Winding small coils

[snoozy]

Hey folks,

I try to wind small coils with a 0.2 mm wire. Due to the required mechanical dimensions, the coil is suppost to have 50 windings with 5 layers. So, 10 windings per layer.

To do the job, I created something with my 3D printer, but it does'nt work so well. Mostly because everything is so tiny and you need to have tension on the wire.

I'm wondering if somebody has an idea, how to improve this thing or any other option to create a coil like that.

I'm working on a new version with a tensioner on the side. Actually it will be only two screws and a plate that should give some preasure on the wire. The wire will be bend around to the coil.

Thank you.

[ChristofferB]

I think the only machine up for the task is a human with a very steady hand.

Also with wire that thin (is it verowire??) You really need something to hold the coil once done. Something ugly like glue.

Cant see the exact diameter from the image but people has used thin drinking straws for small coils for ages, it might be worth trying! Good luck!

[snoozy]

Thanks for the help. You mean the drinking straws as the "core" to wire around, right?

Yes, the wire is verowire. I think it should be good for the task and I had it laying around.
Is there anything better, like more stiff with this diameter?

I will use some super glue After the windings are done. It doesnt need to be perfect, it just needs to work.

I will start printing the version two and then give it another try.

[Conrad Hoffman]

I do a lot of tiny coils. It helps if you can use a bobbin that remains part of the coil assembly. Without a bobbin you can coat an arbor with some kind of release agent, could be silicon, Teflon or even wax. The best wire to use is self bonding magnet wire. It's coated with an adhesive. When the coil is done you fuse the adhesive with alcohol or heat, creating a self supporting coil. IMO, the Verowire will probably drive you nuts. Even regular magnet wire would be better. I usually tension small wire by pinching it in a fold of cardboard or paper with my fingers. The hardest part for me is getting even layers, so I end up with what's called a scramble wound coil. Some coil winders have a mechanism to evenly lay the wires.

https://mwswire.com/magnet-wire/bondable-magnet-wire/ (minimum order will get you unless you can find on eBay or somewhere)
https://www.worthpoint.com/worthopedia/vintage-morris-moreco-coilmaster-hand-1826388508  (these used to be common and cheap for ham radio builders; posted just for ideas)
https://www.windingmac.com/winding-machine/bobbinless-winding-machine-fully-automatic-bobbinless-coil-winding-machine/ (so you can see how the big boys do it)

[Ian.M]

I'd use thin wall PTFE tube on a slightly tapered mandrel for the former, with end cheeks of PTFE washers so you can actually get the coil off the former after applying the superglue!   If you cant get tube the right diameter you'll have to settle for rod, again with a slight taper where the coil goes, turned down either end to fit the cheek washers.  As PTFE has such a low coefficient of friction, you'll need to cross-pin the drive end of the tube on its mandrel or the rod to a close-fitting hollow coupler on the drive shaft.

I'd seriously consider building an automated winder, with a small encoder equipped gearmotor direct driving the mandrel, and running the wire through a hole in the horn of a RC servo as a guide so the controller can lay the wire down neatly across the full width of the former.  To get  consistent tension, run the wire between two felt washers on a threaded stud, with a metal washer bearing on the outer one, compressed by a spring and a wing-nut.  Notch and pin the side of the washers opposite to the wire so they don't rotate, kick out the wire and loosen the wing-nut.   The controller would just be an Arduino driving a MOSFET for the gearmotor, counting encoder pulses on a timer input and adjusting the servo position accordingly.  Control could be as simple as define the coil parameters in the sketch so it winds ONE  coil every time you press a 'go' button, or you could go the whole hawg and use a LCD +buttons shield and implement a fully featured UI that lets you specify coil dimensions, no. of turns and wire dia. and works out the guide advance and number of layers for you.

[mag_therm]

In 2018 I made some DC choke windings ( 100' of turns) on the lathe.
To maintain tension, the field current on the toolpost spindle is set, and the motor is pulled in reverse.
To maintain position, a "press fit" hole was drilled in an epoxy block which is carried on the saddle on an aluminum bar as shown.
A gear ratio was set on the longitudinal feed approximately equal to the wire diameter.

Perhaps Snoopy could miniaturize those items while maintaining the hand crank.
I think an ordinary small permanent magnet dc motor could be pulled in reverse when fed via a variable dc supply.

[SiliconWizard]

Hmm yeah, although a bit annoying, I think I would go purely for hand-winding here. 50 turns is not that bad.
Use some adhesive on the coil's core (if they are air-core only, you can use small tubes of paper or plastic), wind the first layer, then you can put again some adhesive on the first layer, and so on.

I remember having designed a semi-automated winder years ago (must still be in a box somewhere), with a PIC MCU, electric motor with reduction gear, optical detection. Back in the days I was making custom electric guitar pickups, so that was hundreds to thousands of turns, wouldn't have been doable without some kind of winder like this. But for 50 turns... unless you had many, many coils to build, I'd say nope.

[Ian.M]

OTOH the winder on its own is a nice project that is likely to see significant future use.  Once you've got one, it makes custom solenoids, inductors and pot-core transformers a go-to option any time you don't fancy trolling through magnetics catalogs for a part that may not even exist, then having to find a supplier with stock.

[snoozy]

Thanks again for your input.

My first trys are quite promising. Still needs a bit of improvement. Especially for the dimensions of the core.

Do you have any knowledge about how much influence a not so well wired coil has on the inductance?

[cheeseit]

The new setup looks good.

