Hantek 1833C LCR meter, great for pickups

[Antigua]

A few months ago we were talking about measuring the capacitance of pickups with LCR meters set to 100kHz, in Cp mode. A problem at the time was that the DE-5000, the most affordable full-featured LCR meter available, only has a high test frequency of 100kHz, and it's very possible for the guitar pickup to be anomalously inductive around 100kHz, because coils are imperfect in how they're wound. When that happens, the Cp measurement for capacitance at 100kHz gives a bogus result. But I've come across another meter that's also affordable, the Hantek 1833C LCR , selling for less than $180 on Amazon https://www.amazon.com/gp/product/B0...?ie=UTF8&psc=1 and it appears to be even cheaper from some other vendors, which has all the usual features of a good LCR meter, but has several more high frequency test points: 40kHz, 50kHz, 75kHz, and 100kHz, and with the total of 4 data points past the self-resonance of the pickup, anomalies in the coil can be spotted and discarded, giving the true capacitance of the pickup. The capacitance of most pickups is small enough that it doesn't have a big impact on the end result, but if you want to know the capacitance for it's own sake, this will reveal the value with a good degree of certainty.

It's a real nice piece of hardware, having a backlit screen, fast update rate, USB chargeable internal battery. If you're into pickup analysis, this is a good LCR meter to have on hand.

[Joe Gwinn]

I did look at this. It appears to be an extension of the DE-5000 (itself a General Radio design), with more test frequencies and a lithium-ion rechargeable battery replacing the 9v battery of yore..

I'm not sure that the added test frequencies will yield more accurate self-capacitance measurements, because these test frequencies are way out of the audio band.

I would suggest using Terman's Method (discussed in many postings from a few years ago) to measure the self-capacitance directly in band, for comparison.

[max oersted]

Who doesn't like a good cheap meter that does it all, yeah? Joe, I know this "ye olden" discussion and utilize it still to this day. But also as I get older, I become somewhat of a lazy fuck and wish for a meter that does it ALL. Who makes this? Anyone? In these modern times there's no reason why they shouldn't. I want this.

[Helmholtz]

I think that other than with PU inductance, there is little reason to assume that the real capacitance of a PU varies significantly with frequency.
Never saw any evidence for such effect using a sophisticated HP impedance analyzer.
.
Problem with LCR meters is that they typically measure apparent C, not real C - because of the simple 2-component equivalent circuit model they use for evaluation.

And apparent C strongly varies in the vicinity of a resonance.

Measuring at 100kHz gives good real capacitance results except in cases where there is a secondary resonance close to the measuring frequency.

[Joe Gwinn]

I think that other than with PU inductance, there is little reason to assume that the real capacitance of a PU varies significantly with frequency.
Never saw any evidence for such effect using a sophisticated HP impedance analyzer.
.
Problem with LCR meters is that they typically measure apparent C, not real C - because of the simple 2-component equivalent circuit model they use for evaluation.

And apparent C strongly varies in the vicinity of a resonance.

Agree.

Measuring at 100kHz gives good real capacitance results except in cases where there is a secondary resonance close to the measuring frequency.

Has this been validated by experiment? There are lots of 2nd-order effects that become significant at the higher frequencies that are undetectable in the audio band, and these effects will tend to confuse LCR meters.

[Helmholtz]

Has this been validated by experiment? There are lots of 2nd-order effects that become significant at the higher frequencies that are undetectable in the audio band, and these effects will tend to confuse LCR meters.

All I can state is that I typically (i.e. in the absence of a close-by secondary resonance) found good agreement between LCR meter measurements at 100kHz and the results of the HP impedance analyzer which fits the values of a 4-component EQCKT to the complete impedance and phase response data taken over a wide frequency range (40Hz to MHz if required).


Actually I don't care much about PU capacitance accuracy.
I don't think anyone would really notice much of a difference difference with even 50pF more or less with a total load of typically > 500pF.
Means a shift of PU resonance by 5% or less.

And of course some total load capacitance is necessary for the PUs to develop their characteristic sound and lower capacitance is not necessarily better.

[Joe Gwinn]

All I can state is that I typically (i.e. in the absence of a close-by secondary resonance) found good agreement between LCR meter measurements at 100kHz and the results of the HP impedance analyzer which fits the values of a 4-component EQCKT to the complete impedance and phase response data taken over a wide frequency range (40Hz to MHz if required).

Which model were you using?

Actually I don't care much about PU capacitance accuracy.
I don't think anyone would really notice much of a difference difference with even 50pF more or less with a total load of typically > 500pF.
Means a shift of PU resonance by 5% or less.

Well, the question was to measure the pickup's self capacitance, for use in pickup design. But it is certainly true that the cable will add a fair bit of capacitance.

And of course some total load capacitance is necessary for the PUs to develop their characteristic sound and lower capacitance is not necessarily better.

Well, tastes vary widely here.

[Helmholtz]

Which model were you using?

IIRC, what we had.was a HP 4194A.

[Joe Gwinn]

IIRC, what we had.was a HP 4194A.

