Hexaphonic Pickup Project

[cycfi]

[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.]

Here's the full E-Core simulation:



It looks good! Alas, it's not possible with the current bobbin design :-(. There's no way for the magnetic disk to fit at the bottom and touch the ring. I can only fit a small disk underneath as I did in the simulations above. Here's a pic of the new bobbins with coil:



That's 6mm diameter! The total height is 4.5 mm including the SMD terminals. Amazing little creatures! We successfully wound 1000 turns gauge 46 in there. The inside diameter is a mere 2.8mm.

[Mike Sulzer]

A couple of comments:
You are running FEMM in the 2D mode. The 3D equivalent of your simulations has everything continuing back into the page. This is significantly different from the true 3D case.

Simulating the permanent magnet system does not tell you what you need to know. The only purpose of the permanent magnets to to magnetize the strings; it is the field at the strings that counts, not elsewhere. As for leakage, you can get a better idea what is happening by placing a small magnet (the string) over one coil with the field directed towards the coil. Look at the relative levels of flux through the coil below and other coils. It is not very easy to relate such results to reality since you have only the 2D mode and the mode with cylindrical symmetry in FEMM.

[cycfi]

A couple of comments:
You are running FEMM in the 2D mode. The 3D equivalent of your simulations has everything continuing back into the page. This is significantly different from the true 3D case.

Nods. Indeed. There are lots of things missing and can't be taken into account in a 2D simulation. First thing that came to my mind are the rings and the fact that they wrap around. With 2D, you only get a cross-section.

Simulating the permanent magnet system does not tell you what you need to know. The only purpose of the permanent magnets to to magnetize the strings; it is the field at the strings that counts, not elsewhere. As for leakage, you can get a better idea what is happening by placing a small magnet (the string) over one coil with the field directed towards the coil. Look at the relative levels of flux through the coil below and other coils. It is not very easy to relate such results to reality since you have only the 2D mode and the mode with cylindrical symmetry in FEMM.

Hmmm. Interesting... Something like these?

The first one is the coils only:



The second is the coils plus the rings. I can see how the rings deflect the flux away from the adjacent coils:



Makes a lot of sense.

[Mike Sulzer]

Nods. Indeed. There are lots of things missing and can't be taken into account in a 2D simulation. First thing that came to my mind are the rings and the fact that they wrap around. With 2D, you only get a cross-section.



Hmmm. Interesting... Something like these?

Makes a lot of sense.

Yes, but if those are field lines, I think they should be pointing down so that the central ray passes straight through the coil

[cycfi]

Yes, but if those are field lines, I think they should be pointing down so that the central ray passes straight through the coil

Right. I also intensified the colors and displayed more lines to make it more evident. Now it's pretty visible what's going on:

Coils Only:



With Rings:



Pretty neat, if I may say! Thanks, Mike!

[rjb]

The primary question is "how effectively do the shields prevent crosstalk?"
My secondary question is "with the shields in place, how close must the pickup be to effectively magnetize the strings?"
I ask because cycfi mentioned the Roland cited restriction of 1mm from the strings as too restrictive.

[cycfi]

The primary question is "how effectively do the shields prevent crosstalk?"
My secondary question is "with the shields in place, how close must the pickup be to effectively magnetize the strings?"
I ask because cycfi mentioned the Roland cited restriction of 1mm from the strings as too restrictive.

Well, 1) In the simulation above, the the intensity of the signal coming from the strings gets through to the adjacent coils a lot more (The color gradient is not linear) and clearly shows that the shields do their work in deflecting the flux, hence the flux density is a lot less (almost blue). 2) I'd say it's the same as without the shields (judging from the simulation). Of course, as Mike said, the simulation is far from perfect and the bottom line is that only real world testing can give real answers. But it's nice to have a rough idea and visualization in any case.

PS. Hey, call me Joel :-)

[rjb]

PS. Hey, call me Joel :-)

OK, Joel.

[Mike Sulzer]

There is still more to do with FEMM while you are in the mood. Suppose you put a piece of ferrite in the core of each coil. This increases the flux through the coil from the vibrating string. How much does this increase the output? Note that it is the change in flux when you move the "string magnet" up or down a small amount that really counts. Thus in comparing the case of ferrite vs. no ferrite, it might be necessary to do more than just look at the flux.

