OT: Testing Frequency Response of an Audio Source - Pickup Exciter Coils

[YouAre]

Hey There!

Background:
My ultimate goal is to test the frequency response of guitar pickup using an exciter coil per the following article/tutorial:
http://guitarnuts2.proboards.com/thread/7723/measuring-electrical-properties-guitar-pickups

The simplified TLDR for those who don't want to read the article, basically you create a pickup coil without magnets (larger wire, fewer turns), place next to the pickup in question, run all audible frequencies through it, and measure the output at the guitar pickup under measurement.

The software, Room EQ Wizard will run a frequency sweep through the soundcard output, and will measure what's returned at the input. The original application is run the sweep through speaker monitors, and measure it with a flat mic placed where our ears would be when mixing. Setting the levels correctly should allow this to be useful for pedals or other audio circuits.

The purpose of this is to be able to compare pickups under similar conditions so I can visualize differences in components and winding techniques, in addition to listening to differences.

The idea is the create a coil with a low enough inductance but with enough turns to create a sufficient output. But....

The Problem:

I don't know how to confirm that all frequencies under testing are being reproduced at a (relatively) consistent amplitude to ensure an accurate (enough) reading when I run a frequency sweep through the source into the pickup.

I understand that the coil's output may not be perfectly flat, but I'd like to understand where any dips and spikes might be so I can compensate for them somehow.

The Potential Solutions/Alternatives:

-Would there be a benefit of using a flat range speaker or a full frequency audio transformer coupled next to the pickup under measurement?

-What useful information would I get if I ran the frequency response with the coil set up as a series or shunt filter? It'd essentially be an inductor in series or shunted to ground, right?

-I can measure it's inductance and resistance. Could I use the RLC model we use for pickups to test this frequency response?
( http://guitarnuts2.proboards.com/thread/7842/modeling-electric-guitar-ltspice )




Any thoughts you have on how to measure the frequency response of this DIY coil, or alternative ideas would be GREATLY appreciated.

Thank you all!

[Rob Strand]

I don't know how to confirm that all frequencies under testing are being reproduced at a (relatively) consistent amplitude to ensure an accurate (enough) reading when I run a frequency sweep through the source into the pickup.

I understand that the coil's output may not be perfectly flat, but I'd like to understand where any dips and spikes might be so I can compensate for them somehow.

When you look at the details these things never turn out to be simple.

The way I see it is to drive the drive coil with a constant current.   That will ensure you are generating a  constant B-field.   It will also ensure the coil resistance has no effect.  You can *approximate* constant current with a series resistor but you have to very careful that the coil inductance is small enough for this to hold true over the frequency range.

So one mistake I see would be to use a drive coil which has too much capacitance.  Like a pickup coil.   Here the coil capacitance would divert current from coil.  So even though from outside constant current goes through the terminals not all that current goes through the coil.    Another issue is having too many layers on the coil or using thick wire.  This would cause eddies and reduce the external field.

I did test pickups many moons ago and I use a round coil about 1" in diameter and about 25 turns of maybe 0.4mm or 0.5mm wire.  Those details aren't 100% accurate but it was in that order.   That coil is fine provided the pickups don't have too much metal around them.   BTW, you shouldn't really remove shields and metal covers on the pickup as these items contribute to the capacitance and frequency response of the coil.

[Eb7+9]

The Problem:

I don't know how to confirm that all frequencies under testing are being reproduced at a (relatively) consistent amplitude to ensure an accurate (enough) reading when I run a frequency sweep through the source into the pickup.

bingo ...

reminds me of my TA days,
watching students run to a solution w/o understanding the stated problem first

I know you’re not asking for nothing ...
you prob already sense that things are doomed from the get-go

—-

first of all, modelling a coil using a “lumped” RLC model makes for a nice simple view of life, but ... in the real world of engineering the use of finite methods to model layered coils generates a very complex problem ... you’ll be loosing lots of hair trying to justify that kind of coarse equivalence in this case ... assuming you have hair

people, especially in guitar gizmo land, like simple ...
makes it easier to talk big
(a common tell-tale sign of over simplification)

next, what is it you really want to quantify ?! the response of a moving (external) chunk of remanent material or that of an external M field - like the stray fields of a power transformer powering a (nearby) amplifier - ?! ... I suspect you might have a hard time proving, to me anyway, that they are the same

have you read Seth Lover’s PAF pickup patent ?!!

