helical coil on toroid stack

[Anonymous]

I am trying to make a 4-turn helical coil laying on a stack of toroids. I made a crude diagram to try and make it clearer. The wire I'm using is a flat robbon about 0.2 x 2.0 mm cross section. Anyone have any ideas on how to go about this?

Thanks, Greg

[Anonymous]

There is a youtube model on how to make a shoe of all things, and he used tspline "pipe" to make the shoelaces, and adjusted them to the shape he wanted so they looked like real shoelaces.

So I guess if you can't use Coil, you can use Create > Pipe.

I think it works by following a spline or line you draw first, and then setting it to use this line for the path. So you can draw the path with a spline, etc, and then use Create > Pipe to make it look like a wire, etc.

You can make adjustments to the tspline pipe afterwards, as well, so your coil around 4 toroids looks official. If you watch the video, you can see how he eliminates nodes and tugs up in the middle, to make the "shoelaces" criss cross over one another properly.

See the shoelace part, 3/4 through this Youtube video:

Modeling a Shoe in Fusion 360

:

[lnonnato]

Hello coleagues,

At first sight, this is not a complex problem. As valuecnc mentions, just draw a spline and turn it to a pipe. Unluckily, that is absolutely not the case.

Geometrical  modelling of a real-world coil such as the one we want is not easy. I would try to model it mathematically, using a program such as MATLAB and then render it and import to Fusion. Of course, you can get a very rough piece by drawing it by hand, but I don't think this would be very far from reality.

Then, there is a even worse problem. We want to use a flat ribbon and, when you wrap it around the torus, we won't have it flat to the torus surface (see below) - try to cut a thin strip of paper and wrap it diagonally around a donut ... you will see the problem. We would have to deform the ribbon, what adds a second layer of problems.

It's a challenging problem. I will think about it.

Luiz

[lnonnato]

Hello,

Due to bad editing, my previous message states exactly the oposite to my idea. I think drawing the curve by hand WOULD result in a far from reality component.

Sorry.

Luiz

[Anonymous]

Hi Luiz-

Thanks for all the effort you put in. I see what you mean about the helix not 'twisting' with the toroid. I did try the strip of paper experiment, and it wraps just fine, so I think that Fusion is just trying to make a single closed loop instead of continuing with a helix. Maybe there a way to make the 'pipe' twist somehow around the torus curve.

[Anonymous]

The other thing would be to see if you can do the heavy math outside Fusion, then use the resulting parameters to drive a parametric model. That's assuming you can define the shape somehow in the first place. Maybe do this in coil pieces that stack up with a vertical pattern, combine solids, then add two end pieces. I've see threads made that are done starting with one turn and then the turns are stacked vertically. But your shape seems to have more to it, yes. So, this is getting way over my head in how to do the geometry, but I would love to see if someone can pull it off.

The topic of toroidal transformers is drifting towards my core interest here, audio, specifically electrostatic loudspeakers and use of toroids to step up the audio voltage for them.

I have a pile of transformer design books on my computer, heh they are perhaps the most math-heavy thing in analog electronics. I use toroids for driving an electrostatic loudspeaker I designed. Stacking cheap power toroids hooked up in reverse works nicely for that, like 6 volt to 120 volt, 15 to 50 Volt-Amp power capability, single primary and split center tap secondary.

Below is a list of inexpensive toroid transformers. I have played with the little ones, (6 to 9 volt secondary ones as listed) for stepping up audio amp voltage for driving electrostatic loudspeakers. They make very clean and loud sound; they are stunning. A top dog in diyaudio.com planar & exotics forum says to stack small toroids like you are talking about, rather than use two or four larger ones. Has to do with power and with low frequency performance is improved, without cutting the resonant frequency down into the audio highs like much larger toroids would do.

http://www.antekinc.com/grids/

I wish you luck. I think the bad boys of 3d cad need to step in on this one. And they will ask you, are you doing this for looks or for physical realism, I take it more for physical realism. What is the reason for the realism of the coil, to make sure it fits on something, etc, or just for the sake of purity of design? What use factors determine when it fits the accuracy needed?

[donsmac · Accepted solution]

Here's an idea how to go about it (screencast). It's not 100% accurate but adjusting the t-spline should get it pretty close. 

The result:

[Anonymous]

Well its very nice looking, that's for sure, but I think they want to drive the model with parameters somehow, no?

I love how that render looks, nice job. If we can get to a model that takes parameters, that would be even more awesome.

Perhaps we are waiting for future features like defining the path and it's orientation in some mathematical way, perhaps using the core shape as a reference or guide geometry?

