Minimum number of turns needed for an active pickup?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
A typical pickup with 5-10H inductance exhibits a peak impedance in the range of 500kOhm.
Noise density of a 500kOhm resistor is 80nV/sqrt(Hz).
While a 500k resistor is quite noisy, 500k of inductive reactance is not equally noisy.

Only the real part of a complex impedance generates thermal noise. Pure inductive or capactive reactance does not cause thermal noise.

See for example: Johnson–Nyquist noise - Wikipedia

While the reactance of a humbucker coil does not generate thermal noise, the 500k volume pot in the typical humbucker-equipped guitar certainly does. With the volume pot set to half-resistance, the resistive part of the guitar source impedance is a bit over 125 kilo ohms, and that makes plenty of thermal noise.

I agree that the humble TL072 has been quite adequate when it comes to noise, in every guitar circuit I've built using traditional high-Z guitar pickups.

In recent years I've come to believe that opamps are not the right way to amplify traditional magnetic guitar pickups, not because opamps are flawed, but because of the opposite: opamps are precise and accurate, and the electric guitar is fundamentally not a good-sounding instrument if you precisely and accurately reproduce what comes out of the pickups. Instead you need to squash and smooth out and slightly distort the pickup signal if you want it to sound musically attractive.

There are a few plain-Jane JFETs out there with low flicker noise down to well below 100 Hz, and these make nice low-noise input stages for traditional guitar pickups. Quieter than the traditional half-12AX7, though they may not sound quite as good.

-Gnobuddy
 
~1970, I rewound an alnico bar magnet pickup with maybe 600 turns and used a little microphone matching transformer - it would have been good for some electric bass work but too "hi-fi" for guitar - output was adequate for typical commercial tube heads. Transformers are cool for some pickups.
 
...maybe 600 turns...too "hi-fi" for guitar
I think some drastic EQ/filtering will need to be part of the circuitry, otherwise you'll get that "too hi-fi" sound.

Some of the Fishman Fluence demos I heard a few years ago seemed to suffer from the same problem. There was too much extended treble response to sound like a "proper" contemporary humbucker. Fishman was claiming that proper original PAF humbuckers sounded like their Fluence version, with low output and wide frequency range. I've never heard real vintage PAF humbuckers, so I can't comment on that claim.

-Gnobuddy
 

PRR

Member
Joined 2003
Paid Member
While a 500k resistor is quite noisy, 500k of inductive reactance is not equally noisy....

Pure inductance is zero hiss.

But this is not a pure inductance! It is an RLC network with both series and shunt resistors. Do your tuned-circuit math.

10H is not 500K until 8KHz, which is not only above the good guitar band, it is above the typical 3KHz resonance with the winding and cable capacitance. The impedance peaks and falls-off. Like any tuned circuit the impedance (resistance) at the peak can be higher than the reactive impedances.
 
Pure inductance is zero hiss.
Indeed, as shown in the link I cited in my previous post. It is Re[Z], the resistive part, that shows up in the expressions for thermal noise voltage, and not the complex Z itself. Re[Z] for a guitar pickup is just the DC resistance of the windings, which is only a few kilo ohms. That is minor compared to the guitar volume control, except when the volume control is set to full (or to nearly zero.)

On the guitar side of things, source impedance is a bit more than 250k (resistive), worst case, with the volume pot set to half-resistance. The 1M grid bias resistor in the amp will end up in parallel, but thermal noise voltage changes rather slowly with changes in R (~square root of R), so it won't make a whole lot of difference.

Anecdotally, none of my guitar amps hisses louder when you turn down the guitar volume knob (raising the source impedance.) It would seem that either the thermal noise is too low to hear, or other sources of noise in the amp swamp out thermal noise from the guitar itself.

-Gnobuddy
 
Last edited:
On the guitar side of things, source impedance is a bit more than 250k (resistive), worst case, with the volume pot set to half-resistance. The 1M grid bias resistor in the amp will end up in parallel, but thermal noise voltage changes rather slowly with changes in R (~square root of R), so it won't make a whole lot of difference.

Anecdotally, none of my guitar amps hisses louder when you turn down the guitar volume knob (raising the source impedance.) It would seem that either the thermal noise is too low to hear, or other sources of noise in the amp swamp out thermal noise from the guitar itself.