—-

I looked at this problem years ago, and yes read all the over-simplified solutions out there / which are basically the same as presented here ... I came up with a trick for taking the C portion of equation, assuming a lumped model // again, futility for the reasons stated above

I feel that stimulating a coil electrically (using signal generator and large value resistance in series) to uncover its resonant frequency is the only thing that isn’t wrong here // again, not easy to prove or disprove ... but does it get you closer to knowing the profile of that response as you seek ?! ... the answer is a pretty clear nay // and if you think that should be a yes good luck coming up with a proof that goes beyond mere hand waving (something that is typically left un-mentioned)

briefly, the mistake everybody makes is failure to remain in the electro-mechanical domain, the source being mechanical in nature ... I will make a slightly bold claim here and suggest the only correct way of doing this is by swinging a piece of guitar string over the pickup (to “close” the magnetic circuit) up to very high speeds

do the math ... think you can make a piece of guitar string cross the surface of the PU 6000 times a second AND maintain constant swing amplitude ?!

here’s one highly theoretical way of doing it: imagine a large wooden/styrofoam wheel with emended pieces of guitar string of equal length spaced apart by at least twice the width of the PU surface (so one doesn’t interfere w the effect of the previous length as they pass over the PU) ... now let’s say you can construct a well balanced wheel with 100 embedded pieces of string ... you’d have to reach up to 60 rotations per second to attain an equivalent 6000hz upper test limit ... or, you could somehow try to make the motion simulate a square wave and argue based on response of harmonics, but then good luck moving a piece of string back and forth across the PU at those speeds!

I dunno, wishful thinking me thinks ...

I’d give NASA a call

[YouAre]

When you look at the details these things never turn out to be simple.

I know

The way I see it is to drive the drive coil with a constant current.   That will ensure you are generating a  constant B-field.   It will also ensure the coil resistance has no effect.  You can *approximate* constant current with a series resistor but you have to very careful that the coil inductance is small enough for this to hold true over the frequency range.

My buddy, Jackie O, has this circuit:
http://www.muzique.com/lab/superbuff.htm

Might use that for the coil driver.

I'm a bit unfamiliar with the series resistor constant current approximation. Would you be able to point me in the right direction of where to read up on it's application? Much appreciated!

So one mistake I see would be to use a drive coil which has too much capacitance.  Like a pickup coil.   Here the coil capacitance would divert current from coil.  So even though from outside constant current goes through the terminals not all that current goes through the coil.    Another issue is having too many layers on the coil or using thick wire.  This would cause eddies and reduce the external field.

Yup! Too evenly parallel of layers causing excess capacitance. So going to limit it to 200 turns with a random scatterwind to mitigate extra capacitance.

I did test pickups many moons ago and I use a round coil about 1" in diameter and about 25 turns of maybe 0.4mm or 0.5mm wire.  Those details aren't 100% accurate but it was in that order.   That coil is fine provided the pickups don't have too much metal around them.   BTW, you shouldn't really remove shields and metal covers on the pickup as these items contribute to the capacitance and frequency response of the coil.

Curious to see if swapping magnet types affects the capacitance!

Thank you again for your help!