I repaired some tweeter diaphragms by winding new coils using 0.1mm enameled wire. The setup was a jig where a piece of copper around a bit holder was held in a fly tying vice. The wire was wound on a bobbin and held in a bobbin holder which ensured consistent tension. The coil was then wound by rotating the bit holder, which was angled a bit so the wire naturally followed the previous winding tightly.

The most tricky bit was crossing back on top of the first layer while keeping the previous layer and new windings real tight. They had to be epoxied so was wound on PTFE tape. I found high temperature, low viscosity epoxy easier to use than CA glue and used hot air to semi-cure it between layers. After a bit of fine tuning this produced consistently good results in a few minutes.

" alt="" class="bbc_img" />

[Ian.M]

Spriz the coil lightly with dilute sodium bicarbonate solution as an activator and dry with warm air, then when you apply the superglue, it will cure near instantly on contact. 

OT for coil winding: You can also use a little fine sodium bicarbonate powder mixed into colloidal silica as a superglue filler - put the mix where you want to fill up to a mm or two thick, drip thin superglue on it, and repeat till you've got the required thickness. Add cellulose microfibres to the mix and it can be reasonably strong.

[Distinctly Average]

Many years ago I had to wind coils regularly for a little CB modification we were doing for people. We used a winder meant for winding sewing machine bobbins. It only cost about £15 at the time and only small modifications were needed for it to do what we wanted. Might be worth having a look at something like that for inspiration.

[ttx450]

Hi all...

I have similar project.  started over 1yr ago, CV 317P hardrocker crossover coil burned up.  ignore left pict..  I need to rewind a coil that burned up many yrs ago.    I will try and match up henry as close as possible.,  got the wire, ferric core.. do not have original thimble.. will make a good quality paper tube if necessary.
any help or suggestions appreciated?

[ttx450]

how important is the gap between the ferric core to windings?  I would thk significant-few thousands ? - maybe not?

[snoozy]

My coil should have around 19 µH due to a calculation from https://coil32.net/online-calculators/multilayer-coil-calculator.html and the results I get from FEMM.

I'm wondering if there are any experts on FEMM here. Why I don't get a correct result, if I draw each wire itself instead of using a rectangular box with 50 windings?

(Any suggestions for another simulation tool?)

[jonpaul]

Check the standard single and multilayer solenoid formulas and calcs.

YES gap is critical, we use 10,000 u ferrie with mirror lapped mating surfaces.


A gap of 0.01 mm is noticable.

We use Adams Maxwell winders and easily wind #36-#39 AWG on 0.085 WL bobbins.

You need to check the available bobbins for your applications.

Kind Regards,

Jon

[mag_therm]

Hi snoozy,
There are various empirical/ analytical functions that should give accurate results for a coil like that.

If you want to use FEM:

I would suggest to use axisymmetrical, where a cross section is visible above the x axis ( X is at radius = 0)
Y becomes Radius R.

The axis should not be assigned any properties in axisym case.

Then the air needs to be bounded by an edge which is assigned a Dirichlet boundary.
The air edge should be at least 3 coil diameters away from the edges of your coil ( for accuracy).
Set it to 0 wb.

The circular copper turns (blocks) should be assigned conductivity about 45e6 S/m, Ur =1
In my app, you can set a current density [A/m^2]  to a block, or else map the block to a circuit and use
currents or voltages.
Your system may be different.

The air conductivity = 0 , Ur=1

Your mesh spacing looks good

Because the model is a full axisym, the post processor should give actual ( total) values of the coil

I can't find in archive any multilayer solelnoids I have done, but here is an axisym model of two single layers, one with a core.

[ttx450]

Check the standard single and multilayer solenoid formulas and calcs.

YES gap is critical, we use 10,000 u ferrie with mirror lapped mating surfaces.


A gap of 0.01 mm is noticable.

We use Adams Maxwell winders and easily wind #36-#39 AWG on 0.085 WL bobbins.

You need to check the available bobbins for your applications.

Kind Regards,

Jon

thks..  thats some small wire.    no lapping here... lol..  most likely hand made bobbin over the core I purchased.  interesting to know and learn..

[pwlps]

I routinely wind small coils (for NMR detection), the smallest dimensions are about 0.4-0.5mm diameter with a 0.05-0.1mm wire, 20-50 turns single layer. It is easy to wind them by hand on a drillbit using a binocular, a winder is not necessary but the tensioner has to be stable and smooth.

[ttx450]

I will need binoculars..

after you make the first run across, do you cross all the way back and start from the beginning, or do you just start back winding from where you left off.  does it matter?  thks

[snoozy]

Thank you for your input!

I tried out Quickfield in the simulation, but the restrictions of the studends version make it unusable. Sadly, the licences are quite expencive.

I keep trying with my handmade coils, but I will get the option to use a winding machine in the university. It is ment for transformers, so probably it can't handle such small wires. We will see...

The most difficulty I have is to come back with the wire for the next layer. Sometimes it stacks up. I might try with a magnifier, too...

[jonpaul]

Bonjour PWlps, je pense que le traduction une peux mauvaise:

Binocular (FR) >>>Binocular microscope (EN)?

Amenities,


Jon

[ttx450]

Bonjour PWlps, je pense que le traduction une peux mauvaise:

Binocular (FR) >>>Binocular microscope (EN)?

Amenities,


Jon

  no kidding.. i need to binoc to just walk now days

[ttx450]

still on this.. 

I dont have any fancy lcr meter chinese one, hope to be getting a Der dee 5000 soon to read inductor before I start, dont totally trust what I have.  Like to finish the crossovers soon.  thks