Guitars with more than one PU's usually are wired in a way that the hot PU leads are connected to the volume controls' wipers, with the control pots' ends in parallel serving as the common output. Hence minimum Zout occurs at full volume and vice versa.
Best regards!
 
If you guys are worried about both e_n and i_n in a package that needs to be miserly on power, have y'all looked at some newer opamps? The opa1652 (or the single/quad) comes to mind. Unless you're trying to engineer the sound at the pickup, which case that will be very neutral.
 
If you are asking how many turns do you need around the magnet, it's 1.
Than you do need a transformer. Like Lace alumitone pickups.

I tried it and it seems like a neat concept, it does work well, with low noise.
I'd be interested the transformer could be omitted, maybe with a transconductance amplifier or something like that.
 
Than you do need a transformer.
For guitar, the requirements on the transformer will be very relaxed compared to Hi-Fi. All we need is 100 Hz - 5 kHz +/- 3 dB, which should be pretty easy to manage.

I'm wondering if a dirt-cheap surplus output transformer from the days of push-pull transformer-output transistor AM radios would do the trick. Those were designed for about the same bandwidth, and if you use the transformer backwards, there might be enough voltage step-up ratio to be useful.

-Gnobuddy
 
I'm using a 220 to 9 volt transformer for testing. I removed the windings on the 9 volt side. The other side has 2.5k DC resistance, no idea how many turns. It works, but I'm planning to build a test jig to see how it compares to traditional pickups. By ear I have to say it sounds very transparent.
 
https://i.imgur.com/LoJdWz4.jpg

This is how it looks like.
Soldering the 10mm2 copper wire was no easy task.
Naked like this it's pretty noise, but I tried covering it with aluminium tape, and it made it virtually silent. Much quite than the donor pickup was in its original form.

I'm going to get a stronger magnet, that may boost volume. The donor pickup I used have a very weak magnet. The signal is not too hot.

It works fine on the microphone input of my audio interface. With a hi z input it picks up all kinds of noise, I sort of get why that is, but not enough to explain it. Ideally this would need a preamp with an impedance tuned exactly to the transformer.

I already bough a bunch of E ferrite cores to create custom transformers for, try out different configuration.
 
Last edited:
Soldering the 10mm2 copper wire was no easy task.
Well, that's certainly unique!

I don't think there is an advantage to using extremely thick wire like that. You get the same induced voltage for the same number of turns of wire, whatever the thickness of the wire itself is.

It's true that you get a larger induced current in a shorted turn by using thicker wire - but this is going to be somewhat offset by the fact that only a small part of that single turn is actually in a changing magnetic field (the rest is leading off to the transformer.)

Winding, say, a hundred turns of 36 AWG wire on a pickup bobbin is a very simple and easy task. The wire is thick enough not to be fragile, and thin enough to handle easily. And it will generate a hundred times as much signal voltage as a single turn.

-Gnobuddy
 
Guitars with more than one PU's usually are wired in a way that the hot PU leads are connected to the volume controls' wipers, with the control pots' ends in parallel serving as the common output. Hence minimum Zout occurs at full volume and vice versa.
Interesting, though I've never encountered a guitar wired that way myself.

There is a very easy way to determine Re[Z] for a guitar with passive pickups, no matter what the guitar's internal wiring might be: use a DMM wired across the output jack! You can even do this with the guitar plugged into a (tube) amplifier to include the effect of the grid bias resistor of the input stage.

For this particular problem, I think this might be the simplest solution, rather than trying to calculate from first principles, which is hard because there are so many unknowns (pickup inductances, pickup DC resistances, wiring scheme, pot resistances, etc.)

-Gnobuddy
 
...and the Pups are wired out of phase...
Years ago I built my first electric guitar (because I couldn't afford to buy one.) I installed a DPDT toggle switch that let me reverse the polarity of one PU at the flip of a switch.

Sadly, this only led to the discovery that I was not Peter Green. :) I didn't particularly care for the out-of-phase sound, and wound up rarely using it.

-Gnobuddy
 
In a less experimental design, most of the wire would be around the magnet, and just a small piece would lead to the transformer.

The difference between the thickness of the wire is quite drastic, so i would not say it's pointless.

One advantage of this, is because there is only a single loop of wire, it does not pick up interference. Although this is offset by the fact that there is a transformer.

But in the end, i will probably go with a simple wound pickup with around 100 turns as you say, and an active pre-amp.