[YouAre]

bingo ...

reminds me of my TA days,
watching students run to a solution w/o understanding the stated problem first

I know you're not asking for nothing ...
you prob already sense that things are doomed from the get-go

—-

first of all, modelling a coil using a "lumped" RLC model makes for a nice simple view of life, but ... in the real world of engineering the use of finite methods to model layered coils generates a very complex problem ... you'll be loosing lots of hair trying to justify that kind of coarse equivalence in this case ... assuming you have hair

people, especially in guitar gizmo land, like simple ...
makes it easier to talk big
(a common tell-tale sign of over simplification)

next, what is it you really want to quantify ?! the response of a moving (external) chunk of remanent material or that of an external M field - like the stray fields of a power transformer powering a (nearby) amplifier - ?! ... I suspect you might have a hard time proving, to me anyway, that they are the same

have you read Seth Lover's PAF pickup patent ?!!

—-

I looked at this problem years ago, and yes read all the over-simplified solutions out there / which are basically the same as presented here ... I came up with a trick for taking the C portion of equation, assuming a lumped model // again, futility for the reasons stated above

I feel that stimulating a coil electrically (using signal generator and large value resistance in series) to uncover its resonant frequency is the only thing that isn't wrong here // again, not easy to prove or disprove ... but does it get you closer to knowing the profile of that response as you seek ?! ... the answer is a pretty clear nay // and if you think that should be a yes good luck coming up with a proof that goes beyond mere hand waving (something that is typically left un-mentioned)

briefly, the mistake everybody makes is failure to remain in the electro-mechanical domain, the source being mechanical in nature ... I will make a slightly bold claim here and suggest the only correct way of doing this is by swinging a piece of guitar string over the pickup (to "close" the magnetic circuit) up to very high speeds

do the math ... think you can make a piece of guitar string cross the surface of the PU 6000 times a second AND maintain constant swing amplitude ?!

here's one highly theoretical way of doing it: imagine a large wooden/styrofoam wheel with emended pieces of guitar string of equal length spaced apart by at least twice the width of the PU surface (so one doesn't interfere w the effect of the previous length as they pass over the PU) ... now let's say you can construct a well balanced wheel with 100 embedded pieces of string ... you'd have to reach up to 60 rotations per second to attain an equivalent 6000hz upper test limit ... or, you could somehow try to make the motion simulate a square wave and argue based on response of harmonics, but then good luck moving a piece of string back and forth across the PU at those speeds!

I dunno, wishful thinking me thinks ...

I'd give NASA a call

You sound fun You're invited to my next party.

There's enough residual THC in my system that I can take your challenge about the vibrating string....

Take a constant torque ac motor, attach to a shaft long enough that the electromagnetic fields from the motor are isolated enough from our system. Then, just as you described, string end attached. Swing that guy in a circle, applying frequency sweep as AC input through a suitable amplifier.

Or

Just put the motor next to the pickup. Use THAT swinging hunk of metal inside....oh wait. There's multiple "bars" on the squirrel cage, that might mess with the readings. Either way!

[Sooner Boomer]

I did something like this, just not as complicated.  I wound about 10 feet of #32 wire on a 1" form (wound edge-edge, as neatly as possible).  The coil was "fixed" with a couple of coats of thin yellow glue.  After the glue dried, the coil was stiff enough on its own to be taken off the form.  I got lazy, and just used a 555 running at about 1KHz on 9v to drive the coil.  It's not enough to get a frequency response curve, but it helps troubleshoot if a coil is working.

You might re-post your question here:
https://music-electronics-forum.com/forumdisplay.php?f=11&s=8926368e39ba00b093f40c9f75ee9414

[Rob Strand]

My buddy, Jackie O, has this circuit:
http://www.muzique.com/lab/superbuff.htm

Might use that for the coil driver.

OK so that circuit gives you more output current but it is still a voltage output circuit not a current output.
A current output circuit is more like fig 2 on this page where you put the coil in the feedback loop.  The output *current* is proportional to the input voltage.

https://sound-au.com/project34.htm

By the way, this circuit also increases the output current but it uses the output transistors.  Either transistors or parallel opamps can work.

