Build Project: Universal/Progressive RF Coil Winder for Theremins

I guess I'm talking more about a comparison between a zig-zag donut like on the EW chokes, and a straight bobbin wind - where there is no air in either.

Interesting article, thanks!  The author talks a lot about inter-winding capacitance...  He also mentions dielectric losses due to the coil former, which I'm ashamed to say is a new concept to me.  I thought a non-metallic coil former might magnify self-C, but not actually produce losses.  Here's a nice paper that I ran across discussing the subject:

  http://g3rbj.co.uk/wp-content/uploads/2015/11/The-Effect-of-dielectric-inside-a-coil-4.pdf

We're lucky in a way that Theremin coil self-C requirements don't drive us automatically to litz wire baskets and spiderwebs.

A straight wire has inductance - I wonder what its self-C is?

[EDIT] Ha!  It's zero: https://physics.stackexchange.com/questions/558905/capacitance-of-a-single-straight-wire

Dielectric losses can be a big deal when working with microwave materials and in many cases they dominate over conductor losses, and even for theremins lossy dielectrics in the high voltage field will lower the Q.  The dissipation factor or loss tangent (which is actually 1/Q) for RF/microwave substrates can be a key contributor to transmission line losses and a limiting factor for distributed circuit elements like printed resonators.  Dielectric losses dominate at the hot (high voltage swing, high-Z) ends of resonators, while conductor losses dominate at the low-Z drive end, and you sometimes see both types of losses referred to one end of the schematic model (for example, conductor loss at the drive end modeled as a high resistance value at the high-Z end).

"Dielectric losses dominate at the hot (high voltage swing, high-Z) ends of resonators, while conductor losses dominate at the low-Z drive end, and you sometimes see both types of losses referred to one end of the schematic model (for example, conductor loss at the drive end modeled as a high resistance value at the high-Z end)." - pitts8rh

Ah, that seems much like how coil Q can be lump modeled as low R in series or high R in parallel. 

Roger:

1. What do you think of my use of PVC (relative permittivity ~3, or so I have read) as a coil former?  I could get thinner wall tubing (and may for structural reasons) but I believe that's in an area of diminishing returns?  Paper is around 2, but add some phenolic and it could easily go much higher?  It's tough to get below 3, and I'm uncertain as to the value of that as a goal.

2. What are you target inductances for your Pro?  When the antennas are similar in area, I recommend a 1:2 ratio (pitch:volume), and maybe don't go below 0.5mH.  I'm running 1mH:2mH for P1 & P3, 4mH:8mH for P2 (coils I has lying around).  It seems, to some extent, that smaller inductors tend to have higher voltage swing, and the oscillator with the lower voltage swing tends to be the one that experiences interference from the other - if that is the case then you want the pitch axis to have the smaller inductor.

3. Assuming it could be worked out physically, do you think it might be possible for you to use an air core in the pitch arm of your Pro?  Or do you think the concentration of the ferrite is essential to the C-field shape for your style of playing (so that you aren't "playing to the arm" - an active horizontal support could be a compromise in this regard)?

4. Do you understand what's going on with the insulator over the ferrite?  Is it mainly to reduce capacitance?  I can't say I've ever read anything about it, but I abandoned ferrite fairly early in my research.

1) It's hard to beat PVC for cheap and easy, but you could at least see if you can get thinwall PVC pipe at your home center stores.  I bought some 3" and 4" pipe with 1/16" wall thickness instead of the usual 3/16" or 1/4" or whatever. If you can't get smaller diameters in thinwall you can order it in cut sections if you don't mind the concept of paying to ship plastic pipe.  And of course you have a 3D printer and you can print any size tube you want, in any wall thickness.  A single perimeter tube printed with a 0.8mm nozzle in vase mode is actually pretty stiff.

2) The target inductances for the Pro were driven entirely by what could fit in the pitch arm first, and then choosing twice that inductance for the volume.  At the time I built it nothing was firm on which side should be higher frequency, but the space limitation in the pitch arm made me choose the smaller value inductance for that side.

3) I don't have  any ferrite in either of my Pro coils, nor did I ever.   I think you may be remembering some of the ferrite-core sectioned wood-bobbin coils that I was playing with at the time (some of which still look pretty good, actually).  My Pro pitch coil is a single 1.25" long layer of #38 wire on a 3/4" diameter Garolite tube.  The only thing inside is a thin brass wire centered in the core that carries the antenna sense back to the AFE and also incidentally serves as a means of pushing the AFE a safe distance down the tube through and past the windings.

Any concerns that I had about the "hot" antenna sense line running a couple inches down the tube have gone away.  I still think that we could do better with where the capacitive divider is located and possibly use coupled lines instead of a chip cap for the series C as I told you about earlier, but fears of having a distorted pitch field because of this little compromise haven't panned out, at least not on my Pro.  I'm willing to bend the rules some more on the next one to see what I can get away with.

4)  Aside from the obvious need to protect the wire from the abrasive core, I suspect that you are right about minimizing capacitance.  The fringe E-fields between adjacent (and not only immediately adjacent) windings with different potentials run in air outside the coil and through the relative permittivity of the ferrite inside the coil. I don't remember any permittivity numbers for ferrites but I think they can run pretty high - in effect acting like a high dielectric-constant substrate, concentrating the fields in the core and thus increasing inter-winding capacitance.  So it probably makes sense to put a thick-ish spacer under the windings where possible to help keep those fields in air or low-dielectric material.

Ferrites have all kinds of non-ideal characteristics, but if you value compactness they can still look pretty good.  The D-Lev may have special needs,  but a lot of theremins are built around those pi-wound Miller coils.  I'm very interested in seeing what can be done with ferrites and various winding patterns. 

