MIMF

Non-traditional guitar designs are my obsession. I designed and built the Carbon-Fiber-Bamboo guitar. Currently, I use DiMarzio Injector pickups. On my next iteration of the prototype, I am experimenting on hexaphonic pickups for hex processing.



The Cycfi Six Pack Project is an ongoing Open Source Hardware project for the development of an active hexaphonic pickup system designed for hex processing. The pickup has six low impedance coils. The hex pickup is active with six differential low-noise, low-power pre-amplifiers —one for each coil. It has the same footprint as the ubiquitous Strat single coil, with a very low profile: 8mm (0.3 inch).

Eventually, the goal is to have hexaphonic sustain drivers as well. That, and with extensive processing for each string, will give us musicians full control over the dynamics of the guitar. I know hexaphonic sustain has been done in the past with the Moog guitar, but that was a very expensive gear. I want something more affordable. And I want a system that can be adapted to just about any guitar. This IMO is the holy grail and I know this is very difficult to do right, but every journey starts with the first step.

As an Open Source Hardware project, all the designs (schematics, PCB layout, software, bill of materials, CAD drawings and source code) will be freely shared, 100% free.

I would very much love to hear your thoughts and gather ideas while the project evolves!

Hi Joel, interesting stuff.
If I'm reading right, you don't have differential coils, only a differential output amplifier for each coil, correct? You might find it worth it to try differential coils like a humbucker (either in traditional humbucker side-by-side arrangement, or as stacked coils), as keeping the noise floor lowest at the source will provide most improvement in overall signal-to-noise ratio.
I'll be interested to see how you go with a hexaphonic sustainer, and also how well your string-to-string isolation works. You might find it easier to implement a hexaphonic sustainer by using individual piezo pickups in the bridge saddles to provide better string-to-string isolation and reduce crosstalk, taking the final output from the magnetic pickup. A challenge indeed. Hopefully you end up with something that can give Brian May studio style sounds (Brian May would often track each string individually in the studio, so that he could keep the breakup dynamics of each string rather than have them interact as would normally happen when chording. He apparently tried some hexaphonic pickups, but I don't believe he ever stuck with one), individual distortion on each string!

Good luck.

Thank you, Greg!

Greg Robinson wrote:Hi Joel, interesting stuff.
If I'm reading right, you don't have differential coils, only a differential output amplifier for each coil, correct? You might find it worth it to try differential coils like a humbucker (either in traditional humbucker side-by-side arrangement, or as stacked coils), as keeping the noise floor lowest at the source will provide most improvement in overall signal-to-noise ratio.

No Greg, I have differential inputs for each coil and (pseudo differential; floating reference) single-ended outputs. The signals at each end of the coil are 180° out-of-phase while noise propagates equally through the wires in-phase. The differential amplifier subtracts the signals and the noise, hence doubling the signal while eliminating the noise. Even without any shielding, the pickups are already superior by far in terms of SNR in my initial measurements.

Anyway, I intend to do some noise measurements and I'll probably try a stacked double-coil, for science.

Greg Robinson wrote: I'll be interested to see how you go with a hexaphonic sustainer, and also how well your string-to-string isolation works. You might find it easier to implement a hexaphonic sustainer by using individual piezo pickups in the bridge saddles to provide better string-to-string isolation and reduce crosstalk, taking the final output from the magnetic pickup. A challenge indeed. Hopefully you end up with something that can give Brian May studio style sounds (Brian May would often track each string individually in the studio, so that he could keep the breakup dynamics of each string rather than have them interact as would normally happen when chording. He apparently tried some hexaphonic pickups, but I don't believe he ever stuck with one), individual distortion on each string!

Good luck.

For string isolation, I'm thinking about doing it in software with some form of phase cancellation using digital signal processing. I'll see how this works out. I'd like to minimize the components involved, but I'll definitely look into peizo pickups in the bridge if I get stuck.

Brian May, I love that guy!

Let it breathe (part 2): More Frequency Analysis

Last time we did some (very rough) preliminary tests comparing an early version of the hex pickup with a Fender Stratocaster and a Gibson Les Paul. The initial test was done using a 2000 turns AWG 42 coil and the cheapo TL072 Op Amp in differential configuration. The goal basically was to see the frequency spectrum of lower impedance coils side by side with that of the single-coil Strat and Les Paul humbucker pickups. Here we will continue with the tests, this time focusing only on the Hex pickup installed on a Strat while continually improving on various aspects of the pickup design.

Tell me what you think.

So, what does the graph show us? And please show the graph we are supposed to compare it to.

Yeah Joel, that doesn't really give us any information. Is that an FFT of a strum?
By the way, you'd do better with a low impedance pickup to use fewer turns and a larger gauge wire, if you want greater and more even bandwidth. I know you've said that you're using the opamp in floating differential mode for low noise, and that may be adequate, but you would get even better signal to noise ratio if you used a true balanced system, ie a humbucker. Food for thought anyway.

Compare to graphs on the linked blog.
Lots and lots of octave overtones here. Is that what a Guitar looks like?