Or does this kind of thing exist now in some other form overlooked? I noticed that pattern along path takes an argument for orientation, but its more for an array of copies than a continuous coil shape.

Or, independent of the core, using calculations to drive a spiral path with copies that overlap and join prior to using as a path?

I guess the calculations aren't really the issue in attempting this, but applying them to the path and to the orientation of the flat side.

I will be watching this thread to see what folks come up with, I find it to be a very interesting challenge.

Hey, why not have the API folks take a look at this, maybe the coil shape could be done there in a add-in kind of thing for transformers.

One could find the math in a math app, and port it into a plugin code. A coil maker would be a great exercise and demo of the API's capabilities.

Let the user select the shape of the wire, the turns, and a "bobbin" body they draw, and let the API wind the wire around it, hehe.

If someone wants to show off their mad API coding skills, here's your chance...

[Anonymous]

Still thinking about this.

This model is made more difficult to do in some automated way, due to a square coil form, and square wire.

So you can't use the Coil tool, yah.

Maybe you can project the shape of the toroidal core in 2d, perhaps reusing the original 2d parameters to drive the 3d combined curve?

Maybe do one sketch in each plane through the center of the toroidal core, and then use that geometry or variables to define a 3d curve consisting of one turn, which is later rotation-patterned around the axis of the toroid and combined? I have seen folks combine curves in two planes into a 3d combined curve routinely. I can't remember if Fusion does it. But can we use parameters or some constraint trick to drive that final 3d wire path line or make it follow and bend via constraints to the side-view 2d sketches?

OK, I feel in over my head now.

[Anonymous]

Look at the thread below to see more ideas on modeling twisted shapes like your wire, specifically look at the video screencasts in the thread on page 2:

Note the seed points that are driving the circular pattern are defined with dimensions, so you could in theory drive them with parameters, no?

Maybe this is the closest solution yet to address your request on how to do the wire around the core, but with a seed turn of controlled dimensions using Parameters in the dimension. Now the only thing is perhaps how to do this with a squarish shaped path, as looked at from the side, rather than a circular pattern, so it wraps around multiple cores which form that shape of wire turn. Maybe Patten on Path could be used in place of a circular pattern. The seed turn would need to controlled in two planes by the dimensions prior to duplicating the turns, to properly launch into a pattern which overlaps and can be combined or lofted? I will play around some, but others around probably know a way already. The ultimate goal is to drive this thing with dimentions and then Parameters. Might be useful technique also for doing mechanical cams, I'd love to see that screencast.

having a problem with sweep and flex

[Anonymous]

Oh, in the example I just pointed to, the poster did the circular pattern in the sketch, and there is no "pattern on path" in sketch mode, is there?

Close but no cigar I suppose.

Maybe we need a tool which can use a formula or multiple sketches to describe a 3D spline, and a way to make a "pipe" follow it, the profile of which could be defined by another component or sketch (in this case the wire cross section, a square-shaped profile). That way the whole thing could be formula and dimensionally driven.

Doesn't Solidworks have a tool which allows a spline or curved line to be programmed with formulas or a CSV (Comma Seperated Values) text file? Something like that.

I realize now the original poster may simply want a one off, not necessarily programmed or parametric, but I am reaching in that direction because a semi-automated method seems to me like a more consistently reliable way to generate shapes like the coil above, esp when they need to fit in an assembly or under a protective cover and you want to make sure the coil dimentions is close.

Like I mentioned above, mechanical cams need a similiar kind of "weird shape on a circular path" treatment. They can't simply look pretty, the cam profile has to be machined very accurately. I think most cams have one less degree of freedom than this wire wrapping exercise, though.

I will Google mechanical cams in Fusion 360 next, to see what I can dig up.

[Anonymous]

I made a sketch for drawing a cam lobe from the parameters table. I did it thinking if I could project the sketch lines of two of these "cam lobes" to drive a third combined 3d curve, to drive a seed turn of wire, maybe that could get us down the road towards a coil that follows the weird square wrapping pattern.

But at least I found a basic way to do cams.

I drive the rectangular array of lines with the parametric table, (like a bar chart) and duplicated the lengths of the spokes of the wheel in the round cam via equals constraints. (I could not find a way to dimension the angled "spoke" lines, it kept referencing the horizontal or vertical, not the simple line length. Is there a toggle for that? Or is this a bug?

We need a programed way to do circular patterns, with a table for the length of each point from the center. I did it all manually to set this up and it was kind of a pain. But it is pretty cool to think of the stuff you can do with this kind of parametric driven displacement for a round object, ie mechanical cam.

Fusion 360 file attached below.