I've made some tests on a dual coil just for testing on the D and G strings of my guitar. This enables me to get an idea of channel separation and other properties of the concept. I've done these tests using two different diameters on the coils and pole pieces. The coils are wind in my automatic wire winder. Click on images for full resolution.
The first test was done with coils that are 10mm diameter and a 5mm hole where a regular
guitar pickup pole piece fits. Bobbins are home made in my lathe.
The wire winder in action. 0.075mm wire is used for these coils. It's the thinnest copper
wire I've got.
The coil pair looks like this after assembly. I've made tests with and without the outermost assembly screws.
The screw heads can be helpful when holding the device over strings.
Another view reveals the pole pieces.
The device on top of the strings on the guitar.
The performance was quite promising for the device, and the channel separation figure was roughly
estimated to around 15dB or around 6 times which sounds good in clean mode. However it is not nearly good enough for
playing choruses of metal lead solos Thin Lizzy style which is pretty much what I aim for. I would
suspect separation figures of around 50dB at least would be usable. Observe that these figures are calculated
with pickup very close to strings. It would be even worse if pickup was placed at a more usual distance from strings.
Here is a listening test: test1.mp3
I also made smaller coils to see if this would improve channel separation. The amplifier board is
visible at the top.
This view gives an idea of how it's assembled. One of the pole pieces is removed. They are modified in the lathe
to fit this core diameter which is now 1mm.
In fact, the smaller coils improved the performance somewhat as can be seen in this view. Now the channel
separation figure is more like 24dB or around 16 times. Still not near the goal of 50dB-ish performance.
Here is a listening test: test2.mp3
So the quest for the perfect hexaphonic pickup still remains. I've got some ideas where the pole piece goes around the string. Maybe this would need to involve some shields between pickup coils. I've also tested to use software attenuation by subtracting the neighbouring strings and thus cancellate bleed but this has proved to yield neiglible improvement. Another idea involves strings where some current flows through them. Maybe one could pick up signal from a coil going around the string. I'm not sure if a regular coil or a toroid shaped would be optimal for that purpose. I've also thought about hall-sensors around the strings but have no idea if that would work well or if at all.