Pickup Tester, not a driven coil

[R.G.]

I've done a bit of searching on the forum here and on the net, and this one never came up. There's a lot of topics on using drive coils for testing pickups, but then you have the problem of the drive coil's inductance and the actual drive level, coil-to-coil magnetic coupling or lack thereof.

Seems like for testing a pickup, you want to move a string around over it.( No, not a calibrated string plucker... )

Stop me if this has been done before, but imagine:
A rod of nylon, about 0.062 diameter. This has a groove on one side and into that groove is epoxied a length of guitar string. The nylon rod is set into nylon bearings and has a nylon "foot" with an adjustable spacing. The foot lets you set the rod over the pickup under test.

For the test, you mechanically spin the nylon rod over the pickup being tested. For freedom from magnetic issues, this is probably best done with a belt or possibly air pressure. The entire fixture can be made from magnetically transparent materials.

The captive length of string moves in a true sine wave of vertical motion, but the horizontal-quadrature motion will probably keep this from being a pure sine output. But it will be constant. It won't depend on anything except the mechanical motion, which is consistent, and the pickup, which is being tested.

An FFT on the pickup output should then start to tell you whether the pickup is generating harmonics, which ones, and how much.

You can test the sensitivity of the pickup along it's length, not by driving the whole pickup as with a driver coil which encompasses the whole thing. You can tell where there are sweet spots, where there are nulls, and get some idea about the real output of the thing as well as an accurate frequency response to strings.

There's a lot of refinement to be done; as witness, while I was typing this, I figured out another couple of ways to drive the piece of string.

But as I said, probably someone else has figured this out before.

[David Schwab]

Interesting idea.

The thing about pickups and strings is that the motion of the string parallel to the top of the guitar produces different harmonics from the motion perpendicular to the top. So you want the motion to simulate the string's normal motion.

Another idea might be to have a tensioned string fixed between two points, and at one end have an audio speaker couple to the string's mount. You can induce the string to vibrate by exciting it with a signal to the speaker, and by varying the frequency of the signal, get varying harmonic nodes from the string. This could be swept from low to high.

Bartolini uses a mechanical string plucker.

[SteikBacon]

I have had the exactly the same idea.... I don´t agree that the different harmonics, at different spots at the strings, should be in such a test. A true test must have as pure sinecurve as possible.


But to agree on how to make or read the info isn´t the biggest issue. The problem would anyway to get a rig that can handle the frequencies.

[Joe Gwinn]

I've done a bit of searching on the forum here and on the net, and this one never came up. There's a lot of topics on using drive coils for testing pickups, but then you have the problem of the drive coil's inductance and the actual drive level, coil-to-coil magnetic coupling or lack thereof.

These are solvable problems, but OK.

Seems like for testing a pickup, you want to move a string around over it. ( No, not a calibrated string plucker... )

Stop me if this has been done before, but imagine:
A rod of nylon, about 0.062 diameter. This has a groove on one side and into that groove is epoxied a length of guitar string. The nylon rod is set into nylon bearings and has a nylon "foot" with an adjustable spacing. The foot lets you set the rod over the pickup under test.

For the test, you mechanically spin the nylon rod over the pickup being tested. For freedom from magnetic issues, this is probably best done with a belt or possibly air pressure. The entire fixture can be made from magnetically transparent materials.

It's not uncommon to use spinning samples in testing of magnetic materials, and a spinning coil is used in the exploration of very faint fields. Never heard of it for pickups, but it ought to work, at least for low frequencies.

One can use an electric motor, so long as the shaft is long enough. This is the usual approach. The motor can also be in a soft steel box, to shield it.

The captive length of string moves in a true sine wave of vertical motion, but the horizontal-quadrature motion will probably keep this from being a pure sine output. But it will be constant. It won't depend on anything except the mechanical motion, which is consistent, and the pickup, which is being tested.

Many pickups are sensitive more to one direction of motion than the orthogonal direction, and a few claim to be good for both. A mechanical arrangement that yields linear motion would be cleaner, and would yield easier-to-interpret data.

An FFT on the pickup output should then start to tell you whether the pickup is generating harmonics, which ones, and how much.

Yes, although we do know that the magnetic material in a pickup is not a significant source of harmonic distortion. The shape of the magnetic field is the key.

You can test the sensitivity of the pickup along it's length, not by driving the whole pickup as with a driver coil which encompasses the whole thing. You can tell where there are sweet spots, where there are nulls, and get some idea about the real output of the thing as well as an accurate frequency response to strings.

