Hexaphonic Pickup Project

[rjb]

for analog processing... is -20db good enough ...? I am getting -38 db, so should I put the cross-talk issue to rest?

Just for reference, does anyone know the cross-talk specs of any of the commercial hex pickups?

[cycfi]

Ralph, you might be interested to know that I will be attempting a (hopefully) final experiment with:

1) Smaller diameter magnets (2.8 mm; down from 5 mm)
2) Smaller diameter coils (6 mm; down from 10 mm)
3) Mu-metal shielding around the coil (to minimize flux leakage)
4) Magnets inserted perpendicularly between the coils (as suggested here)

I'm already getting -38 db crosstalk, which is already quite good, but Steve has a very valid point with the hex-fuzz issue. A distortion effect can have gains as high as 100 (e.g. Ibanez tube screamer) which amounts to 40 db. With that in mind, you'd want to have as little cross-talk as possible. I repeatedly say that I intend to do cross-talk reduction in software, but with the nice results I am getting, I am inclined to think that it is a good idea to further reduce (if not eliminate) cross-talk in the pickup itself. That would make cross-talk cancellation in software an unnecessary step and allow pure analog applications and maybe even MIDI conversion.

Thanks for the comments, Y'all!

[cycfi]

Just for reference, does anyone know the cross-talk specs of any of the commercial hex pickups?

BTW, I asked the same question in the VG guitar forum. Elantric replied that with a properly setup GK-3 hex pickup, greater than -45dB adjacent string crosstalk rejection is obtainable. That means for example that the PU should not be greater than 1mm away from the strings.

FWIW, in my tests, I am getting -38 db without any extra effort (I have my hex-pickup at around 3mm from the strings).

[salvarsan]

... a properly setup GK-3 hex pickup, greater than -45dB adjacent string crosstalk rejection is obtainable.

FWIW, in my tests, I am getting -38 db without any extra effort ....

The GK-3 midi pickups are hum-canceling types, small blades with a focused magnetic field.
Good to know the design tallies at -45dB crosstalk.

For comparison, a properly set up turntable arm is good for -34dB crosstalk between channels on a stereo vinyl record.

[Joe Gwinn]

3) Mu-metal shielding around the coil (to minimize flux leakage)

Be aware that mu metal loses most of its magnetic properties when bent or formed. A hydrogen anneal is required to restore the magnetic properties.

Also be aware that mu metal is easily saturated, magnetically.

Mu-metal - Wikipedia, the free encyclopedia

A less fussy material may work better. Like mild steel.

[cycfi]

A less fussy material may work better. Like mild steel.

Thanks, Joe. I'll try mild steel.

[rjb]

The GK-3 midi pickups are hum-canceling types, small blades with a focused magnetic field.

I'm just tossing out some mild entertainment while we wait for cycfi's latest research results....

I was somewhat surprised to learn that some game controllers contain an actual (albeit cheaply made) hex pickup:
Unusual source for Hex pickups

Note how the magnetic field is focused, using a single-coil per string design (see the steel bar next to each coil).
I have no idea how good the crosstalk specs are.



Compared to the Roland GK-2A, the Gigmaster pickup looks pretty tall- too tall for surface mounting.

[Mike Sulzer]

I

Note how the magnetic field is focused, using a single-coil per string design (see the steel bar next to each coil).

Why just a steel bar rather than a tube surrounding the coil?

[peskywinnets]

I'm just tossing out some mild entertainment while we wait for cycfi's latest research results....

I was somewhat surprised to learn that some game controllers contain an actual (albeit cheaply made) hex pickup:
Unusual source for Hex pickups

Woah...that manufacturer must have been totally scorched by that product ....that's one helluva lot of R&D/production/marketing costs for a product that's crashed & burned (for all I'm sure it sucks as a guitar, it's still an impressively well finished product at first glance - http://www.youtube.com/watch?v=rdyQRLn5TKA) ...still, nice find & got to be worth a punt on a secondhand one for the hex pickup alone!

[rjb]

Why just a steel bar rather than a tube surrounding the coil?

I presume the main reason is because it's cheaper and/or easier to assemble.

But I'm really not sure a tube would work better.

In the Gigmaster design, the core and steel bar are aligned along the string- as are the oppositely charged magnets in the Roland humbucking design. AFAIK, the Roland design doesn't use any additional magnetic guarding between strings.

Would a tube cause more flux to spread out nearer adjacent strings? I don't know.

I might ask "Why not use smaller, rectangular cores?"- and I'd guess the answer is because it's easier to use off-the-shelf cylindrical inductor bobbins.

[cycfi]

Here are some updates (towards v1.3). We have new SMD bobbins (5.9mm diameter, 2.8mm core diameter, 3.3mm height) with new neo-cores (customized N48, 2.7mm diameter, 3.5mm height). In addition, I ordered matching ferrite rings (10mm outside diameter, 6mm inside diameter, 3.5mm height) that perfectly wrap the coils. Also in transit are 3.5mm mu-metal strips as well as permalloy strips that could also wrap around the coils (possibly multiple layers). Finally, I also have new small neo bars 4mm x 10mm x 2mm that could serve as fringe flux barriers. I'll try each of these solutions for effectivity.

