of the coil along the core. The cross-sectional area and coil length terms tend to cancel. Just on turns alone, the inductance ratio would be something like (5/5000)^2= a million to one.

[Antigua]

I'm confused. I thought we were talking about those pickups that yielded about 500 pF of capacitance well above 10 KHz.

It measures high if there is a secondary resonance at the test frequency. 500pF+ is just an error on the part of the LCR meter because the pickup is not behaving capacitively at that particular frequency. Since we know that no Strat pickup measures 500pF, unless is soaked in water, then it's for sure that the reading is in error, but now with 40kHz, 50 and 75 in addition to 100kHz, there are ways to get around that. Since the four high test frequencies have strong agreements, within 5pF, I think it's safe to say the value is stable with respect to test frequency, and not dependent on it. I don't know what it would measure at 500kHz or 1 MHz, but these four test points seem to be working well.

I would not think that a few loops in an eyelet could possibly produce sufficient inductance to do any such thing.

Inductance of a solenoid is proportional to the product of core cross sectional area times the square of turns count, divided by the length of the coil along the core. The cross-sectional area and coil length terms tend to cancel. Just on turns alone, the inductance ratio would be something like (5/5000)^2= a million to one.

In some of the eyelets of my test pickups I would loop it about ten times, I'm not sure what the inductance works out to. The solder in the eyelet might short out the turns of wire, too. That was just a theory. I know for a fact that the secondary resonances are there, it's recorded data, but I don't know what causes them to be there. It seems to be more likely to happen with home made pickups, and less frequent in mass produces machine made pickups.

[Helmholtz]

[SIZE=12px]


It measures high if there is a secondary resonance at the test frequency. 500pF+ is just an error on the part of the LCR meter because the pickup is not behaving capacitively at that particular frequency. Since we know that no Strat pickup measures 500pF, unless is soaked in water, then it's for sure that the reading is in error, but now with 40kHz, 50 and 75 in addition to 100kHz, there are ways to get around that. Since the four high test frequencies have strong agreements, within 5pF, I think it's safe to say the value is stable with respect to test frequency, and not dependent on it. I don't know what it would measure at 500kHz or 1 MHz, but these four test points seem to be working well.

If you still have the PU from post #25, why not re-measure it with the Hantek?
Also using Termans method for reference would be a good idea. Should be easy using the Velleman Bode plotter and a few caps.
I see a chance to stop this seemingly never ending discussion about C-measuring methods and the "real" C.

[Joe Gwinn]

It measures high if there is a secondary resonance at the test frequency. 500pF+ is just an error on the part of the LCR meter because the pickup is not behaving capacitively at that particular frequency. Since we know that no Strat pickup measures 500pF, unless is soaked in water, then it's for sure that the reading is in error, but now with 40kHz, 50 and 75 in addition to 100kHz, there are ways to get around that. Since the four high test frequencies have strong agreements, within 5pF, I think it's safe to say the value is stable with respect to test frequency, and not dependent on it. I don't know what it would measure at 500kHz or 1 MHz, but these four test points seem to be working well.

So the theory revolves around extra resonances. This can be seen most clearly in a phase plot.

In some of the eyelets of my test pickups I would loop it about ten times, I'm not sure what the inductance works out to. The solder in the eyelet might short out the turns of wire, too. That was just a theory. I know for a fact that the secondary resonances are there, it's recorded data, but I don't know what causes them to be there. It seems to be more likely to happen with home made pickups, and less frequent in mass produces machine made pickups.

The ratio of the inductances is easily estimated, if not the actual inductances in H. Basically, it's the square of the turns ratios. So for ten turns (eyelet loops) and ten thousand turns (main winding), the ratio of turns is a thousand, and the inductance ratio is a million. So, the main coil dominates. It would take very good equipment to even see such an effect by measuring inductance.