I'm a bit unfamiliar with the series resistor constant current approximation. Would you be able to point me in the right direction of where to read up on it's application? Much appreciated!

It's a very simple idea.   Suppose you have a voltage source V and put a large resistance Rs in series with it, then you connect that to your load ZL.  The output current is, for complex impedances,

    I   = V  /( Rs + ZL)

*If* Rs is much larger than ZL that reduces to,

   I  = V / Rs

which is constant.    Notice though the choice of Rs depends on ZL.   So if Rs turns out to be large then you need a very large V to get the current.   So it's not always possible to use that trick.

Curious to see if swapping magnet types affects the capacitance!

So if the magnets aren't conductive acts like a dielectric but if the magnets are conductive then you have stray capacitance to the metal surface and in that case it will depend if the magnets are floating, grounded or close to another metal surface.    I suspect these are small compared to the winding capacitance but careful measurements show-up all sorts  of things.

next, what is it you really want to quantify ?! the response of a moving (external) chunk of remanent material or that of an external M field - like the stray fields of a power transformer powering a (nearby) amplifier - ?! ... I suspect you might have a hard time proving, to me anyway, that they are the same

The final response is the whole package, which is affected by the magnetic aperture of the pickups as well.  However, since the mechanical and electrical are strongly grrr weakly coupled (ie. the inductance and capacitance aren't largely affected by the strings) the response will be the product of the mechanical and electrical responses.

An interesting thing about pickups is the string moves up and down and side to side.  The up/down motion is asymmetrical from the perspective of the changing field but side to side can be symmetrical or asymmetric depending if there is a single pole or a pole each side.    You can get double frequency components from one but not the other.

[YouAre]

OK so that circuit gives you more output current but it is still a voltage output circuit not a current output.
A current output circuit is more like fig 2 on this page where you put the coil in the feedback loop.  The output *current* is proportional to the input voltage.

https://sound-au.com/project34.htm

By the way, this circuit also increases the output current but it uses the output transistors.  Either transistors or parallel opamps can work.

The second you said constant current, I immediately thought of ESP's reverb article! Just to clarify, are you saying I should put the coil in place of R7? Or should I be replacing the jumper between R7 and C8 with this coil?

It's a very simple idea.   Suppose you have a voltage source V and put a large resistance Rs in series with it, then you connect that to your load ZL.  The output current is, for complex impedances,

    I   = V  /( Rs + ZL)

*If* Rs is much larger than ZL that reduces to,

   I  = V / Rs

which is constant.    Notice though the choice of Rs depends on ZL.   So if Rs turns out to be large then you need a very large V to get the current.   So it's not always possible to use that trick.

According to the following:
https://www.powerstream.com/Wire_Size.htm

Let's assuming 36AWG, which allows for 0.035 amps for power transmission. I can assume that the ampacity is constant throughout all frequencies, right?

So in order to accomplish this, I'd want to:

1. Wind my coil, determine resistance (ZL in this case)
2. Determine maximum output voltage of my amplifier in this setup (V)
3. Using the above given max ampacity, solve for the safest Series Resistance RS that won't burn out my coil?

Sound about right?

Would this setup be used in conjunction with the circuit above, either with the resistor and coil in the feedback loop, or on the output of that driver circuit?

So if the magnets aren't conductive acts like a dielectric but if the magnets are conductive then you have stray capacitance to the metal surface and in that case it will depend if the magnets are floating, grounded or close to another metal surface.    I suspect these are small compared to the winding capacitance but careful measurements show-up all sorts  of things.

Sonically, I can hear the resonant peak shifting or flattening/sharpening as I swap magnets, would be interesting to see visually. I actually meant the inductance, and was confusing my properties. Increasing the amount of ferrous content, like pickup covers, affects the inductance, right? It would make sense to me that changing the metallurgy of the whole system, by swapping the magnet, would affect that same property no? So theoretically, I can measure these differences in inductance after swapping covers and magnets? Haven't done this yet, but I want to.