A small coil, wound on the end of a soda straw, will do the same thing.

There's a lot of refinement to be done; as witness, while I was typing this, I figured out another couple of ways to drive the piece of string.

How about gluing the piece of steel string to one tine of an aluminum tuning fork? Or drilling a small hole in the tine and inserting a piece of music wire? (You will need to glue an equal mass to the other tine, to maintain balance.)

The big issue with the motor method is the high RPMs needed. To achieve 1,000 Hz the nylon rod needs to turn at 60,000 rpm. An air motor can do this easily, but this is far beyond ordinary electric motors without belts or gears, and it's hard to keep things glued together at that speed.

If one is happy with 200 Hz, then 6,000 rpm is needed. Small electric motors can do this. However, a long thin shaft at 6000 rpm will tend to whip, and so will need to be routed through a larger rigid tube with bearings every inch or two. Nor can the shaft be nylon. Stainless steel (non-magnetic) will be needed. Tubing may work.

One can get small DC motors that will spin at 10,000 rpm, so this part isn't hard.

The tuning-fork approach may be easier, and cheaper, although one loses the ability to accurately know the amplitude of motion. Unless one mounts a piezoelectric sensor on one tine; perhaps it would be sensitive enough.

[David Schwab]

I have had the exactly the same idea.... I don´t agree that the different harmonics, at different spots at the strings, should be in such a test. A true test must have as pure sinecurve as possible.

This will test the coil's response, but not the way the pickup senses the strings due to their motion.

Check out the two Bartolini patents to see how the shape of the magnetic field effects the tone of the pickup. It's about the shape of the field as it relates to the lines of flux the string cuts through.

3983777

3983778

[David Schwab]

Many pickups are sensitive more to one direction of motion than the orthogonal direction, and a few claim to be good for both. A mechanical arrangement that yields linear motion would be cleaner, and would yield easier-to-interpret data.

That was my point, but you put it much more eloquently.

[R.G.]

The thing about pickups and strings is that the motion of the string parallel to the top of the guitar produces different harmonics from the motion perpendicular to the top. So you want the motion to simulate the string's normal motion.

That's true, and one reason that having the pseudo-string move in a circle is more like a real string.

Another idea might be to have a tensioned string fixed between two points, and at one end have an audio speaker couple to the string's mount. You can induce the string to vibrate by exciting it with a signal to the speaker, and by varying the frequency of the signal, get varying harmonic nodes from the string. This could be swept from low to high.Bartolini uses a mechanical string plucker.

Hmm... OK, I have another tester design. We have a wooden box here. It's about three feet long and maybe four inches wide, two inches high. Inside is a set of... sure, why not? ... six guitar strings, tensioned to standard tuning. Each string is driven at one end of the box with an e-bow-like driver coil, and the amplitude is sensed by a pickup coil, perhaps the one on the pseudo-e-bow driver, and electronically limited. Now I can start a string oscillating and with a knob set the level it's driven to. I don't think I've seen a driver coil used to make a constant-frequency/amplitude mechanical oscillator before, but I'm sure it works. One could even have a mechanical slider on the top of the box that moves the internal "nut" to sweep the frequency mechanically. In effect, the string becomes its own tuning fork. (see below)

I think I could get one of these working in under a day. And I think it is probably a better tester than I had proposed. DOH!!

Now my forehead hurts.

I have had the exactly the same idea.... I don´t agree that the different harmonics, at different spots at the strings, should be in such a test. A true test must have as pure sinecurve as possible.

Yes, that's correct. Actually, I'd say that a true test must have as pure a sine excitation as is practical. Hence the spinning string segment. However, if you can't have a pure sine wave, as long as what you have is constant, it works - you can record the result digitally and subtract out the imperfections. For instance, you could record the signal from a hall effect sensor above the coil, then subtract that out of the test signal. What's left is what the pickup added/subtracted.

But to agree on how to make or read the info isn´t the biggest issue. The problem would anyway to get a rig that can handle the frequencies.

Yeah. See above for the new one and below for the old one.

How about gluing the piece of steel string to one tine of an aluminum tuning fork? Or drilling a small hole in the tine and inserting a piece of music wire? (You will need to glue an equal mass to the other tine, to maintain balance.)

That's a problem. You'd have to use something like fiberglass, as the spinning aluminum itself will have eddy currents and distort the M-field it's running in, especially at high speeds.

