Extended response pickups for bass

[Mike Sulzer]

This discussion concerns extended frequency range pickups for basses. Some questions are the following. Does it require a low impedance pickup? If not, how can it be done otherwise? Can it be done with a passive system, or is a preamp necessary? How should an extended frequency range pickup be designed?

I have made and compared a couple of pickups, to be considered later, but first here is a review of what causes frequency limitation in a pickup. The pickup circuit and the load it operates into make up a low pass filter. The pickup coil is an inductor, but there is also some capacitance associated with it. The most important element of the external load is the cable capacitance, although the resistance can matter, too, and in some circumstances additional resistance can be added to help shape the response. The effect of the cable capacitance is usually more important than than that of the coil.

An inductor looking into a shunt resistor is a first order low pass filter. When it looks into a capacitance, it is a second order filter and can have a resonance; this means the response can increase with increasing frequency before falling off quickly. However, the second order filter can also be damped with a resistor so that the response does not exhibit a significant peak.

There is a conflict between high output (lots of turns, high inductance) and broad frequency response (must have fewer turns, and has a lower output). This is usually not a problem for an electric guitar, where distortion fills in the higher harmonics, usually by preference. However, certain kinds of bass playing need a clean signal with wide frequency range. How can this be achieved?

The most obvious way is to put a lot fewer turns on the coil and use larger wire, resulting in a so-called low impedance pickup with both low inductance and resistance. Such a pickup requires additional amplification, and is well-suited to connection to a standard low impedance microphone input, especially if it uses a two-sided balanced configuration.

Another way is to put a preamp in the bass. Of course, this removes the cable capacitance from the load, and it is no longer necessary to use such a low impedance since the load on the inductance is a higher impedance, that is, a smaller capacitance. In fact, it is not a good idea to use a very low impedance pickup with a preamp in the bass since preamps designed to have good good signal to noise ratio with a very small voltage require, in general, more current than preamps designed to work with a higher voltage from a higher impedance pickup. High current is bad for battery life.

There is another way that might get good enough frequency response, a compromise between high and low. Reduce the number of turns to put the resonance with a typical cable just above the highest required frequency, and then load the pickup with a resistor (located in the bass) to flatten out the resonant peak.

So to test this idea I have made two pickups, one very low impedence, the other medium as just described. These pickups are the same except for the winding. So, how should such a pickup be designed? First, one should minimize conductors, no steel cores, in order to avoid eddy currents that can affect the frequency response. These pickups are air core except for a 3/16" by 1/32" diskAcomp.mp3Bcomp.mp3 neo under each string embedded in the top of the bobbin near the string. Without a high permeability core, one loses the ability to direct the flux from the vibrating string through the coil. So I use a shallow wider coil, only 3/16" deep and 15/16" wide, in order to match the more spread flux.

The low impedance coil has 250 turns total of #30 wire wound bifilar, 125 turns each side of the center tap; and so it is set up to feed a low imp. mic. input, high gain, low noise. The inductance is 2.8 mH, and the resistance is 10.5 ohms. The medium impedance coil has 2700 turns of #42 an it feeds a standard single ended instrument input. Its inductance is 350 MH and its resistance is 2K. When used with a 630 pf cable, it has a resonance of about 10KHz, and loading it with a resistance in the 16K to 20K range makes the response nearly flat, depending on how you want to set it up. The ratio of the inductances is very nearly the square of the turns ratio as one should expect. The gain of the low impedance coil is flat as high as I can measure (96KHz sampling rate). This is measured using a small coil to excite the pickup coil.

How did I decide on 10 KHz? I measured the spectrum of the bass using the low impedance pickup. It falls at about 30 db per decade above 1KHz, and you can see string harmonics to about 20 KHz. However, the SNR is quite poor out there, and so I am assuming that somewhere between 5 and 10 KHz is good enough response as a practical matter.

I have recorded a little bit of crappy bass playing with each pickup. (This is my old aluminum bass, only one I have.) Can you decide which coil is which?