The final response is the whole package, which is affected by the magnetic aperture of the pickups as well.  However, since the mechanical and electrical are strongly grrr weakly coupled (ie. the inductance and capacitance aren't largely affected by the strings) the response will be the product of the mechanical and electrical responses.

An interesting thing about pickups is the string moves up and down and side to side.  The up/down motion is asymmetrical from the perspective of the changing field but side to side can be symmetrical or asymmetric depending if there is a single pole or a pole each side.    You can get double frequency components from one but not the other.

So the angle of pick attack is partially why tone is in the fingers!

I'll consider that confounding data. Whenever we see the frequency response of speakers, we look at it in isolation and not as part of the system. The poweramp that drivers the speaker, the microphone that measures the response. These are all contributing factors that we assume to be constant whenever a manufacturer provides this data. If I were comparing celestion's published frequency responses to eminence's, I'd be comparing apples to apples, but each apple is in a different pie, using a different recipe, baked by different bakers.


Thanks again for all your help!

[YouAre]

I did something like this, just not as complicated.  I wound about 10 feet of #32 wire on a 1" form (wound edge-edge, as neatly as possible).  The coil was "fixed" with a couple of coats of thin yellow glue.  After the glue dried, the coil was stiff enough on its own to be taken off the form.  I got lazy, and just used a 555 running at about 1KHz on 9v to drive the coil.  It's not enough to get a frequency response curve, but it helps troubleshoot if a coil is working.

You might re-post your question here:
https://music-electronics-forum.com/forumdisplay.php?f=11&s=8926368e39ba00b093f40c9f75ee9414

A pickup making forum?! Thank you sir! I will sign up and I'll report my findings.

[PRR]

I think you are making this complicated. Your measurement system adds complexity which you don't seem ready to understand.

You do NOT want a "flat response" on a guitar. The higher string overtones are not exact harmonics and sound bad, dissonant. A steep high-cut is rather essential. Yes, you could do this in the amplifier, but....

There is no free lunch. Increased output inevitably means a reduced top frequency. Especially if you have a cable (capacitance) after the pickup. Below I have plotted the response for a fixed input through a loosely coupled "transformer" (which is what a pickup is) into a cable, for various turn-ratios (winding turns). The 100H (many-many-many turns) winding has the most below 1KHz but then falls. The 1H (few turns) is flat to the upper audio band (that peak won't be so bad) but is 20dB 10:1 weaker than the 100H winding.


We see the same thing, reversed, in woofers. A Fifteen can be efficient above 150Hz or flat low-efficiency down to 50Hz.

We see this in microphones. Here the diaphragm diameter scales the tradeoff.


The "optimization" for a guitar pickup is to pick a happy top-frequency and wind-up until the resonance is there.

The resonance can be shifted a part-octave up by putting a buffer AT the pickup. However the pickup winding has significant capacitance (hundreds of pFd) so it doesn't go up far.

[j_flanders]

Some potentially useful links from my "pickup frequency response" bookmarks folder:

https://guitmod.wordpress.com/2016/09/26/diy-simple-measuring-of-a-pickups-frequency-response/
http://guitarnuts2.proboards.com/user/4605/recent?page=3 (pretty much all of antigua's posts, they are all worth a read)
https://bedlamguitars.wordpress.com/technical-info/pickup-inductance/
http://kenwillmott.com/blog/archives/152

I have used the setup from the first link.
With this setup I have tested some of the pickups antigua has tested with his setup and I got the same frequency responses.

I don't know if it delivers the actual frequency response, but it is very useful to compare the same pickup with different magnets, with or without an electrically conductive cover, with or without pole pieces, baseplate etc. And it surely beats my deceiving ears.

