Magnetic Isolation for a Hexaphonic pickup

[Ikedude]

Hi all,

I've been working on this hexaphonic pickup for a while but I'm having trouble with cross-talk between coils.



Each string has two inductors (http://search.digikey.com/scripts/Dk...27&k=M10029-ND) wired in series with one "reverse-wound" (swapped the leads). I put one of these (http://www.kjmagnetics.com/proddetail.asp?prod=D21B-N52) on the bottom of each one, with opposite polarities inline for each string (humbucker style).

My problem is the signal from each pair of coils is mainly the string its under but also quite a bit of the neighboring strings. Is there any kind of shielding arrangement that might help isolate the coils?

I was thinking of a strip of steel between each set of coils but I'm not sure that would help. Something like this:



Would the strips need to be tall enough that the line in red is interrupted? I'm still working on trying to visualize the magnetic field.

I also built a bass one but these are far enough apart I didn't run into the same problem.

[peskywinnets]

IMHO you're gonna struggle with such bulky individual pickups.

The reason Roland make theirs so small, is so that the overall slim pickup can be slotted between the void between the bridge pickup and bridge .....even with the most 'iron man' of string benders, the string doesn't move much from it's resting postion there, which is what allows Roland to use tiny magnets/coils (therefore less crosstalk) but still be confident that the string will always be over the top of each hex pickup when it's vibrating.

So, in my opinion if you want decent separation, you got to go tiny (even if your proposed magnetic shileding route worked - and I think that might be both expensve & challenging - what happens when a string is bent? Yep, the string will travel 'off axis' and then be outside the small window/aperture you've created by magnetic shielding....result = massive string fade)

[David Schwab]

I'm surprised you are picking up the adjacent string. That's a big distance between them. I've made similar pickups had had no cross talk at all. Mine even shared the same magnet.

Maybe your cross talk is coming front he circuit board? Are you using a lot of gain?

Maybe the magnets are too strong?

[salvarsan]

The coils are too close together.
You could reduce crosstalk by staggering/offsetting adjacent coils from each other.

Excellent execution on the PCB.

[David King]

What about putting the magnets on top of the coils, if there's an iron core it shouldn't make much difference to the coil. Are the all the adjacent poles opposites?

ie.
NSNSNS
SNSNSN
EADGBE

That might help with crosstalk.
The only other thing I can think up is re-directing the adjacent fields with Mu-metal or steel to focus the fields onto the strings.

[Mike Sulzer]

Hi all,

I've been working on this hexaphonic pickup for a while but I'm having trouble with cross-talk between coils.

How much crosstalk? The pickups I have made have adjacent string pickup about 20 db down. Steel between the coils improves it by maybe 3 db if I remember right. Do you have them close enough to the string? You want the field from the vibrating string to couple well with the core of the coil below it in order to make the crosstalk low.

Have you measured the static field at string height? I think your magnets might be a bit small for location on the bottom. (This does not affect the cross talk, but it does affect the signal level.) On top, I use magnets half as thick. Below, I use a bit larger diameter.

[David Schwab]

The coils are too close together.

You think so? Look how close Roland guitar hex pickup are to each other.

These have quite a bit of space in there.

[Ikedude]

I don't really see how staggering the coils would make a difference. The D string coils pick up the A and G strings regardless of the placement of the adjacent coils.

Right now the magnets are set up:

NNNNNN
SSSSSS
EADGBE

Would alternating make the problem worse? Wouldn't the field lines couple between adjacent coils more than they currently do?

My problem might be from what Mike Sulzer mentioned, I'm just too far away from the strings. Magnetic fields drop off in intensity with r^2 right? So, the closer I can be to the main string, even if it puts me a little closer to the adjacent strings, the better off I am? A separation of 20 dB would be great, more is always better, but with some of the testing I've done that should be enough. I haven't measured the field above each coil, I don't have equipment to do that, but my signal is around 60 mV for each pair of coils.

The circuit might be bringing in some crosstalk, I'm gonna try and think of a good way to test that.

