Post by antigua on Jan 1, 2018 21:48:44 GMT -5
The "L" in LCR meters...
The aftermarket guitar pickup industry trades almost exclusively in DC resistance measurements when attempting to meaningfully compare pickups, and that's not terrible, because this value can lead you to a calculation of how much wire was wound onto the pickup, if you're also able to guess the gauge of wire that was used.
But the more "perfect" measurement the tell one pickup apart from another is the inductance. From a circuit topology standpoint, a pickup is an inductor and a resistor in series, with a capacitor across both:
Though a pickup doesn't literally have a resistor or a capacitor under the cover, it has (usually unwanted) resistances and capacitances that can be accurately modeled as if it did actually have a resistor and a capacitor under the cover. The coil(s) of wire are the inductors, the several thousand feet of tiny wire is the resistor, and the proximity of that wire to itself is the capacitor. This circuit topology results in what is called a "low pass filter with resonance", and when you put an signal (moving guitar strings) into this type of filter, you get a transfer function that looks like this:
The reason this curve arises is because, below a certain frequency, the inductor, or the pickup's coil(s), is "inductive", and above a certain frequency, the inductor / coil(s) is "capacitive", and at that frequency, it "resonates", or puts up a lot of resistance. The amount of resonance that you get is inversely proportional to the resistance caused by the long length of the copper wire, as well as eddy current losses that are caused by steel parts in and around the coil(s) of the guitar pickup.
Here's is an actual bode plot I've created of three Fender CS 69 pickups, you can see that the response curve conforms to that the transfer function depicted above:
So you've got inductance (L), capacitance (C) and resistance (R) working together to define the transfer function of a guitar pickup. Why is the inductance so important when you have C and R to contend with? The reason is because, for a given pickup type, C and R are relatively fixed values, where as L inductance changes wildly from one pickup to the next. Most pickup has a capacitance between 100 and 200 picofarads , and a resistance between 6 and 15 kiloohms. The significance of that rather small capacitance is mitigated by the large capacitance of the guitar cable (around 400pF). The differences that exist with respect to DC resistance is mitigated by the parallel resistance of the volume and tone pots in the guitar. Therefore, the ranges of R and C are fairly narrow in situ, but then you consider that the inductance can vary widely, and it's not mitigated by anything. For example, you can measure as low as 2 henries for a Gretsch Filter'tron, or as high as 8 henries for a Seymour Duncan JB, and beyond.
This difference in inductance not only varies the output voltage generated by the pickup by a wide margin, but it also causes the resonant peak to vary from 2kHz up to 5kHz, which represents a substantial difference in presence and treble. When you think about the difference between and SSL-1 and and SSL-3 or an SSL-5, it comes down to a wide variance of inductance, and little else. The case is similar with "P.A.F." replicas, where nearly all the same components are used to assemble the pickup, and the turns of wire on the coils, or the inductance, is the only substantial feature that sets one apart from the next.
As a practical example of how inductance is more useful than resistance, suppose you have two nearly identical Tele or Strat pickups, and one has a DC resistance that is 500 ohms higher than the other. You might assume that the pickup with the higher DC resistance is the hotter pickup. But what you might not realize is that the pickup with lower resistance has AlNiCo 2 pole pieces which, due to a higher magnetic permeability, actually cause it to have a higher inductance, and in fact become the hotter pickup. A good LCR meter will reveal that difference, where has a DC resistance measurement wouldn't. There are a handful of such reasons why the inductance and the DC resistance diverge.
tl;dr, inductance indicates how loud and dark a pickups will be, far more accurately than DC resistance.
Different LCR meters options...
The popular LCR meter among pickup testers and makers has been the Extech 380193, and it costs around $180. The high cost is part of the reason that inductance values are not more widespread. Additionally, a lot of people might not be sure how to configure the LCR meter to give a good result, but I'll cover that below.