Ok, had another look at the blog. Still, with only a single plucked note, it doesn't really tell us anything about bandwidth. With a large "plucked" sample group to account for the problems of consistency of plucking, and averaged results, we could gather data on comparitive bandwidth, but to actually measure the bandwidth you would want a very flat electromagnetic transducer to apply a white noise signal to the coil, and then take an FFT from that.
Easier yet would be to measure the the properties of the finished coil and then calculate or simulate the bandwidth. This is how coil design is usually undertaken.

Mark Swanson wrote:So, what does the graph show us? And please show the graph we are supposed to compare it to.

Mark, please see the second link in the post http://www.cycfi.com/2013/04/frequency-analysis/.

Greg Robinson wrote:Ok, had another look at the blog. Still, with only a single plucked note, it doesn't really tell us anything about bandwidth. With a large "plucked" sample group to account for the problems of consistency of plucking, and averaged results, we could gather data on comparitive bandwidth, but to actually measure the bandwidth you would want a very flat electromagnetic transducer to apply a white noise signal to the coil, and then take an FFT from that.
Easier yet would be to measure the the properties of the finished coil and then calculate or simulate the bandwidth. This is how coil design is usually undertaken.

Agreed. I am working on this incrementally and this is just a step in the development. I just wanted to have a quick 'taste' of the pickup's sound compared to pickups I know. I do have the lo-Z driver-coil test setup and I am starting to measure the response of the coil+differential amp using a sine sweep for 500, 1000, 2000 and 3000 turns. This is work in progress. I guess I should have made that clearer in the blog.

Joel de Guzman wrote:Agreed. I am working on this incrementally and this is just a step in the development. I just wanted to have a quick 'taste' of the pickup's sound compared to pickups I know. I do have the lo-Z driver-coil test setup and I am starting to measure the response of the coil+differential amp using a sine sweep for 500, 1000, 2000 and 3000 turns. This is work in progress. I guess I should have made that clearer in the blog.

Ah, yes. I just re-read it, guess I missed that before.
Joel, are you experimenting with winding patterns at all? The parasitic capacitance will be the main limiting factor with bandwidth, and your winding pattern will greatly influence this value.
If you want FLAT frequency response across the entire range, you'll need to take the inductance of the coil into account as well, as the LC resonance will become a problem, you'll want it to fall outside the audio band, or be sufficiently damped.

Keep up the good work!

Greg Robinson wrote:

Joel de Guzman wrote:Agreed. I am working on this incrementally and this is just a step in the development. I just wanted to have a quick 'taste' of the pickup's sound compared to pickups I know. I do have the lo-Z driver-coil test setup and I am starting to measure the response of the coil+differential amp using a sine sweep for 500, 1000, 2000 and 3000 turns. This is work in progress. I guess I should have made that clearer in the blog.

Ah, yes. I just re-read it, guess I missed that before.
Joel, are you experimenting with winding patterns at all? The parasitic capacitance will be the main limiting factor with bandwidth, and your winding pattern will greatly influence this value.
If you want FLAT frequency response across the entire range, you'll need to take the inductance of the coil into account as well, as the LC resonance will become a problem, you'll want it to fall outside the audio band, or be sufficiently damped.

Keep up the good work!

Experimenting with winding patterns sounds like a splendid idea! I haven't thought of that. How do you do it?

Anyway, preliminary tests with the sine-sweep through the driver coil reveals that the resonance peak is way beyond 20kHz for the 500, 1000, 2000 and 3000 turns (75-Ohms, 150-Ohms, 300 Ohms and 450 Ohms at D.C. respectively). That's as I expected anyway from these low-Z coils. Apart from the cutoff frequency, I was also worried about irregularities (dips and peaks) in the response. The test reveals nothing of that as well. I'll post my findings and graphs in the blog.

Greg Robinson wrote:I know you've said that you're using the opamp in floating differential mode for low noise, and that may be adequate, but you would get even better signal to noise ratio if you used a true balanced system, ie a humbucker. Food for thought anyway.

Thanks, Greg. To be honest, I've been pondering on that for quite some time now.

Humbuckers are not perfect. The CMRR is not that good and I have graphs to show (below). To get good CMRR, you need to have good matching of the coils and that is very difficult to achieve. The tolerance could go as high as 10%-20%. Differential Op Amps, on the other hand, have very good CMRR. The CMRR is limited by the matching of the resistors and you can get matched 1% or even 0.1% resistors. For these tests, I use 1% matched resistors.

In these tests, the strat and the LP go single-ended through the high-impedance input of the audio interface (as intended) while the hex pickup goes through the balanced input (again as intended). I made sure the gain of both inputs are the same with a test reference signal (1Khz sine). I am touching the bridge or strings of the Strat and LP to keep the noise down, but the hex pickup is left alone "no hands". The strings are damped with some cloth.

The first graph is the noisy single coil. The nasty peaks are the 60 Hz harmonics and there are some bad high frequency EMF as well (the shop is quite a noisy environment).



The second is the LP. It is a lot better than the strat of course, but it is not perfect. The noise is still very present.



