Design of Solenoid Former.

[dzarren]

Hello there! This is really more of a mechanical question.

I am making a cryogenic Heat switch. This is like an electrical switch, but for heat flow instead of current.
It involves a superconducting solenoid, and a superconducting junction to act as the switch.

My concern lies in the design of the former, the brass part you see here that the magnet wire winds around.

There are 4 layers of 100micron wire. There are 200 turns per layer. The wound section is 2cm long, which gives us a turn density of roughly 40,000 turns/meter.
1 Amp will be flowing through this superconducting wire when the coil is energized.


The wire comes in and immediately undergoes a sharp bend via a slot. Then, the wire is wound, and after the 4th layer, the wire is back at the side the winding started. it leaves the windings through a similar slot.
How can I design this such that there are no sharp bends in the wire that may cause it to kink, or damage the insulation of the wire? I have tried filing and scraping at the corners, but they still have quite a small radius of curvature.

If you have any design considerations, i would love to hear them!

thank you for reading.


you can see here that the wire is fed in, goes into a sharp bend, then gets wound.



for those interested.

[kony]

Do you have rotary table for your mill? What is the smallest endmill you can use with your spindle?
You can mill there (almost) arbitary escape radius (best case being 4th axis on CNC mill, second best case bit iffy clamping & aligment job on rotary table under conventional mill, worst case lot of filing).

How about teflon tubing over the ends in critical spot?

BTW, those flat headed screws are making me cringe a bit - they do not sell DIN912 or DIN916 in brass?

[CM800]

how about this:

[David Hess]

How about teflon tubing over the ends in critical spot?

In transformers where this is a problem, I usually see tough small diameter tubing (or sometimes a bit of the same tape used between windings) added to the wire along the length where it approaches the winding.  I have had good results using very small diameter polyolefin heat shrink tubing for this which can be heated during and after winding to relieve stress and form it into the proper shape which it retains upon cooling but I have no idea what would be suitable for a cryogenic application.

[ConKbot]

Keep on filing till you get a nice radius (use a radius gage if you want to shoot for something specific) filing work is necer fun, but patience pays off. If your you're feeling crafty, but some abrasive paste/compound on a string and use that on the corner to round it over and polish it to whatever finish you desire.

[dzarren]

Do you have rotary table for your mill? What is the smallest endmill you can use with your spindle?
You can mill there (almost) arbitary escape radius (best case being 4th axis on CNC mill, second best case bit iffy clamping & aligment job on rotary table under conventional mill, worst case lot of filing).

How about teflon tubing over the ends in critical spot?

BTW, those flat headed screws are making me cringe a bit - they do not sell DIN912 or DIN916 in brass?

I do have a rotary table I can use. I can hold down to a 1/16th endmill. I have no access to CNC.

As for the teflon tubing, I don't really have room for other stuff in the design, its also going to be subject to very low temperatures. The bend radius increasing features must be inherent in the geometry of the design.

Ive never seen the socket machine screws in brass, actually.
Although the flathead drive is not crucial, the button head aspect is, and that is what I had access too and i think it'll work great.
The lower set of screws is for clamping wire to between the screw and the flat of the solenoid, in order to thermally couple it to something.
The top set of screws is just for retention later. But a socket headed button head screw would be great

how about this:

PICTURE

I considered this, but this still has the problem of the wire going into a very sharp bend.

Keep on filing till you get a nice radius (use a radius gage if you want to shoot for something specific) filing work is necer fun, but patience pays off. If your you're feeling crafty, but some abrasive paste/compound on a string and use that on the corner to round it over and polish it to whatever finish you desire.

I filed for quite some time, maybe about 40 minutes, plus some scraping with a razor, (i cut myself by accident doing this), yet only to accomplish the minimal radiusing you see in the picture. There isnt enough room for me to get a decent stroke on the file, im scraping one or two teeth of the file, maximum, with each stroke. I will try your sting idea, if i cannot fix this by redesign.

[floobydust]

I thought a hard former might not work. Problems with the wire breaking due to contraction at cryogenic temps.

You wind it tight at room temp 25°C and then cool it to say ?200°C which is a delta of 225°C?

I do not know the thermal expansion coefficient if this is specialty superconducting wire.

200T/layer * 4 layers *14mm dia. * Pi = 35.2m of wire? Copper wire that long would shrink 135mm in length. You could do a per turn calc. Or try it with cheap copper wire and see.
I don't know how much you can stretch the stuff.

For the wire entry/exit points, consider the designs of some plastic bobbins for inductor/transformers. You could have pegs at the coil form end and terminate winding to that, then have the flying lead come off the peg. The peg has say 10 turns on it and is tiny. I can draw that if it seems reasonable. 

[ejeffrey]

This is for a nuclear demagnetization refrigerator?  You poor bastard

Honestly, I wouldn't worry about it. As long as the edge isn't actually sharp and you are careful when assembling you shouldn't damage the wire.  If you can't get a file in there try just a thin strip of sandpaper, or even something like a piece string loaded with abrasive material.  The goal isn't really to make a large radius.  You just need to break any burr/knife edge left from machining.  The Niobium or NbTi core wire is very strong, so the wire itself isn't going to break.  The varnish insulation is ductile enough it doesn't care about the bend radius -- the problem with tight bends is that it causes work hardening of metal and makes it brittle.  You just have to make sure that the edge isn't sharp enough to actually cut the insulation.  Also you want to anchor the wire in place in multiple locations so that it doesn't actually rub back and forth after winding.

Just wind the coil, anchor the wire to the support rod with kapton tape, and test it with a multimeter for faults.  Do a couple of liquid nitrogen dunk tests, and check again.  If it still works, it should be good for life. it if fails, check for burrs and hit any rough spots with sandpaper again then wind it again.

[jbb]

Sorry, I don't have 3D CAD SW, so I'll try to paint a picture with words. Or something...

Your initial effort used a trench, which would have a sharp edge.  How about you drill a hole perpendicular to the face near the corner of the trench and insert a smooth round pin (with appropriate fastener)?

[floobydust]

Um, I did another calculation.

I get roughly 35m of wire needed, about #38 AWG to get 76 for copper wire.
If OP is passing 1A through this, that's a lot of power! 76V for 76W.
The fusibility limit for #38AWG is 2.4A-3A.

[kony]

Why do you care about copper wire resistance when it is not used in the final design? Only time the power dissipation from wire would be worth considering is quench event.

[ejeffrey]

Why do you care about copper wire resistance when it is not used in the final design? Only time the power dissipation from wire would be worth considering is quench event.

Even in a quench event you don't care much about the resistance (assuming your power supply doesn't keep trying to push 1 amp through it).  All that matters is the stored energy 0.5 * LI^2 which will be dissipated during a quench.  As a practical matter, that isn't a problem for solenoids of this size.  A quench will ruin your day by warming up the fridge, but it isn't going to ruin your week by breaking anything. It isn't like a big 9 tesla magnet with 100 amps flowing, and it certainly isn't like the huge magnets that quenched at the LHC when it first came online causing massive damage.

[T3sl4co1l]

And that brass core will trap magnetic field something fierce, for example limiting the flyback voltage in a manner.

Tim