(By the way, it was necessary to set the pickups to the same distance from the strings with very good precision since the different magnetic field strength at the strings makes a large different on the level of higher string harmonics, due to the magnetic damping.)Acomp.mp3Bcomp.mp3

[cnb77]

I feel like I hear more high end in the B clip, so I think that's the special pickup. I don't have my headphones nearby, just using my built in speakers, so I could be way off

[David Schwab]

They both sound the pretty much same to me. It's hard to tell with you using what sounds like a rather thin pick, but I'd guess clip A is the heavier wire.

In the pickup world, a pickup with 2700 turns of #42 would be considered low impedance, even though it's medium, since typical high impedance bass pickups are in the 8-14k range. So it's either you have a high z pickup, or you don't.

Alembic pickups have about 1500 turns of 40 gauge wire. The stacked Alembic pickups read about 8k, and are pretty low output. The Bill Lawrence designed low Z pickups for the Les Paul Signature bass were 2.2k (humbucker, in series), and wound with 42 gauge wire. MusicMan Stingrays read about 2k, and are two 4k coils in parallel. All these pickups use either a preamp or transformer to boost the level and match to high z amps.

So the real comparison would be any of those, or your pickups, compared to a standard off-the-shelf high impedance pickup, like a fender Jazz (8-9k), Fender P (11k) or EMG HZ (8-11k), since all of the above medium Z pickups fall into the not high Z realm. The difference in tone will be night and day.

[Mike Sulzer]

They both sound the pretty much same to me. It's hard to tell with you using what sounds like a rather thin pick, but I'd guess clip A is the heavier wire.

A it is, but you might not be hearing the wire alone. By the way, that is a finger, not a pick. They sound so similar because the frequency responses in the important range are nearly the same, and other differences have been minimized.

In the pickup world, a pickup with 2700 turns of #42 would be considered low impedance, even though it's medium, since typical high impedance bass pickups are in the 8-14k range. So it's either you have a high z pickup, or you don't.

Then the pickup world is out of touch with the real world. In the real world a low impedance pickup is one that can be easily used with an appropriate cable without loss of bandwidth. My 2700 turn pickup must be loaded carefully in order to get the full BW, and so it is not low impedance by normal definition.

So the real comparison would be any of those, or your pickups, compared to a standard off-the-shelf high impedance pickup, like a fender Jazz (8-9k), Fender P (11k) or EMG HZ (8-11k), since all of the above medium Z pickups fall into the not high Z realm. The difference in tone will be night and day.

Pickups themselves are not different by night and day, but pickup and load can be: that is, the pickup and how it is used. You are comparing capacitive loaded high impedance pickups to lower impedance pickups operating into a much lower C, perhaps resistively loaded as well, that is, a preamp input. Both those differences matter, and differences cannot be attributed to the pickup alone, since you can use the same pickup in different ways.

Suppose that I took my design, increased the bobbin size to allow 8,000 turns and ran it into a preamp with minimal cable. How close would it get to the extended range sound? I think pretty close. The L would rise by less than the square of the turns ratio since the bobbin shape must change, and the C would drop by almost 10. This would leave the resonant frequency similar, and the optimum load would be higher, maybe 80K. This is all somewhat uncertain, and you would have to try it to prove it, but it is certain that there is no inherent "night and day" difference in the pickups themselves.

[bbsailor]

A it is, but you might not be hearing the wire alone. By the way, that is a finger, not a pick. They sound so similar because the frequency responses in the important range are nearly the same, and other differences have been minimized.


Then the pickup world is out of touch with the real world. In the real world a low impedance pickup is one that can be easily used with an appropriate cable without loss of bandwidth. My 2700 turn pickup must be loaded carefully in order to get the full BW, and so it is not low impedance by normal definition.



Pickups themselves are not different by night and day, but pickup and load can be: that is, the pickup and how it is used. You are comparing capacitive loaded high impedance pickups to lower impedance pickups operating into a much lower C, perhaps resistively loaded as well, that is, a preamp input. Both those differences matter, and differences cannot be attributed to the pickup alone, since you can use the same pickup in different ways.