Part of the reason I'm doing this write up is because I came across a cheaper alternative to the Extech that works really well, the Global Specialties LCR-58 "tweezer LCR meter", costing about $90 on Amazon.
The important quality of the Extech and G.S. LCR-58 is that they have "series" and "parallel" test modelling modes, and you can set the test frequency. In general, you want low test frequencies, because the high eddy currents caused by guitar pickups with steel parts are substantial, so substantial that as the test frequencies become higher, they will fool the LCR meter into making a bad calculation of inductance. Most LCR meters should have several test frequencies, as low as 100Hz and as high as 10kHz. The Extech has only 120Hz and 1kHz, where as the LCR-58 tweezer meter has 100Hz, 120Hz, 1kHz and 10kHz. Since most guitar pickups have a resonant peak below 10kHz, it's useless to attempt to determine inductance at that frequency anyway, since the coil behaves like a capacitor beyond that point.
The other important aspect of these LCR meters is that they have equivalent series (SER) and parallel (PAL) modes. What this is asking is "is the real resistance best modeled as being in series or in parallel with the reactance?" The logic behind which mode is most appropriate for a given circumstance is confusing, and to be honest I don't have the firmest grip on the matter, but one thing that is reasonably well established is that guitar pickups seems to be best modeled as a series resistor and inductor. This circuit diagram below is borrowed from one of Helmuth Lemme's blog posts, he's also written a book about how guitar pickups work:
Good results seem to come from running LCR meters in series "SER" mode, to indicate that the model above is the one we are assuming applies.
There are some DMM's on the market, or "digital multi meters", that also have built in inductance testers, and the upside to these meters is that 1) they are very cheap, and 2) they are able to measure standard DC resistance as well as inductance, and sometimes capacitance (though never the capacitance of a pickup, since it's not technically a capacitor). The downside is that the inductance test mode generally has no selectable test frequency, not a series or parallel equivalent circuit selector. It's not even clear to me how their method of determining inductance differs from a traditional LCR meter.
Some measurements...
So I have three meters on hand, one DMM and two LCR meters, all featuring a capacity to measure inductance. They are the UNI-T UT58D digital multi meter, the Extech 380193 LCR meter, the Global Specialties LCR-58 tweezer LCR meter.
I have five pickups out for testing, a Seymour Duncan SSL-1, SSL-3, a 59' neck, a JB, and a Fender Fidel'tron Filter'tron clone.
What is known going in is that the SSL-1 and SSL-3 are Fender style single coils with no steel parts, and only AlNiCo pole pieces, meaning that they have low eddy current involvement. On the other hand, the three others are all humbuckers with steel parts, such as screws, slugs and "keeper bars", and do have significant eddy current involvement.
Here are measurements of the five pickups, using the three different meters. For the LCR meters, each available test frequency is used, and in both cases they are set to "series" mode. The DMM has neither a SER/PAL mode, nor a test frequency setting, so it's one and only measurement value is listed by itself.
The table making code for these forums is too tedious to mess with, so I just took a screen shot of my spread sheet:
Comparing just the to true LCR meters, it can be seen that they are in most agreement when both are set to the 100Hz or 120Hz modes, but when in 1kHz mode, there is much less agreement, and of course the 10kHz mode of the LCR-58 yields completely useless values.
Note how the single coils, the SSL-1 and SSL-3 show relatively little different in measurement between 100kHz, 120Hz and 1kHz. This reflects the fact that the low eddy current losses allow for the higher test frequency of 1kHz to be somewhat effective. Note that the super-high inductance SSL-3 does show more divergence than the more typical SSL-1.
Compare the effect of the test frequency in relation to the three humbuckers and their various steel parts, and it can be seen that when the test frequency is increased to 1kHz, they all vary by a few hundred henries, with the '59 and the Fideli'tron being especially divergent. This goes to show that, while 1kHz test frequencies might be usable for Fender style single coils, you need a low test frequency if your pickup has steel parts, which accounts for most other pickup designs on the market.