The big issue with the motor method is the high RPMs needed. To achieve 1,000 Hz the nylon rod needs to turn at 60,000 rpm. An air motor can do this easily, but this is far beyond ordinary electric motors without belts or gears, and it's hard to keep things glued together at that speed.

Yeah, that's an issue, OK.

If one is happy with 200 Hz, then 6,000 rpm is needed. Small electric motors can do this. However, a long thin shaft at 6000 rpm will tend to whip, and so will need to be routed through a larger rigid tube with bearings every inch or two. Nor can the shaft be nylon. Stainless steel (non-magnetic) will be needed. Tubing may work.

One can get small DC motors that will spin at 10,000 rpm, so this part isn't hard.

Stainless has the same issue as aluminum, eddy currents. I think the shaft has to be something like fiberglass now that you mention the mechanical issues. The shaft probably needs to be solid, with one imbedded string. I envisioned the whole shaft as maybe 1 or 2 inches long, with a bearing support at each end, so whipping might not be a problem. An open high E string is up a bit over 320Hz and the twelfth fret is around 640 Hz, so the highest rpm you'd need is about 40K. Not trivial.

The tuning-fork approach may be easier, and cheaper, although one loses the ability to accurately know the amplitude of motion. Unless one mounts a piezoelectric sensor on one tine; perhaps it would be sensitive enough.

Check the second shot above. With an e-bow-like driver, I can electronically servo the size of the motion to be what I want it to be, speed is not an issue since the string is doing what it would otherwise do. By sensing the string output with a servo coil, I can cut back on the drive when some specific level is reached, and I can change that amplitude limit with a single knob. I can dial in the string amplitude.

So I think I'm off after variant 2 - the electronically servo'ed string. A guy who can wind pickups can wind both sense and drive coils. The sense "coil" could also be hall effect, or specially wound low signal / low impedance to sidestep the issue of the sense coil's own characteristics.

I'm certain that people have used "victim" guitars before, but by not using an actual guitar, you can put all the strings in the same plane, for testing pickup-to-string distance variation, you can use more than six strings by using a the selectable string damper I just invented in my head to only let one string vibrate at a time.

Between the slider "nut" on the strings, the selectable damper on the slider and the electronic servo on the string motion, I think it's a pretty workable pickup tester. About all that remains is to make it easy to fix the pickup over (easier than under) the strings at a fixed height, probably on a plastic sheet suspended over the strings on a table with variable height.

I think this is a workable idea. I'll go do some drawings and see if I can make up a test rig.

[David Schwab]

That's true, and one reason that having the pseudo-string move in a circle is more like a real string.

I think the rod with the groove is a good idea. It would catch the string moving in all directions. Maybe make the rod cone shaped?

For the other idea I was actually thinking of an eBow, since I have one. You'd need a magnet under the driver coil to help it along.

[Joe Gwinn]

That's a problem. You'd have to use something like fiberglass, as the spinning aluminum itself will have eddy currents and distort the M-field it's running in, especially at high speeds. ... Stainless has the same issue as aluminum, eddy currents.

The aluminum would not be spinning, it would be vibrating back and forth sinusoidally, but in an almost straight line. I'll have to think about it, but some well-placed slits ought to greatly reduce the eddy currents.

As for stainless steel, the proper alloy has a very low eddy current coefficient. Same discussion as for metallic covers on pickups.

I think the shaft has to be something like fiberglass now that you mention the mechanical issues. The shaft probably needs to be solid, with one embedded string.

Only the last 3/4 inch would need to be non-metallic.

I envisioned the whole shaft as maybe 1 or 2 inches long, with a bearing support at each end, so whipping might not be a problem.

It's the length-to-diameter ratio that matters: 2"/0.625"= 32:1. That's a very slender shaft. I'd use a 1/8" or 5/32" shaft for all but the last part. This makes the ratio a more manageable 16:1 or less.

An open high E string is up a bit over 320Hz and the twelfth fret is around 640 Hz, so the highest rpm you'd need is about 40K. Not trivial.

Even at 10:1, intermediate bearings will likely be needed. Keeping these bearings lubricated at 40,000 rpm could be a challenge. It's not impossible, especially for short periods of time, but requires some careful mechanical engineering. My guess is that a workable approach is a steel shaft running in a steel tube with some Oilite bearings between tube and rod.