We can get 1000 turns into these tiny bobbins with 46 gauge. Winding was a feat with our simple in-house winder, but the folks at Chipsen are properly equipped and are experts in this field (they manufacture SMD inductors).

I'm very excited! I expect these new stuff to rock! :-) I'll keep you guys informed on the developments. I was just informed that the new (v1.3) PCBs are about to be shipped.

[cycfi]

AFAIK, the Roland design doesn't use any additional magnetic guarding between strings.

Note that with the GK, you'll have to have the pickup 1mm away from the strings to get -45dB crosstalk. People at the VG forum find that very difficult to do especially with different string/bridge radiuses on different guitars. You'll also have to have the PU located no more than 20mm away from the bridge. IMO, these constraints are too limiting. I got the decent -38 db crosstalk with v1.2 even without trying anything and with the poles at around 3mm away from the strings and with the PU at the Strat's bridge PU position.

[rjb]

We have new SMD bobbins (5.9mm diameter, 2.8mm core diameter, 3.3mm height) with new neo-cores (customized N48, 2.7mm diameter, 3.5mm height). In addition, I ordered matching ferrite rings (10mm outside diameter, 6mm inside diameter, 3.5mm height) that perfectly wrap the coils.

I guess I should be embarrassed for not knowing the answers, but...

1) For best isolation, should the neo cores and surrounding ferrite rings sit on 10mm disks of some ferrous material?

2) If such disks were employed, would higher permeability material result in higher output?
Ref http://music-electronics-forum.com/t22460/#post188148

[rjb]

Dumb guy raises his hand again

Just trying to test my understanding, or lack thereof.
Could someone with working knowledge of fringe/leakage flux in various physical configurations tell me if these statements are true?
Thanks.

- In the new design, the ferrite rings serve two functions: to divert flux from adjacent strings from the coil; and to control the flux path of the neo pole.
- If the pole and the ring sit on a disk of ferrous material, then the pole and ring will have opposite magnetic polarities at the top of the pickup.
- You then have a working gap between the magnet and ring of 1.65mm (1/2 the difference between magnet diameter and ring inner diameter).
- The string will be magnetized by the fringe flux across the working gap- which will extend 1.65mm (the length of the gap) above the pole.(????)
- You can increase the working distance between the pickup and string by increasing the ferrite ring's ID- but at the expense of increased cross-talk.
- The permeability of the ring material has (great, little, no) effect on the working distance between the pickup and string.
- The mentioned ferrous disk is totally unnecessary because (fill in the blank).

Refs
http://music-electronics-forum.com/t26380/#post226819
http://www.kgmagnetics.org/APNOTES-06/An-115.pdf

[cycfi]

Hello RJB,

These are very good questions. I'll try to answer these as best as I can. Anyone please correct me if I am wrong.

Just trying to test my understanding, or lack thereof.
Could someone with working knowledge of fringe/leakage flux in various physical configurations tell me if these statements are true?
Thanks.

- In the new design, the ferrite rings serve two functions: to divert flux from adjacent strings from the coil; and to control the flux path of the neo pole.

Yes.

- If the pole and the ring sit on a disk of ferrous material, then the pole and ring will have opposite magnetic polarities at the top of the pickup.

Yes.

- You then have a working gap between the magnet and ring of 1.65mm (1/2 the difference between magnet diameter and ring inner diameter).
- The string will be magnetized by the fringe flux across the working gap- which will extend 1.65mm (the length of the gap) above the pole.(????)
- You can increase the working distance between the pickup and string by increasing the ferrite ring's ID- but at the expense of increased cross-talk.
- The permeability of the ring material has (great, little, no) effect on the working distance between the pickup and string.

Not sure about these. The base is not level with the top of the magnet, so there's a lot more gap there. It very possible that I am misunderstanding you. [Edit: Oh I see, you want the ring and the base to form a W (or an E turned 90 degrees). Alas, that's not possible with the current bobbin. I'll try to simulate that if I can later.] (more below).

- The mentioned ferrous disk is totally unnecessary because (fill in the blank).

It's very difficult to really see where the flux lines go and analysis like this is not trivial (IMO). The best I could do is to visualize what's happening using FEMM (thanks to DarkAvenger in the Project Guitar forum for introducing me to this fine software). So here are a series of images from my analysis (the dimensions are accurate and the materials used are those that I will be experimenting on).

Here goes...

Here's the Coil Only plus the Neo magnet Core:



Now, I added ferrite rings around the coils:



If I add a stainless-steel base at the bottom, here's what happens:



Is there an improvement? I'm not sure. The top flux lines and density seems to be the same, although there's less leakage at the bottom, which I'm not sure is important. Anyway, such a large base will not fit under the SMD bobbins.

Now it gets interesting if I use magnets as barriers. Here's what happens if the barriers have the same polarity as the pickup cores:



Ugh. Doesn't look good, right? But if the barriers are reversed, then I get:



Now what do you think about that?

This FEMM stuff is so cool. I spent a lot of time with it I find it irresistible! So, for one final try, I added Mu-metal shielding and a stainless steel base at the bottom (this time I used a smaller base that can fit under the bobbin). Here it is:



Tell me what you guys think. I'd love to hear from anyone here more knowledgeable in magnetics provide a better analysis from these simulations!

At any rate, while simulations are cool, this has to be put to the test and actually try it in the real world.