Overall, a lower test frequency makes for more consistent measurements with guitar pickups, but how about the $30 "digital multi meter" with an inductance measurement option? Surprisingly, it's in the least agreement with the LCR meters when measuring the the SSL-1, showing about 10% more inductance, but it was remarkably close to the mark when measuring the humbuckers, showing less than 5% difference. It baffles me how it seems to get a closer agreement when it comes to the eddy current heavy humbuckers, and less when it comes to the more "pure" Fender style single coils. Though it seems to be OK for getting rough numbers from humbucking pickups, it doesn't seem to be the most accurate, especially if you're measuring style Fender single coils.
Also note, as an aside, that the SSL-3 has nearly the same inductance as the JB, and that the Fideli'tron has even less inductance than the SSL-1. The in-situ resonant peaks of these pickups correspond closely with their inductance values, and so they have surprisingly similar transfer functions , even though people tend not to associate them otherwise.
The "C" in LCR meter...
So this thing is supposed to measure capacitance also, seems like a two for one deal, but sadly an LCR meter can't measure the intrinsic capacitance of an inductor, because a unlike an actual capacitor, an inductor has electrical continuity, which interferes with the the LCR meter's ability to measure capacitance. It will show a capacitance value that is orders of magnitude beyond what it actually is. However, the capacitance mode can be used to measure the capacitance of guitar cables, which is a very useful application, as the guitar cable introduces more capacitance to the passive circuit than another other one component, including the pickup itself.
The best way to determine the capacitance of a guitar pickup is to find the resonant peak of the pickup with an oscilloscope and a function generator, find the inductance with an LCR meter, or through other means, and then solve for capacitance with XC = 1 / (2 · π · f · C) , or do as I do and use an online calculator.
UPDATE: it turns out you can measure pickup capacitance with LCR meters that feature test frequencies of 100kHz or higher. More on that here guitarnuts2.proboards.com/thread/8271/measuring-pickups-capacitance-lcr-meter LCR meters such as the DER EE DE-5000 which feature a 100kHz test frequency are able to reliably measure a pickup's intrinsic capacitance. I'm told that it's preferable to set the LCR meter to "PAL" mode, but it appears that the capacitance readings are fairly accurate in both SER and PAL mode for both humbuckers and Fender type single coils.
The "R" in LCR meter...
Even though LCR meters measure resistance, as is implied by the "R", it's not "DC resistance", but rather"AC resistance", and though "AC resistance" is interesting for other reasons, it's not to be confused for "DC resistance".
An interesting thing about AC resistance is that when you increase the test frequency, and the resistance climbs, it indicates eddy current losses are at work. Eddy currents are essentially modeled as resistance that rises with frequency. For example, an SSL-1 with AlNiCo slugs has rather low eddy currents. The standard multi-meter shows 6.72k ohms resistance for the SSL-1 I have on hand, but the Extech LCR meter shows an AC resistance of 6.71k ohms at 120Hz, and 6.874k ohms at 1kHz. Not much difference, not much eddy current. Now try it with the Fender Fideli'tron, which with its sizeable steel parts causing significant eddy current losses. The Fideli'tron shows a DC resistance of 4.46k ohms with a typical multi-meter, but the Extech shows a resistance of 4.521k ohms at 120Hz, a little higher, abd 7.112k ohms at 1kHz... a lot higher, indicating that eddy currents are "pushing back".
This isn't a good way to quantify eddy current resistance, though, and just by looking at a pickup and seeing how much steel is in and around the coils, you can get just as good of an idea how extreme the eddy current losses will be. The best way to see exactly how the eddy currents rise with frequency is to create an impedance bode plot with a network analyzer, such as the Velleman PCSGU250. When plotting a pickup with an oscilloscope and function generator, you should see a +6dB per octave slope before the resonance, but eddy currents will cause it to dip below +6dB per octace. The degree of eddy current losses will be proportionate to the downward divergence from +6dB/oct as the test frequency rises.