I also replicated Ken Wilmotts experiments in slicing up pickup covers to avoid eddy currents.
http://kenwillmott.com/blog/archives/246
Although I could clearly hear the difference it's nice to also have proof through the frequency response.

[Rob Strand]

The second you said constant current, I immediately thought of ESP's reverb article! Just to clarify, are you saying I should put the coil in place of R7? Or should I be replacing the jumper between R7 and C8 with this coil?

The way I would say it is the coil goes where the reverb coil is.   You may need to add a high value R7 if the circuit oscillates but if you make R7 too low it will affect measurements in the upper frequencies.

So in order to accomplish this, I'd want to:

1. Wind my coil, determine resistance (ZL in this case)
2. Determine maximum output voltage of my amplifier in this setup (V)
3. Using the above given max ampacity, solve for the safest Series Resistance RS that won't burn out my coil?

Sound about right?

If the coil was only resistive the current would be V / (Rs + Rcoil) which is already constant.  The problem actually is the coil has inductance and the ZL varies with frequency which means the current with vary with frequency.   Rs needs to be large enough so the current at the lowest frequency is (almost) the same as at the highest frequency.

For a sample coil of 25turns on a 25mm diameter coil with say 0.4mm wire the inductance will be about 25uH and the DC resistance about 0.3ohms.   The inductive reactance of a 25uH inductor at 20Hz is 3 milliohms, which then increases to about 3 ohms at 20kHz.   If the coil was placed near some iron that might go up by a factor of 3.   So where in the realm of 10 ohms.  If you chose Rs to be 1k ohm then current variations would be relatively small.     Different coils will have different inductances.  If your inductive reactance was 100 ohms at 20kHz then that would require Rs to be 10k ohm.  So with 15V drive that limits you drive current  15V/10k = 1.5mA.     Don't read too much into the specifics here I'm just outlining the problem.   When you design the test jig and consider these effects you have to determine if this method can be used or not.

So the angle of pick attack is partially why tone is in the fingers!

You can ear that effect on an acoustic guitar which has no pickups. It's actually that's got to do with the shape of the string before it is released.   The pick has a sharp bend where as the fingers have a smooth bend at the contact point.   In physics you analyse the string using fourier series and the sharp bend has more higher harmonics.   Reality isn't that straight forward but it does give you an idea.   You will find something on the web about this for sure.

[Gus]

Pickup cover eddy currents

Has anyone used thin stainless steel. It has higher resistivity compared to copper and other metals

[Eb7+9]

Come to think of it, you can do a linear test - no need to use pulses
and 6khz equivalent is likely unnecessary

by using a std tuned high E string as reference tension
We can possibly get close to what we need

rethinking the practical numbers

say we assume a useful rest range of  4khz
that is equivalent to a 48th fret frequency to give an idea
or half way to the bridge from a 24th fret

so using a sliding fret system to sweep test frequencies
an exciter coil and variable gain amplifier
some kind of method to maintain constant (visual) string displacement

provided the exciter circuit can drive the string at
those rates (or higher) and you keep displacement
low you might have yourself an accurate linear testing method

silly not to think of that first
one thing that would be interesting to see is how the same pickup rewound differently
would alter the response on a graph

[YouAre]

I think you are making this complicated. Your measurement system adds complexity which you don't seem ready to understand.

If I haven't bitten off more than I can chew, it's just not full days work.

You do NOT want a "flat response" on a guitar. The higher string overtones are not exact harmonics and sound bad, dissonant. A steep high-cut is rather essential. Yes, you could do this in the amplifier, but....

I don't want a flat response from my guitar. I'd like it (reasonably) for my exciter coil. I imagined the response of the coil would be similar to that of a pickup, being a similar design.

My concern is that if I can potentially wind the exciter coil so that it's resonant frequency can be too low and can skew the results of the pickup measurement. Assuming the typical pickup RLC model for arguments sake, I can measure inductance and resistance, but not capacitance. My worry is that a capacitance

There is no free lunch. Increased output inevitably means a reduced top frequency.