[cycfi]

There is still more to do with FEMM while you are in the mood. Suppose you put a piece of ferrite in the core of each coil. This increases the flux through the coil from the vibrating string. How much does this increase the output? Note that it is the change in flux when you move the "string magnet" up or down a small amount that really counts. Thus in comparing the case of ferrite vs. no ferrite, it might be necessary to do more than just look at the flux.

Here's the simulation with ferrite cores:



A quick observation with the static simulation above is that, while the ferrite can increase the output (assuming the magnetization of the strings is the same as in the non-ferrite case), there might be more crosstalk.

It is possible to do AC simulation with moving strings (the PU and strings are basically an AC generator) using LUA scripting in FEMM 4.0. I'm not sure I want to go there though. It might be easier to just test the real thing.

[Mike Sulzer]

The true 3D case is probably a lot different in this case since there is much less contact between the pieces of ferrite. In any case, it the the component of field along the axis of the coil that matters.

You can use an ac gnerator, yes, but that does not necessarily reproduce the effect of a moving magnet accurately.

[Rick Turner]

All the requirements for this project would be much easier served by using individual piezo pickups. You can get fantastic frequency response, a small package, and incredible string to string isolation with no worries about bending strings from one pickup's coverage over into it's neighbor's. Piezos can be designed for great lateral string response, too, and the response is really quick because of that if you choose to use a synth interface. And...you're not limited to steel strings.

Piezos rule!

[cycfi]

All the requirements for this project would be much easier served by using individual piezo pickups. You can get fantastic frequency response, a small package, and incredible string to string isolation with no worries about bending strings from one pickup's coverage over into it's neighbor's. Piezos can be designed for great lateral string response, too, and the response is really quick because of that if you choose to use a synth interface. And...you're not limited to steel strings.

Piezos rule!

If piezos are all that great, then why do electric guitars still use magnetic pickups? The sound of course! Piezos sound differently due to a lot of factors. If I were to simply track the notes and generate MIDI, I'd use piezos.

Magnetics rock! :-)

[Rick Turner]

It's highly unlikely that a hex magnetic pickup that has any kind of decent separation...with these tiny coils...is going to sound anything like a traditional Strat pickup. Your stated requirements do not add up to what we know as a Strat sound, and you're looking for individual string sound processing.

I never said that piezos sound like magnetic pickups...though most people haven't heard what good piezos properly buffered can do...please actually read my previous post to see why I suggested that in this particular application, piezos are worthy of serious consideration. Once I figured out the trick of coupling to both pressure and shear modes with piezo crystals, I found an incredibly great sounding system with "whales to bats" frequency response and equally incredibly phase response. The sound is very, very musical as audio, and it drives synth converters like nobody's business...significantly better and more glitch-free than anything else I've seen or heard. You can hear how well it works with nylon strings on any of the live Lindsey Buckingham/Fleetwood Mac versions of "Big Love". He's able to absolutely thrash the guitar at the end of the song without causing synth freak out. Yes, he does mix the audio with the synth, too...

Of course I realize that because you've put so much time into the magnetic direction, you have a lot invested in it being the best approach. I think you might want to open your mind to other possibilities. Piezos also take up very little real estate.

BTW, I made individual string magnetic pickups as far back in ancient history as 1973...Phil Lesh's "Mission Control" bass...yes, I was "there" forty years ago. The pickups were crude but sounded great. I did a fully adjustable hex guitar pickup...magnetic with Samarium cobalt magnets wound with 46 ga. wire...for a no-go guitar synth project for Gary Kahler in 1984. I'll dig it up and post a picture...the individual pickups were adjustable for height and string spacing, and that version of the pickup could mount to a Tunamatic bridge. Great separation, string to string.

All that said, there's a big difference between essentially using a mag pickup as an effect and needing a phase accurate pickup with great string to string separation. When people talk about liking "the sound" of a magnetic pickup, they're essentially talking about a using transducer as an effect...and don't get me wrong, I do this too. I wind my own for my Model 1 guitars and for my Electroline series of basses. But when I want individual string output or stereo mixing, etc., I now greatly prefer piezos...properly designed and properly buffered, of course...which not many are...which is why I make my own.

[RedHouse]

Wurd.