Pics:
The aftermarket guitar pickup industry trades almost exclusively in DC resistance measurements when attempting to meaningfully compare pickups, and that's not terrible, because this value can lead you to a calculation of how much wire was wound onto the pickup, if you're also able to guess the gauge of wire that was used.
But the more "perfect" measurement the tell one pickup apart from another is the inductance. From a circuit topology standpoint, a pickup is an inductor and a resistor in series, with a capacitor across both:
Though a pickup doesn't literally have a resistor or a capacitor under the cover, it has (usually unwanted) resistances and capacitances that can be accurately modeled as if it did actually have a resistor and a capacitor under the cover. The coil(s) of wire are the inductors, the several thousand feet of tiny wire is the resistor, and the proximity of that wire to itself is the capacitor. This circuit topology results in what is called a "low pass filter with resonance", and when you put an signal (moving guitar strings) into this type of filter, you get a transfer function that looks like this:
The reason this curve arises is because, below a certain frequency, the inductor, or the pickup's coil(s), is "inductive", and above a certain frequency, the inductor / coil(s) is "capacitive", and at that frequency, it "resonates", or puts up a lot of resistance. The amount of resonance that you get is inversely proportional to the resistance caused by the long length of the copper wire, as well as eddy current losses that are caused by steel parts in and around the coil(s) of the guitar pickup.
Here's is an actual bode plot I've created of three Fender CS 69 pickups, you can see that the response curve conforms to that the transfer function depicted above:
So you've got inductance (L), capacitance (C) and resistance (R) working together to define the transfer function of a guitar pickup. Why is the inductance so important when you have C and R to contend with? The reason is because, for a given pickup type, C and R are relatively fixed values, where as L inductance changes wildly from one pickup to the next. Most pickup has a capacitance between 100 and 200 picofarads , and a resistance between 6 and 15 kiloohms. The significance of that rather small capacitance is mitigated by the large capacitance of the guitar cable (around 400pF). The differences that exist with respect to DC resistance is mitigated by the parallel resistance of the volume and tone pots in the guitar. Therefore, the ranges of R and C are fairly narrow in situ, but then you consider that the inductance can vary widely, and it's not mitigated by anything. For example, you can measure as low as 2 henries for a Gretsch Filter'tron, or as high as 8 henries for a Seymour Duncan JB, and beyond.
This difference in inductance not only varies the output voltage generated by the pickup by a wide margin, but it also causes the resonant peak to vary from 2kHz up to 5kHz, which represents a substantial difference in presence and treble. When you think about the difference between and SSL-1 and and SSL-3 or an SSL-5, it comes down to a wide variance of inductance, and little else. The case is similar with "P.A.F." replicas, where nearly all the same components are used to assemble the pickup, and the turns of wire on the coils, or the inductance, is the only substantial feature that sets one apart from the next.
As a practical example of how inductance is more useful than resistance, suppose you have two nearly identical Tele or Strat pickups, and one has a DC resistance that is 500 ohms higher than the other. You might assume that the pickup with the higher DC resistance is the hotter pickup. But what you might not realize is that the pickup with lower resistance has AlNiCo 2 pole pieces which, due to a higher magnetic permeability, actually cause it to have a higher inductance, and in fact become the hotter pickup. A good LCR meter will reveal that difference, where has a DC resistance measurement wouldn't. There are a handful of such reasons why the inductance and the DC resistance diverge.
tl;dr, inductance indicates how loud and dark a pickups will be, far more accurately than DC resistance.
Different LCR meters options...
The popular LCR meter among pickup testers and makers has been the Extech 380193, and it costs around $180. The high cost is part of the reason that inductance values are not more widespread. Additionally, a lot of people might not be sure how to configure the LCR meter to give a good result, but I'll cover that below.
Part of the reason I'm doing this write up is because I came across a cheaper alternative to the Extech that works really well, the Global Specialties LCR-58 "tweezer LCR meter", costing about $90 on Amazon.