[David King]

Raising the pickups would bring each coil closer to it's intended string relative to the adjacent strings by a substantial proportion. If the adjacent magnetic fields are opposite there should be less signal from adjacent strings coming through but only after the pickups are summed so that signal cancelation can occur, not so useful here probably.

Perhaps you could sacrifice one of the coils by sawing through it top to bottom or dissecting it for us to see. I'm curious if the core is small enough to focus the magnetic field tightly under the string. (If focusing a field is even a possibility, I'm not sure at all.)

[Mike Sulzer]

Right now the magnets are set up:

NNNNNN
SSSSSS
EADGBE

Would alternating make the problem worse? Wouldn't the field lines couple between adjacent coils more than they currently do?

I think the coupling is directly from an adjacent magnetized string. It is fun to think of ways to locate the coils and some "dummy magnets" so that a coil always sees two parts of an adjacent string magnetized with opposite polarity, but equal amplitude. I have not figured out a practical way to do this.

[jason lollar]

cross talk is something I have always had to deal with- weaker magnets with steel plates between coils will dramatically reduce it

[Mike Sulzer]

cross talk is something I have always had to deal with- weaker magnets with steel plates between coils will dramatically reduce it

Both the crosstalk and the main signal vary with the strength of the magnets; I do not see why the field strength would change the relative level of the crosstalk. Do you mean that when you use the steel, you can get the same field strength at the strings with weaker magnets?

I did notice some improvement with steel separating the coils, but it is not that many db. All the steel can really do is divert some of the field from an adjacent vibrating string down through it (the steel) and back up to that adjacent string through the adjacent coil, rather than letting it pass through the coil you are trying to keep it out of.

[peskywinnets]

At the risk of repeating & stating the obvious - guitars by there very nature are in the main gonna be prone to a lot of crosstalk.

If you achieve perfect separation for each string, then this *will* only result in string signal fade as & when the player bends a string out of one hex pickup's 'sweet spot'...and into another's, you've only two options...

1. Ask the guitarist not to bend the string (!)

2. Put the hex pickup in a location where the string can't be physically bent away from the 'sweet spot' (as close to the bridge as possible)

...it helps to get each hex pickup as close to the string as possible & use weak magnets, thereby ensuring that only the string sitting immediately above generates any meaningful signal. (a roland hex individual humbucker only measures 50 ohms - they use a fair amount of onboard gain to get the signal up to a useable level)

[CRU JONES]

Very Nice Ikedude!

I like it! I see the USB cord there in the background... I think you might be up to something

You are correct to align all your poles the same way, this will keep the flux running the length of the string.

Here is one rule of flux path to think about: Flux travels the path of shortest distance from North Pole to South Pole.

If you could shorten the distance between your poles sets to make each pole set's working air gap significantly smaller than your string spacing, isolation would improve. (look up "magnetic gradient" in magnetic circuit book)

Fringe and Leakage flux is the flux that is not ending up in your working air gap. Where you put your magnets, for instance, could possibly reduce your leakage flux. Mike Sulzer has experience with magnet placement and hex pickups that is greater than my own, so if he says,"No", I am inclined to believe him that it will not lead far.

You could use field repulsion to isolate your working gaps. The "repulsion magnets" (oooh! ) would have their poles aligned perpendicularly to your pole sets. In this case it would be necessary to alternate pole set alignment. This diagram might be confusing...

Ss|nNn|sSs|nN
Nn|sSs|nNn|sS

S
N are your pole pieces

n|s represents a magnet inserted perpendicularly

Anyway... Just an idea! That type of thing IS done in magnetic circuit design. Examples would be in the chapters dealing with fringe flux.

Cheers,
Ethan

[Joe Gwinn]

The coils are too close together.
You could reduce crosstalk by staggering/offsetting adjacent coils from each other.

Excellent execution on the PCB.

I agree with Salversan, but with one addition: It's coil pairs (one pair per string) that I would stagger.

Also, is there a magnetic path between coils of a pair other than air and plastic? If not, there will be far more cross-coupling.

Smaller coils will also help. The closer the two coils of a pair are, the more focused the sensitive volume.

The 1/r^2 law applies only to isolated poles. For a dipole (like the two coils of a pair), the sensitivity falls off with 1/r^3.