Suppose that I took my design, increased the bobbin size to allow 8,000 turns and ran it into a preamp with minimal cable. How close would it get to the extended range sound? I think pretty close. The L would rise by less than the square of the turns ratio since the bobbin shape must change, and the C would drop by almost 10. This would leave the resonant frequency similar, and the optimum load would be higher, maybe 80K. This is all somewhat uncertain, and you would have to try it to prove it, but it is certain that there is no inherent "night and day" difference in the pickups themselves.

Mike,

Yes, the load can make pickups sound like night and day. Try making a simple source follower type buffer with a 2 to 3 meg input impedance. The pickup is isloated from both the cable capacitance loading as well as the 1 meg ohm amp impedance loading and gives the upper frequencies a very clean boost. This is a good thing to build because you can make any instrument active by using this cable.

I built mine by using a jumbo metal .25 inch plug housing with the input capacitor, input resistors and FET point-to point soldered. Then I used 2 conductor mic cable, one conductor for the audio out (low impedance) with a cap on the amp end of the plug plus a piggy backed .125 inch metal jack housing that I hose clamped to the amp end .25 inch metal plug housing. Then I used a wall-wart type power supply to send the voltage to power the buffer up the second audio conductor through the .125 inch jack. This is not too much different from the Tillman design. JFET Buffer and buffer cable. If you really want to hear the effects of pickup loading, build this buffer cable, it will give you a whole new range of sounds from a guitar or bass. Just use double the capacitance values for bass to not attenuate the low frequencies too much.

Joseph Rogowski

[David Schwab]

A it is, but you might not be hearing the wire alone.

Do you know why I picked A? Is has a slightly rounder spongier tone and a little more extension on the low end. That's what larger diameter wire sounds like. This is also true comparing 40 or 41 to 42.

Then the pickup world is out of touch with the real world. In the real world a low impedance pickup is one that can be easily used with an appropriate cable without loss of bandwidth. My 2700 turn pickup must be loaded carefully in order to get the full BW, and so it is not low impedance by normal definition.

Les Paul would have agreed with you, but no one seemed to like the tone of his low z pickups very much. Then of course you have to interface it with a standard high z amp, so you will need a transformer or preamp. If you have a preamp in the bass, then the cable doesn't matter much, but still, the medium/low and high z pickups don't sound the same.

Suppose that I took my design, increased the bobbin size to allow 8,000 turns and ran it into a preamp with minimal cable. How close would it get to the extended range sound? I think pretty close.

Not really.

Here's two identical pickups. One is wound to about 2k, and the other is around 9k. They are both humbuckers using the exact same bobbins and blades, etc. Both were wound with 42 gauge wire and use ceramic magnets. Both are connected directly to the same JFET buffer. There is no coupling cap on the input, so they are DC coupled. They are also both installed on the same bass and using the same strings.

Low Z

High Z

Note that the high Z clip also includes a neck pickup later in the clip. You can hear that the low Z pickup has a more extended high end. You can easily make it sound like the high Z pickup by using a small value cap, but that defeats the purpose of winding it that way.

You can clearly hear that the high z pickup is fuller in the lows and mids and has a distinct upper midrange resonant peak. The high end doesn't extend as high ether. The low Z pickup is flatter and the resonant peak it higher. I don't see much value is flattening the resonant peak, as that is what defines the tone of the pickup.

These kinds of tests have to be done with the pickups wound past 4-5k. Most pickups wound that low will have a nice wide range tone with a bright top end. But players complain about low output pickups, and the unfortunate trend has been for louder pickups, which never sound as good IMO. Once you get enough wire on the pickup you start to develop more of a band pass response.

[Mike Sulzer]

Not really.

....

I do not understand how your clips and discussion respond to what I wrote. All I am saying is that an 8K pickup operating into a small capacitance (coil, short lead, preamp) can have a resonant frequency similar to a 2K pickup operating into the larger cable capacitance, and that they can be loaded to have similar nearly flat band widths. How well this would work out needs to be verified by making one.