The important quality of the Extech and G.S. LCR-58 is that they have "series" and "parallel" test modelling modes, and you can set the test frequency. In general, you want low test frequencies, because the high eddy currents caused by guitar pickups with steel parts are substantial, so substantial that as the test frequencies become higher, they will fool the LCR meter into making a bad calculation of inductance. Most LCR meters should have several test frequencies, as low as 100Hz and as high as 10kHz. The Extech has only 120Hz and 1kHz, where as the LCR-58 tweezer meter has 100Hz, 120Hz, 1kHz and 10kHz. Since most guitar pickups have a resonant peak below 10kHz, it's useless to attempt to determine inductance at that frequency anyway, since the coil behaves like a capacitor beyond that point.
The other important aspect of these LCR meters is that they have equivalent series (SER) and parallel (PAL) modes. What this is asking is "is the real resistance best modeled as being in series or in parallel with the reactance?" The logic behind which mode is most appropriate for a given circumstance is confusing, and to be honest I don't have the firmest grip on the matter, but one thing that is reasonably well established is that guitar pickups seems to be best modeled as a series resistor and inductor. This circuit diagram below is borrowed from one of Helmuth Lemme's blog posts, he's also written a book about how guitar pickups work:
Good results seem to come from running LCR meters in series "SER" mode, to indicate that the model above is the one we are assuming applies.
There are some DMM's on the market, or "digital multi meters", that also have built in inductance testers, and the upside to these meters is that 1) they are very cheap, and 2) they are able to measure standard DC resistance as well as inductance, and sometimes capacitance (though never the capacitance of a pickup, since it's not technically a capacitor). The downside is that the inductance test mode generally has no selectable test frequency, not a series or parallel equivalent circuit selector. It's not even clear to me how their method of determining inductance differs from a traditional LCR meter.
Some measurements...
So I have three meters on hand, one DMM and two LCR meters, all featuring a capacity to measure inductance. They are the UNI-T UT58D digital multi meter, the Extech 380193 LCR meter, the Global Specialties LCR-58 tweezer LCR meter.
I have five pickups out for testing, a Seymour Duncan SSL-1, SSL-3, a 59' neck, a JB, and a Fender Fidel'tron Filter'tron clone.
What is known going in is that the SSL-1 and SSL-3 are Fender style single coils with no steel parts, and only AlNiCo pole pieces, meaning that they have low eddy current involvement. On the other hand, the three others are all humbuckers with steel parts, such as screws, slugs and "keeper bars", and do have significant eddy current involvement.
Here are measurements of the five pickups, using the three different meters. For the LCR meters, each available test frequency is used, and in both cases they are set to "series" mode. The DMM has neither a SER/PAL mode, nor a test frequency setting, so it's one and only measurement value is listed by itself.
The table making code for these forums is too tedious to mess with, so I just took a screen shot of my spread sheet:
Comparing just the to true LCR meters, it can be seen that they are in most agreement when both are set to the 100Hz or 120Hz modes, but when in 1kHz mode, there is much less agreement, and of course the 10kHz mode of the LCR-58 yields completely useless values.
Note how the single coils, the SSL-1 and SSL-3 show relatively little different in measurement between 100kHz, 120Hz and 1kHz. This reflects the fact that the low eddy current losses allow for the higher test frequency of 1kHz to be somewhat effective. Note that the super-high inductance SSL-3 does show more divergence than the more typical SSL-1.
Compare the effect of the test frequency in relation to the three humbuckers and their various steel parts, and it can be seen that when the test frequency is increased to 1kHz, they all vary by a few hundred henries, with the '59 and the Fideli'tron being especially divergent. This goes to show that, while 1kHz test frequencies might be usable for Fender style single coils, you need a low test frequency if your pickup has steel parts, which accounts for most other pickup designs on the market.