   
The ferrite core coils of high quality factor Q (for antenna extension / linearization) I currently use are sectional wound (other members of TW use similar types of coil constructions) apply 3D printed bobbins which are adjustable and provide e.g. 65mH / 0.76pF/ 716kHz SRF and 122mH / 0.77pF / 518kHz SRF. As much air as possible is provided to reduce dielectric losses introduced by the coil former and thus not to reduce Q. The ferrite rod (generally applied in radio controlled alarm clocks) has a typical Al factor (magnetic conductance) of approx. 40. In general, to keep the influence of temperature low, this factor shall be chosen quite low.
The dielectric losses of the large vintage phenolic tubing coil formers (e.g. RCA) must have been quite large, which, in addition to the high DC resistance, leads me to the assumption that the quality factor Q of those coils must be quite low (the Q generally rises with coil-former diameter).

If my mechanical abilities would be higher, I probably would have tried the universal/progressive approach which requires the complicated winding machine. Nevertheless I build the so called ‘Tiny Coil Winder’ which I use to speed up the winding process of the 3D printed sectional coils remarkably.

I own several vintage RF inter-frequency filters (e.g. 470kHz) of AM radios (due to my collection), which are based of the universal/progressive coils with adjustable ferrite cores. The so called (AM) ferrite antennas make also use of it.
I visit worldradiohistory.com for years now; they provide lots of interesting literature (which I appreciate) – I wonder if there are no problems concerning copyright infringement.

I've published a fairly detailed article for this coil winder describing the operation of the various sections and showing some different types of coils that I have made so far.  Since I have to host the images on my website anyway I chose to post the story over there instead of in pieces here.

But I did promise to post a short video, so here it is.  There's no narration so I would encourage you to check out the written article if you are interested.  If there is any Q&A please post here as I had to turn off my website comments (thanks to a party that shall remain nameless).

Roger, you are just all around amazing!  Fantastic movie, pix, and write-up as well!

Like you state, the more delicate the wire gets the more manual operation becomes essential.  Which for fancier coils pretty much means gears I suppose.

I imagine you agonized somewhat over the design of the wire feeder?  Interesting how it does double duty mashing the windings down.  Could you get it to dispense a tiny amount of dope too?

I haven't been able to think of any way of applying a binder that would not still be sticky on the next revolution and consequently gum up the works.   You almost need a heat-sensitive coating with a heater right at the guide exit.  The wire passes through the guide, gets sticky at the heater, bonds to the layer underneath and freezes before the core revolution is complete. 

I'm thinking of some type of comb arrangement on the midstock that would keep the walls of the coils tidy.  But it would have all sorts of problems that I haven't worked out yet.

I also thought about making a mold where you insert a core and then cast wax "washers" onto the core.  They would provide the walls to contain the helical windings and you could just hit it with a heat gun to melt the washers out and pot the windings at the same time (I think they would hold together for this short duration). But unless you are making repeats of the same coil the mold making process gets tedious.

But maybe the simplest thing yet is to just give up and 3D print bobbins for coils when necessary.  I wound a .5" diameter by about 1"  four-section coil that's a little larger than the obsolete 10mH Hammond/Bournes/Miller but it compares almost identically for SRF and has a higher unloaded Q (I used a couple gauges heavier wire).  I guess that's something in the forward direction...

I just received a bunch of different types of Litz wire for when I think I have something worth using it on.  At the very least it should be much easier to wind in self-supporting structures because of the nylon jacket.

"But maybe the simplest thing yet is to just give up and 3D print bobbins for coils when necessary.  I wound a .5" diameter by about 1"  four-section coil that's a little larger than the obsolete 10mH Hammond/Bournes/Miller but it compares almost identically for SRF and has a higher unloaded Q (I used a couple gauges heavier wire)."  - pitts8rh

If you could use just a few walls and minor infill, then the rest would be air and probably not a huge SFR killer.  You could easily control the radial distance too.  What orientation would you use to get vertical walls?  Supports?

"I just received a bunch of different types of Litz wire for when I think I have something worth using it on."

Something I haven't experimented with at all.  It would be quite interesting to compare similar Litz and non-Litz coils for specs.

"If you could use just a few walls and minor infill, then the rest would be air and probably not a huge SFR killer.  You could easily control the radial distance too.  What orientation would you use to get vertical walls?  Supports?" - Dewster

I used 2 walls and 10% infill and on this scale the parts felt like solid plastic. One layer would work if you could get full coverage but that would be difficult.   

There is no single printing orientation that works.  If you print vertically you need supports between the section walls, and they get too snaggy for winding when broken out.  Print sideways and the layer lines create the same problem unless you do some prep sanding.  I ended up printing a two-wall cylinder (could be a single wall) vertically for the core and then an array of spacer-washers flat.  The washers were pressed onto the core and the layer lines on both create a nice ratcheting friction.  I added little 45 degree notches to each washer to give the wire at the top of a section winding to be able to return to the base of the next section with at least a little separation from the subsequent layers.

I've never tried printing PVA but I presume you could make the same sort of bobbin, wind the coils, and then just wash away the bobbin.

I think bobbins hold promise for small inductors on the caliber of the Etherwave pi-section coils, if for no other reason than to have a suitable replacement for the obsolete inductors.  None of these coils will compete with your large air-core coils for SRF and Q, but I'm not sure that you need or even want 300+ volts on the antennas?  Anyway, I'm just playing around at the moment trying to learn more about these darn lumped element inductors.  It may seem odd that it can be easier to create a simple and clean inductance at 4GHz than at 1MHz.