By the way, this is a similar design to Lace alumitone pickups. Without an active preamp, there pickups are very picky i hear, with some amps they sound great, with other they sound terrible.
 
The difference between the thickness of the wire is quite drastic, so i would not say it's pointless.
Interesting, I would not have expected that.

Can you share more about this? Did you actually hear a big change in tone when the wire was made thicker?

I've never heard an Alumitone, though I've seen ads for them in the guitar magazines. I didn't realize they were single-turn pickups.

-Gnobuddy
 
Disabled Account
Joined 2010
To achieve similar results as conventional wound pickups, copper fill factor should be the same.
Otherwise reduced copper fill factor increases resistive loss of winding, damping the resonant peak, making sound dull and boring.
So to come close to the sound of conventional pickups, the 1 turn winding requiries real bulky "wire".
 
Last edited:
To achieve similar results as conventional wound pickups, copper fill factor should be the same.
There are several scaling laws at work when you change the number of turns in an inductance. So many things change, some very dramatically, that I'm having a hard time seeing how maintaining the same fill factor will get you the same resonant frequency and Q.

For example: inductance scales as the square of the number of turns of wire. So 1 turn will have twenty-five million times less inductance than 5000 turns. :zombie:

For the same capacitance, resonant frequency goes like the square root of the inductance. So if the inductance is 25,000,000 times less, the resonant frequency will be 5000 times higher. :eek: Or you need to add 25,000,000 times as much capacitance as a conventional guitar cable + guitar amplifier input capacitance provides. :zombie:

Resonant Q is trickier. The Q is (2 pi f L/R), where L is the inductance and R the resistance in the circuit.

We want the same resonance frequency, so (2 pi f) is the same as before. But L is 25,000,000 times less, so the only way to get the same Q would be to make R 25,000,000 times smaller as well.

If we assume 8000 ohms for the conventional pickup, dividing that by twenty-five million gets you down to 0.00032 ohms. :xeye: That's far too low to be practical - maybe a single turn of very fat wire can be that low, but you have to connect a guitar cable and external capacitance, and those things will have too much resistance.

Maybe I made a mistake somewhere in all of this (I didn't consider what effect a step-up transformer might have, for instance.)

But, unless I made a mistake, I think the only way you can get a 1-turn pickup to sound like a conventional 5000 - 8000 turn pickup would be to electronically create the right frequency response peak in the preamp. In other words, it has to be an active pickup, or have the preamp built into the guitar.

Earlier in this thread, we estimated that those Fishman Fluence pickups might have about 500 turns in the coil. Fishman chose to make the pickups active, rather than putting transformers into them. I'll bet the electronics includes heavy frequency-shaping, to generate a traditional-sounding resonant peak, and then to remove the excessive high frequency response which you're going to get with a 500-turn coil.

But I have never made a 1-turn pickup, so I have no experimental data to go on. I've only wound two guitar pickups in my life. Both those had about 4000 turns each, and they did sound too clean, and too Hi-Fi compared to normal guitar pickups.

-Gnobuddy
 

PRR

Member
Joined 2003
Paid Member
I didn't consider what effect a step-up transformer might have, for instance.

But that is key of the concept. We can, in theory, transform any number of turns to any other number of turns. The prime parameters all scale correctly.

Of course there are parasitics everywhere. And when a system is pushed "to the most", it will become parasitic-limited.

Here Nemuikuma has pushed to a practical limit. He took the copper of a thousands-turns winding and beat it into a heavy bar of one turn. When run into a 1:1000s transformer, the output will be essentially like a thousands-turns winding. At this re-scaling he hit one practical limit (can hardly have less than one turn in a closed circuit) and another practical limit (bending heavy copper rod is hard work; and joining needs more than a Princess solder-iron).

He obviously knows he can't bring this 0.001 Ohm source out through 30 feet of thin wire; he has the transformer right AT the pickup, without even a wire-splice.

What he's done is move the stray L and C troubles/flavor from the pickup over to the transformer. A guitar-impedance transformer is crunched between parasitic L and parasitic C. There may be more leeway on dimensions and materials away from the pickup point. It may be much work to optimize.

FWIW, the hum sensitivity is almost unchanged. The one turn induces a very low hum voltage, but the transformation to bring signal up to useful level (with available amplifying devices) brings the hum up the same amount. Ribbon microphones have a similar problem, and must use a balanced return path so the conductor-triplet is fairly hum-bucking for hum sources more than a few inches away.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.