Oilite (oil-impregnated porous bronze) bearings can go up to 5 meters per second surface speed, which gets us to about 30,000 rpm. Smaller bearings can probably go faster, as smaller things are easier to cool. And there are other materials. http://www.oilitebearings.com/

For a trove of suitable mechanical components, I usually get stuff from SDP/SI: https://sdp-si.com/index.asp.

I would order their paper catalogs. Browsing one the website is difficult and incomplete.

[SteikBacon]

To get a true result we got to get ALL frequencies the PICKUP can read. To just get the string vibrate (like with an e-bow) will just get low frequencies, and not the high frequencies that occur just as a pick attack and snap a string. The high freq is an important part of my opinion on how "open/sharp/warm/cold..etc" the pickup feels like. (I don´t test pickups with the tonecontroll on.)

To be able to have pieces of string, in the stringsizes, vibrate at up to 16kHz (my limit), sinuscurve, close to the pickup with no interference from the system itself.... hm,

[R.G.]

To get a true result we got to get ALL frequencies the PICKUP can read. To just get the string vibrate (like with an e-bow) will just get low frequencies, and not the high frequencies that occur just as a pick attack and snap a string. The high freq is an important part of my opinion on how "open/sharp/warm/cold..etc" the pickup feels like. (I don´t test pickups with the tonecontroll on.)

I should make my aims clear - right now I don't see ANY attempt to make a pickup tester in the open forum. Are either of these testers perfect? I'd be the first to tell you "No."

But I just came up with two different designs that can be used to (a) repeatably and (b) analytically test pickups that don't rely on "... um, I think this one is smoother, more highs and clearer mids, with a hint of blackberry at the finish" - the old subjective ear test.

Here's the second one as a sketch: GEO Pickup Tester. It uses actual strings, not necessarily exactly six or at the same frequencies, although I've shown it that way for simplicity and ease of hooking it to a tuner for setup. This one lets you study the steady state frequency response.

I can hypothesize an addition to that same setup that rotated a guitar pick past one string to "twang" it. It's not a calibrated pick stroke, but it's at least a repeatable one. With one of these in your shop, you could for sure tell what X winds and Y scatter with Z resistance and ... er... W inductance was doing to a pickup and how that pickup was different from the last one you wound.

And how about this - you can go whip an Excalibur-Mach 7 Firebreather pickup into your rig and find out how that one differs from your pickups on tests that you can ... repeatably... do at home for fun and profit.

To be able to have pieces of string, in the stringsizes, vibrate at up to 16kHz (my limit), sinuscurve, close to the pickup with no interference from the system itself.... hm,

It's daunting, isn't it?

But I don't think that should keep us from trying. Do you?

As far as I can tell, nothing is ever tested perfectly. It's only tested well enough to do what you need. It may be impossible to do a perfect test. So we test what we can.

Here are two new (as far as I've ever seen) testers for your pickups that you can cobble together for substantially zero cost and maybe an afternoon of work that will let you get some kind of grip on what pickups are really doing. And you can repeat the test tomorrow and get similar number. Well, OK, you should be able to get close at least.

Nope, they're not perfect. But then I've only been thinking about this for a few days. Give me a week or two...

P.S.
In typing this out, I came up with a calibrated string picker mechanism. OK, OK, calibrate-ish; probably not perfect either, just repeatable within some range.

Captive pick on a pivot. You pull on the pivot with a string attached to a spring-scale and dink with the pick until it plucks the string at X force on the scale. You can not only do this with a pick, but you can measure the force for changing picks and get back to a consistent measurement. Capture the pickup response into a computer mic input on the sound card then go play analysis games with Audacity, etc.

[Ampclutz]

Hmmmm....

Sounds sort of like a pickup holding fixture supported over an autoharp to me...

Call me crazy but that is simple except I don't have my autoharp anymore

AC

[R.G.]

It is indeed a pickup holding fixture supported over an autoharp. Except -
- you make it out of a plank of lumber
- it has a string driver providing constant amplitude on the string motion
- you can sweep frequency

The simplest version only needs one string and a modified ebow.

Multiple strings, sliding frequency changing bar, string selector and calibrated plucker are all embellishments for the people who insist that they have to have all possible tests.

[David Schwab]

Hmmmm....

Sounds sort of like a pickup holding fixture supported over an autoharp to me...

Call me crazy but that is simple except I don't have my autoharp anymore

No, it's more like a lap steel.

[R.G.]

workbench steel?