[David Schwab]

I do not understand how your clips and discussion respond to what I wrote. All I am saying is that an 8K pickup operating into a small capacitance (coil, short lead, preamp) can have a resonant frequency similar to a 2K pickup operating into the larger cable capacitance, and that they can be loaded to have similar nearly flat band widths. How well this would work out needs to be verified by making one.

I posted clips of two actual pickups to use as a real life example. It's very likely that no one is going to be using a 2k pickup passive, since the output will be very low. So cable capacitance is not a factor. Also, I don't understand why you want to load the pickup that way in the first place. The only reason someone would use a low Z pickup is for the extended high frequency response. Why knock that down with a capacitor?

Alembic realized that people would want to get a similar response as a high Z pickup, so they used variable low pass filters with adjustable Q. Tuning the filter down a bit and adding that peak sounds like a pickup with more wire wound on it.

So what I demonstrated is that a pickup wound around 8k will not sound like a pickup wound around 2k, which is what you said:

How close would it get to the extended range sound? I think pretty close.

You can clearly hear that the range is not as extended, and the response is not as flat. It's not a huge difference, but it's different.

[Mike Sulzer]

This discussion is about loading with a resistor (not a capacitor), comparing to a very low impedance pickup (one that requires no particular loading to be flat), to get flat response, that is, good high frequency response.

You have not demonstrated anything about a pickup with a flat response. By the way, one reason for using a flat response is that it is easy to equalize to get whatever else you want. Also by the way, Alembic's 1500 ohm pickup into a preamp probably has a resonance above 50KHz and is nearly flat in the upper part of the bass guitar audio range, similar to what I am doing here, although I have never measured one so I could be off.

[rjb]

This is a Hijack. (The great extended response bass pickup design challenge)

The low impedance coil... is set up to feed a low imp. mic. input. ...gain...is flat as high as I can measure.
The medium impedance coil...feeds a standard single ended instrument input. ...it has a resonance of about 10KHz.

In the OP, it appears that Mike wished to demonstrate practically no audible difference between his low impedance and "medium impedance" pickups. The advantage of the "medium impedance" pickup is that it can be plugged directly into a standard bass amp, without a transformer or active buffer.

To Mike:
Have you tested the "medium impedance" pickup with a bass amp?

To the pickup gurus:
How would you design, construct and load an "extended response" bass pickup that requires no transformer or buffer?

Since I've never built a pickup from scratch, I'll exclude myself from the contest.

[Mike Sulzer]

To Mike:
Have you tested the "medium impedance" pickup with a bass amp?

No, I do not have easy access to one. The main disadvantage would be mental. Amps have plenty of gain, and the SNR would be adequate, but some people might be bothered by having to run the gain higher to get the same power as with a normal pickup.

[David Schwab]

This discussion is about loading with a resistor (not a capacitor), comparing to a very low impedance pickup (one that requires no particular loading to be flat), to get flat response, that is, good high frequency response.

You have not demonstrated anything about a pickup with a flat response. By the way, one reason for using a flat response is that it is easy to equalize to get whatever else you want. Also by the way, Alembic's 1500 ohm pickup into a preamp probably has a resonance above 50KHz and is nearly flat in the upper part of the bass guitar audio range, similar to what I am doing here, although I have never measured one so I could be off.

Mike, the title of your thread is "Extended response pickups for bass". I demonstrated an extended response pickup for bass. I also showed how the same pickup wound high impedance, even when connected to a high Z input on a buffer, does not have the same extended frequency response. I was commenting to your assertion that they would sound about the same. But they don't, which is what I demonstrated. I don't really care if it was flat or not, but low Z pickups tend to be pretty flat. You also said:

The most obvious way is to put a lot fewer turns on the coil and use larger wire, resulting in a so-called low impedance pickup with both low inductance and resistance. Such a pickup requires additional amplification, and is well-suited to connection to a standard low impedance microphone input, especially if it uses a two-sided balanced configuration.