Hence the desire to be able to determine the low pass cutoff. I want to optimize the output of the exciter coil while not affecting frequency.

I understand that this exciter coil test, with it's frequency sweep, is not going to give the same response as a harmonically complex guitar signal. I get that a 440hz test tone is not going to make the pickup react the same as the equivalent A note on guitar.

The goal is to optimize the output of the coil, via winding count and input signal strength, to reasonably match that of the disturbance that a plucked string creates.

Especially if you have a cable (capacitance) after the pickup.

For the purposes of this exercise, we're focusing on the pickup element alone, before moving on to rest the system.

Below I have plotted the response for a fixed input through a loosely coupled "transformer" (which is what a pickup is) into a cable, for various turn-ratios (winding turns). The 100H (many-many-many turns) winding has the most below 1KHz but then falls. The 1H (few turns) is flat to the upper audio band (that peak won't be so bad) but is 20dB 10:1 weaker than the 100H winding.

Thanks for providing this!

Do we need to add a capacitive element to the equation though? It's certainly a concern with pickups, and I'm wondering if it's a factor of if it's negligible with the exciter coil.

Can we assume a similar model for the setup i'm describing, the exciter coil placed adjacent to the pickup? If so, I can measure the inductance and resistance of the exciter and pickup, but not the inherent capacitance.

We see the same thing, reversed, in woofers. A Fifteen can be efficient above 150Hz or flat low-efficiency down to 50Hz.

We see this in microphones. Here the diaphragm diameter scales the tradeoff.

Again, valuable insight! Thank you!

So it looks like I need to wind my coil for maximum frequency output, and increase the voltage input to the coil within the limits of what the coil will withstand. Understanding that the lower resistance coil will have the higher frequency response, but can withstand less of an input voltage.




Would I potentially be better off getting an appropriate amperage rated - low value unshielded inductor, and coupling that next to the pickup?

If that's the case, can I rip the voice coil out of a full-range frequency speaker? Understanding that the magnet/cone/material all affect the "full range response" of a speaker, can we reasonable assume that the coil of one of these speakers will do an OK job?

Thank you!

[Rob Strand]

Would I potentially be better off getting an appropriate amperage rated - low value unshielded inductor, and coupling that next to the pickup?

At first it seems like a good idea but  when you start putting ferrites near strong magnets the field saturates the ferrite and you might as well just use an air cored inductor.   I think the ferrite will add to the uncertainty in the inductance:  high inductance when not saturate and low when saturated.  The purpose of using a drum core inductor (when designing inductors) is to *stop* the field firing out the ends and keep the field between the drum sides.   If you want to try a cored inductor a short ferrite rod might be the best.

If that's the case, can I rip the voice coil out of a full-range frequency speaker? Understanding that the magnet/cone/material all affect the "full range response" of a speaker, can we reasonable assume that the coil of one of these speakers will do an OK job?

Destroying a speaker for that is crazy.   It's only a simple coil.  The cone of a full-range speaker makes it full-range not the coil.   Just make an air-core coil like the one I made or like the ones in the links you posted.

[YouAre]

Some potentially useful links from my "pickup frequency response" bookmarks folder:

https://guitmod.wordpress.com/2016/09/26/diy-simple-measuring-of-a-pickups-frequency-response/
http://guitarnuts2.proboards.com/user/4605/recent?page=3 (pretty much all of antigua's posts, they are all worth a read)
https://bedlamguitars.wordpress.com/technical-info/pickup-inductance/
http://kenwillmott.com/blog/archives/152

I have used the setup from the first link.
With this setup I have tested some of the pickups antigua has tested with his setup and I got the same frequency responses.

I don't know if it delivers the actual frequency response, but it is very useful to compare the same pickup with different magnets, with or without an electrically conductive cover, with or without pole pieces, baseplate etc. And it surely beats my deceiving ears.