Yes, that ought to do it, if one chooses the correct Equivalent Circuit, which would be "B" for pickups.

But these standbys still sell for thousands of US dollars for used units. But I recall there was one list denizen that had one, or the like.

[Helmholtz]

[SIZE=18px]

Yes, that ought to do it, if one chooses the correct Equivalent Circuit, which would be "B" for pickups.

One of the nice things was that it allowed direct comparison of the EQVCKT's impedance response with the measured one. Sadly I don't have access anymore.

But anyone having an LCR meter that measures at 100kHz could compare the results to Terman's method.
Problem with Terman's method is that it assumes a constant L. But real PU inductance drops with increasing frequency to different extent, so you won't get a straight line, which introduces some extrapolation uncertainty.

[Joe Gwinn]

But anyone having an LCR meter that measures at 100kHz could compare the results to Terman's method.
Problem with Terman's method is that it assumes a constant L. But real PU inductance drops with increasing frequency to different extent, so you won't get a straight line, which introduces some extrapolation uncertainty.

Actually, no. If L varies, the straight line becomes curved, so one will know, and also know how large the effect. This is one effect of eddy currents, which are not a lumped-element effect.

Hmm. How does the 4194A handle inductance varying with frequency?

[Helmholtz]

Hmm. How does the 4194A handle inductance varying with frequency?

Hmm, I guess not very well either. Would have to use some average value. But I think the analyzer doesn't need the inductance to calculate the capacitance, as capacitance is given by the decay of the impedance at frequencies far above the main resonance where the influence of inductance tends to zero.

As said, I don't consider PU self-capacitance a primary influencer of PU sound, so accuracy doesn't matter much to me.
Vintage type HBs and SCs typically have a self-capacitance between 50pF and 150pF.
This adds to cable capacitance and ampliffier input (Miller) capacitance of around 150pF - when directly connected to amp.

As a player I prefer guitar cables having a capacitance of around 1nF. Just don't like my LPs and strats with lower capacitance cables.
If I thought a PU had too much self-capacitance (or inductance) I would try a somewhat shorter/lower capacitance cable.

[Joe Gwinn]

Hmm, I guess not very well either. Would have to use some average value. But I think the analyzer doesn't need the inductance to calculate the capacitance, as capacitance is given by the decay of the impedance at frequencies far above the main resonance where the influence of inductance tends to zero.

Well, the inductance is involved in the computation, because inductive reactance cancels capacitive reactance, both varying with frequency. The variation with frequency allows one to isolate tha C and L contributions.

I'd venture that for pickup maker use, Terman's Method is more practical to perform than using the outputs from an impedance meter at various test frequencies.

As said, I don't consider PU self-capacitance a primary influencer of PU sound, so accuracy doesn't matter much to me.
Vintage type HBs and SCs typically have a self-capacitance between 50pF and 150pF.
This adds to cable capacitance and ampliffier input (Miller) capacitance of around 150pF - when directly connected to amp.

As a player I prefer guitar cables having a capacitance of around 1nF. Just don't like my LPs and strats with lower capacitance cables.
If I thought a PU had too much self-capacitance (or inductance) I would try a somewhat shorter/lower capacitance cable.

Yes, this kind of approach is typical in use.

[Helmholtz]

[SIZE=18px]

Well, the inductance is involved in the computation, because inductive reactance cancels capacitive reactance, both varying with frequency. The variation with frequency allows one to isolate tha C and L contributions.

Other than with a series resonant circuit, in a parallel resonant circuit capacitive and inductive reactances are not simply additive.

PU impedance is essentially represented by a parallel circuit of self-capacitance and inductance in series with the DCR.
Sometimes more elements are necessary for a representative model, but the rest of the circuit is always shunted by the capacitance.

As capacitive impedance drops with increasing frequency (and inductive impedance increases), total impedance at HF is dominated by the capacitance as with all parallel resonant circuits.
When the PU impedance curve eventually drops with -6dB/octave (or -20dB/decade) at high enough frequencies, impedance is almost purely capacitive and the capacitance value can be calculated from the asymptote.
With this method actual inductance doesn't matter.

[Joe Gwinn]

Other than with a series resonant circuit, in a parallel resonant circuit capacitive and inductive reactances are not simply additive.

PU impedance is essentially represented by a parallel circuit of self-capacitance and inductance in series with the DCR.
Sometimes more elements are necessary for a representative model, but the rest of the circuit is always shunted by the capacitance.

As capacitive impedance drops with increasing frequency (and inductive impedance increases), total impedance at HF is dominated by the capacitance as with all parallel resonant circuits.
When the PU impedance curve eventually drops with -6dB/octave (or -20dB/decade) at high enough frequencies, impedance is purely capacitive and the capacitance value can be calculated from the asymptote.
With this method actual inductance doesn't matter.

Aren't we going in circles here? Eddy currents in the nearby metal causes the inductance to vary with frequency, and inductance cancels capacitance, to a degree depending on frequency for two reasons, not the single reason assumed by LCR tests. And eddy current effects can only be approximated with lumped-parameter models. Better to just measure things directly.