The third is the hex pickup using the LME49726 with 6.9nV/√Hz input noise. You can see how silent it is. There's still the 60 Hz and its harmonics but the levels are vanishingly low. The high frequency EMF is totally obliterated (what you see there is the self noise of the Op Amp which increases with frequency -- I can see that same noise with the inputs shorted).



Will the use of double coils make it even better? Perhaps. It may well be that the noise we see in the humbucker graph is just from the wiring since LPs have single-ended outputs and this article: http://naiant.com/about/guitarwiring.html shows that balanced wiring on a standard electric guitar can give a lot of improvement.

I could certainly try to find out, but I am already quite happy with the noise level of the current design and a double coil design will complicate matters. One easy idea is to have the odd-even coils alternate from S-N / N-S and CW / CCW. That way the sum of the pickups will cancel some more of the noise.

Joel de Guzman wrote:

Greg Robinson wrote:I know you've said that you're using the opamp in floating differential mode for low noise, and that may be adequate, but you would get even better signal to noise ratio if you used a true balanced system, ie a humbucker. Food for thought anyway.

Thanks, Greg. To be honest, I've been pondering on that for quite some time now.

And the post above turned out to become a blog entry:
Singles or Doubles?

Here, we talk about the limitations of passive humbuckers and propose possible alternatives such as the Bifilar Coil winding. As always, I'd love to hear your thoughts!

Joel,

Thanks for your sharing this. I have to take some time to read everything, but I like it a lot so far.
I like it when new / other things get tried.
Good luck!

Hans

Joel, just a reminder that we don't allow discussions that draw content to other sites, commercial or otherwise (rules 1c, 1d).

Charlie Schultz wrote:Joel, just a reminder that we don't allow discussions that draw content to other sites, commercial or otherwise (rules 1c, 1d).

OK, understood. I'm sorry for violating your rules.

Some comments in no particular order...

If you are matching resistors, the tolerance (1%, 0.1% etc) is not relevant other than making your matching task easier. You should pick a resistor for it's noise generation characteristics (thermal noise, boltzmann etc) and then match.

Differential input amplifier reduce the common mode signals, i.e. the single that is imparted to both inputs traces/leads between the source and amplifier input. Since the amplifier is close to the pickup, there is almost no common mode pickup, and with a gain of 5, it almost defeats the purpose since the source impedance is a few hundred ohms. One should put effort into winding the pickup coil to cut out common mode EM field pickup. This can be done by careful hand winding and counter winding on a turn by turn basis. It is possible.

For a gain of five, an op-amp would seem over kill whilst also introducing more noise. Each transistor in the op-amp will introduce noise and distortion. In this case a well designed fet stage will give the gain required, low noise, low part count, small size, low current and reduce the thermal noise and distortion potential. With the amplifier close to the pickup, I see no pressing need to use a differential stage.

A strat or LP pickup might not perform to your desire but there is more to most players spec than clean frequency response.

There is some interesting data to be gleaned from a number of patents on the response of low impedance pickups, with Lace being the lowest. I know EMG/Steinberger did some low-Z work also. However, in our world, low-z usually means low H and low C which moves or removes any resonant in the response leading to a sound quality many players do not like. (One could achieve low-z whilst retaining higher H and C, but the magnets and copper would make such a pickup somewhat cumbersome!)

The FFT shows one note at one point in time (although the point is wide, compared to zero, because of the way sampling theory and FFT math works, however it is a short time span).

Veronica Merryfield wrote:Some comments in no particular order...

If you are matching resistors, the tolerance (1%, 0.1% etc) is not relevant other than making your matching task easier. You should pick a resistor for it's noise generation characteristics (thermal noise, boltzmann etc) and then match.

Agreed. I keep resistance as low as possible, without consuming too much power. I also use Vishay TNPW series with very low electrical noise: http://www.ttiinc.com/object/vishay_res ... t_thinfilm

Veronica Merryfield wrote: Differential input amplifier reduce the common mode signals, i.e. the single that is imparted to both inputs traces/leads between the source and amplifier input. Since the amplifier is close to the pickup, there is almost no common mode pickup, and with a gain of 5, it almost defeats the purpose since the source impedance is a few hundred ohms. One should put effort into winding the pickup coil to cut out common mode EM field pickup. This can be done by careful hand winding and counter winding on a turn by turn basis. It is possible.

Disagree. Keep in mind that the coil itself is a very long wire that acts like an antenna. Also, the floating connection obviates any connection to noise carrying ground or reference point. Finally, it is virtually impossible to get good matching coils even with low winding count and achieve good common mode rejection. I am not allowed to link to my blog in this forum, but I've shown that I achieved the optimal noise figures in my subsequent tests.

BTW, EMG active pickups have two coils connected to an Op Amp in a differential manner, also very close to the coils.

Veronica Merryfield wrote: However, in our world, low-z usually means low H and low C which moves or removes any resonant in the response leading to a sound quality many players do not like.

Again agreed. The point here is that you want to regain control over the resonant frequency (e.g. using state-variable filters, or in my case FIRs and IIRs through DSP). Not having the tone hard-wired for you by the pickup manufacturer gives you optimum control and freedom.