I also replicated Ken Wilmotts experiments in slicing up pickup covers to avoid eddy currents.
http://kenwillmott.com/blog/archives/246
Although I could clearly hear the difference it's nice to also have proof through the frequency response.

These were fantastic resources, thank you! And yes, Antigua seems to be the top dog doing this, but I would like to check his methods on the coil construction.

I didn't realize Ken Willmott did so much more work! Thank you again!

[YouAre]

The way I would say it is the coil goes where the reverb coil is.   You may need to add a high value R7 if the circuit oscillates but if you make R7 too low it will affect measurements in the upper frequencies.

If the coil was only resistive the current would be V / (Rs + Rcoil) which is already constant.  The problem actually is the coil has inductance and the ZL varies with frequency which means the current with vary with frequency.   Rs needs to be large enough so the current at the lowest frequency is (almost) the same as at the highest frequency.

For a sample coil of 25turns on a 25mm diameter coil with say 0.4mm wire the inductance will be about 25uH and the DC resistance about 0.3ohms.   The inductive reactance of a 25uH inductor at 20Hz is 3 milliohms, which then increases to about 3 ohms at 20kHz.   If the coil was placed near some iron that might go up by a factor of 3.   So where in the realm of 10 ohms.  If you chose Rs to be 1k ohm then current variations would be relatively small.     Different coils will have different inductances.  If your inductive reactance was 100 ohms at 20kHz then that would require Rs to be 10k ohm.  So with 15V drive that limits you drive current  15V/10k = 1.5mA.     Don't read too much into the specifics here I'm just outlining the problem.   When you design the test jig and consider these effects you have to determine if this method can be used or not.

At first it seems like a good idea but  when you start putting ferrites near strong magnets the field saturates the ferrite and you might as well just use an air cored inductor.   I think the ferrite will add to the uncertainty in the inductance:  high inductance when not saturate and low when saturated.  The purpose of using a drum core inductor (when designing inductors) is to *stop* the field firing out the ends and keep the field between the drum sides.   If you want to try a cored inductor a short ferrite rod might be the best.

Destroying a speaker for that is crazy.   It's only a simple coil.  The cone of a full-range speaker makes it full-range not the coil.   Just make an air-core coil like the one I made or like the ones in the links you posted.

Rob, I wanted to thank you again personally for all your help. I think I got some valuable insight that will help!Combining quotes from both your posts. My misconception about resistance vs. impedance of the actually might be the key to this....

From this thread, I'm gathering that the RLC model for this coil is a reasonable approximation.


Can we assume the following frequency response or something similar?



So theoretically, I want to scatterwind my coil to reduce C as much as possible, choosing my number of winds to dictate L and R. The idea is to push that resonant peak outside of 20Khz.

For a moment, let's ignore sufficient output level, and focus on frequency response:
Let's say I screw up and have a peak at 10Khz, and I drive it with a voltage amplifier that doesn't put out constant current.
So let's measure the output of my amplifier to the coil on a scope and let's put a current meter ahead of the coil.
1. I run a 100Hz test tone into my amplifier, observe the amplifier output, and measure the current.
2. Then I run a 1K test tone, theoretically the output amplitude and current draw should be the same.
3. Then run a 10K (same as resonant peak of my coil), the output and current draw should be different, right?

Could this be a decent test for frequency response?


Now you touched on an interesting topic. Inductor "saturation." I'm assuming that in this case, it means that up to a certain point (assuming that we're not running more current than the wire is capable of handling) that the magnetic field output of this coil is proportional to the size of the signal, then past a threshold the field doesn't get any stronger despite the increasing voltage input. That sound about correct?


And don't worry, not taking apart any speakers! I was considering coupling the back of a full range speaker to the pickup under testing in lieu of the exciter coil, but the speaker magnet seems like it would skew the results based on the above.