Another way is to put a preamp in the bass. Of course, this removes the cable capacitance from the load, and it is no longer necessary to use such a low impedance since the load on the inductance is a higher impedance, that is, a smaller capacitance.

But it doesn't really, which is what I demonstrated. Both of your pickups were too similar to have any big difference in tone. This was similar to the thread you started on the multi coil pickups. They were both wound too low to really show the difference between a single large coil and multiple coils, and they sounded almost the same.

So maybe I'm not getting what this thread is about. It seemed to have branched off from the other thread on bass tone and the mention of low impedance pickups. All I'm saying, and have demonstrated, is that low or medium impedance pickups sound quite different from high impedance pickups, even with buffering. One is not better than the other, but if you want extended response, low impedance wins every time.

[Mike Sulzer]

The pickup and the load are a voltage divider, a circuit with a series and shunt element. The response versus frequency is a function of both elements. Therefore, in some cases, you can get similar frequency responses when the series element changes by changing the shunt element.

For example,

1. The low impedance pickup has a low series impedance, and the shunt element (cable, and amp or interface input impedance) is large enough so that there is little attenuation over a very wide frequency range.

2. The medium impedance pickup has a larger inductance (series element), and so the shunt element (cable capacitance) matters and you get a resonance and high frequency attenuation. However, if you use a shunt resistor as well, you can flatten out this response, that is, construct a voltage divider that has nearly the same frequency response over the necessary frequency range as the low impedance pickup.

3. Suppose that you continue to increase the inductance. You get a lower frequency resonance; if you load this with a resistor, the frequency response flattens out to medium frequencies, but falls at a too low a frequency to sound like the other two. (need to finish up this test.)

4. If you raise the shunt impedance (with a preamp), then using the inductance of (3.), then you still can flatten out response to a high enough frequency to have a similar sound to (1.) and (2.). It takes a larger value resistor than in (2.) to do this. (Need to complete this test)

[rjb]

3. Suppose that you continue to increase the inductance. You get a lower frequency resonance; if you load this with a resistor, the frequency response flattens out to medium frequencies, but falls at a too low a frequency to sound like the other two. (need to finish up this test.)

4. If you raise the shunt impedance (with a preamp), then using the inductance of (3.), then you still can flatten out response to a high enough frequency to have a similar sound to (1.) and (2.). It takes a larger value resistor than in (2.) to do this. (Need to complete this test)

There's nothing wrong with hands-on DIY. But, using the data you have, the results can be somewhat predicted with theory.
(Yadda, yadda, Fres equals one over two pi times square root of L times C).

As you add turns to the coil, the inductance eventually becomes so high that you cannot make the shunt capacitance low enough to avoid an Fres below your desired value.

In David's demo with the 2K and 9K (DCR) pickups, the high z pickup's inductance was presumably about 20 times higher than the low z pickup.
(9/2)^2 = 20.25
Even when DC coupled to a JFET buffer, Fres for the 9K pickup was in the "upper mids" range.

But David's pickups are blade humbuckers - which are surely higher inductance than Mike's air core pickups.

The question is how low the pickup's inductance needs to be to assure Fres above 10 KHz when loaded by a particular preamp.

[Mike Sulzer]

The question is how low the pickup's inductance needs to be to assure Fres above 10 KHz when loaded by a particular preamp.

Exactly. In addition there is the question of the effect of eddy currents on frequency response with a steel core. It is simpler to get theory, measurement, and listening to all agree, if you leave out that complication!

The 8,000 turn version of my pickup has a 2.1 something H inductance, and resonates at about 12KH when operated into an op amp follower with a few inches of lead. This should sound very much like the other two when flattened with about 100K, but no cheating, that is not tested and recorded yet! Maybe some unexpected factor will creep in. Need to compare apples to apples. I do know that if you run it straight to the cable and flatten it with a resistor, you get about 3.7 KHz BW, and that is distinctly different in sound from the wider BW pickups, of course, but I have to do the recording.