Overall, a lower test frequency makes for more consistent measurements with guitar pickups, but how about the $30 "digital multi meter" with an inductance measurement option? Surprisingly, it's in the least agreement with the LCR meters when measuring the the SSL-1, showing about 10% more inductance, but it was remarkably close to the mark when measuring the humbuckers, showing less than 5% difference. It baffles me how it seems to get a closer agreement when it comes to the eddy current heavy humbuckers, and less when it comes to the more "pure" Fender style single coils. Though it seems to be OK for getting rough numbers from humbucking pickups, it doesn't seem to be the most accurate, especially if you're measuring style Fender single coils.
Also note, as an aside, that the SSL-3 has nearly the same inductance as the JB, and that the Fideli'tron has even less inductance than the SSL-1. The in-situ resonant peaks of these pickups correspond closely with their inductance values, and so they have surprisingly similar transfer functions , even though people tend not to associate them otherwise.
The "C" in LCR meter...
So this thing is supposed to measure capacitance also, seems like a two for one deal, but sadly an LCR meter can't measure the intrinsic capacitance of an inductor, because a unlike an actual capacitor, an inductor has electrical continuity, which interferes with the the LCR meter's ability to measure capacitance. It will show a capacitance value that is orders of magnitude beyond what it actually is. However, the capacitance mode can be used to measure the capacitance of guitar cables, which is a very useful application, as the guitar cable introduces more capacitance to the passive circuit than another other one component, including the pickup itself.
The best way to determine the capacitance of a guitar pickup is to find the resonant peak of the pickup with an oscilloscope and a function generator, find the inductance with an LCR meter, or through other means, and then solve for capacitance with XC = 1 / (2 · π · f · C) , or do as I do and use an online calculator.
UPDATE: it turns out you can measure pickup capacitance with LCR meters that feature test frequencies of 100kHz or higher. More on that here guitarnuts2.proboards.com/thread/8271/measuring-pickups-capacitance-lcr-meter LCR meters such as the DER EE DE-5000 which feature a 100kHz test frequency are able to reliably measure a pickup's intrinsic capacitance. I'm told that it's preferable to set the LCR meter to "PAL" mode, but it appears that the capacitance readings are fairly accurate in both SER and PAL mode for both humbuckers and Fender type single coils.
The "R" in LCR meter...
Even though LCR meters measure resistance, as is implied by the "R", it's not "DC resistance", but rather"AC resistance", and though "AC resistance" is interesting for other reasons, it's not to be confused for "DC resistance".
An interesting thing about AC resistance is that when you increase the test frequency, and the resistance climbs, it indicates eddy current losses are at work. Eddy currents are essentially modeled as resistance that rises with frequency. For example, an SSL-1 with AlNiCo slugs has rather low eddy currents. The standard multi-meter shows 6.72k ohms resistance for the SSL-1 I have on hand, but the Extech LCR meter shows an AC resistance of 6.71k ohms at 120Hz, and 6.874k ohms at 1kHz. Not much difference, not much eddy current. Now try it with the Fender Fideli'tron, which with its sizeable steel parts causing significant eddy current losses. The Fideli'tron shows a DC resistance of 4.46k ohms with a typical multi-meter, but the Extech shows a resistance of 4.521k ohms at 120Hz, a little higher, abd 7.112k ohms at 1kHz... a lot higher, indicating that eddy currents are "pushing back".
This isn't a good way to quantify eddy current resistance, though, and just by looking at a pickup and seeing how much steel is in and around the coils, you can get just as good of an idea how extreme the eddy current losses will be. The best way to see exactly how the eddy currents rise with frequency is to create an impedance bode plot with a network analyzer, such as the Velleman PCSGU250. When plotting a pickup with an oscilloscope and function generator, you should see a +6dB per octave slope before the resonance, but eddy currents will cause it to dip below +6dB per octace. The degree of eddy current losses will be proportionate to the downward divergence from +6dB/oct as the test frequency rises.
Pics:
