Coil winders Ideas

[R.G.]

I hadn't been by in a long time, but yesterday, I did some catchup reading here.

It struck me that a good winding spindle would be a medium-frame stepper motor with 200 or more steps per rotation. The motor itself should produce a good mechanical spindle to mount the necessary winding head on. If this is done well, the motor itself is the whole mechanics of the rotating spindle for single-sided winders. Constructing a tailstock would take a bit more effort.

If you needed gearing up/down, or wanted to introduce some damping to keep step-jitter out of the wind, you could use a DC motor with a double shaft and put a timing belt gear on the back shaft, use the front shaft for mounting the winding head, and then run the stepper on the timing belt. That should dampen steps a lot. The DC motor is only a mechanical spindle, it's not powered; just a convenient way to mount the spindle.

I would put a toothed disk on the back shaft as a counter and for CNC a position encoder.

Driving the traverse is not an issue for hand-guiders, of course. But if you want a machine controlled traverse, you could gen one up from a second stepper driving a 1/4-20 leadscrew inside a guide tube. The guide tube is slit and the actual traverse carrier is mounted to the side of the leadscrew nut and protrudes out of the guide tube. A square-section guide tube works, as does a delrin nut. The guide tube can be embellished with optical interruptors for limit switches.

These two assemblies can then be independently bolted down to a mounting plate. Being steppers, they are already ripe for driving them with a $100 used laptop computer running some of the cheap CNC ware to drive steppers.

I liked this approach because it did away with all of the machining to make a winder. The motors themselves provide the spindle support for the winding head and the traverse. Making the guide tube and traverse is the most complicated thing and this can be done with ordinary garage tools.

It's also possible to drive the stepper motors from a 555 oscillator and hard-logic stepper encoders so you could put a resistor in a footpedal for spindle speed and use a joystick for traverse position.

This seems like a simple way to get a winder up and running, and have the mechanism ready to go to computer control.

I had some ideas about controlling the rotational speed of the spindle to speed up and slow down the spindle as a first order approximation to keeping constant wire speed - and hence tension - so that the necessary compliance on a tensioner was smaller. That needs more work.

[Ampclutz]

R. G.,

That sounds like my winder... The traverse doesn't take into consideration that stepper motors tend to miss steps and they can accumulate. Also, most controllers need to stop before changing directions so the spindle would start and stop. My fix was to make a heart shaped cam for the traverse control. That makes the traverse motor only have to turn in one direction and the relative speed can be varied for different spacings. Of course you can still wind anything you want by reversing the direction of the traverse motor. It can't get lost enough to wind the wire off the bobbin, an advantage...

AC

[R.G.]

After I posted here, I did some more searching and certainly the idea isn't new. I found essentially the same thing at cnczone as well.

I was primarily enraptured with using the motors themselves as the main mechanical support for the rotating spindle, saving having to do the machining to make a close-tolerance spindle.

On the traverse, yes, stopping on a time is a really important thing. Since there is so little force involved, you probably want the highest power-to-inertia ratio motor possible there. There will definitely be some math involved in figuring out the highest speed for which a given motor can successfully reverse from step to step. If your fastest traverse is under the stop-on-a-dime speed, then you can simply use the stepper as is.

I didn't go into the other things that would be needed like limits and such.

This makes me wonder if a brushless DC motor set up to run as a servo is a better way to do the traverse. Hmmm... I think I remember something about linear steppers. ACK! more research to do!

[Peter Naglitsch]

I used to work for a company selling, among other things, brushless DC motors. That particular brand was nothing else than a stepper motor with integrated controll elektronics. The nice thing with this motor was that it had an integrated analog 0-10V DC input to controll the speed. I still have two of those and have been contenplating the use of those if I ever get around to do a new winder. Their newer version alo have a integrated logical controller (PLC) so that you can controll ramp up and ramp down times, mowement length and such. Check out the Crouzet motormate http://www.crouzet.com/filiales/moto...e/en/index.htm

[R.G.]

Wow, neat stuff. The only thing I can think of as a problem with those is the likely price. I'm guessing that they get a fair return on a fantastic device.

After my last post I went haring off on the net looking for linear actuators which had quick reversing ability. I found that there do exist linear stepper motors. These are honest-to-god stepper motors that step in a straight line. They are very good - but pricey.

I found voice coil actuators available as a separate assembly. Pricey.

What I was after was a linear traverse solution that did some of what the coil turning motor did - form all of the precision mechanics without a lot of precision work to make it into a winder.

So far the leading contender for easy and cheap is to strip apart an old printer. The price is right - they can easily be found free - and you get a precision linear slide you can put a motor on. The motor that's on it is likely to be too slow. What might be ideal is putting a brushless DC motor on the (usually) belt driven printer mechanism. A lot of them will even have encoders.

One of the reasons I was after direct mounting of the coil to one stepper and running the traverse from another is that you can go directly to CNC. Used laptops go for $100-400 here, and you could easily turn an old laptop into a captive controller for your winder. Just dedicate the laptop to the winder entirely.

[R.G.]

This one is almost what I was looking for:

http://www.alltronics.com/cgi-bin/it...tepper%2DMotor

It's a stepper with the rotary-to-linear converter inside the motor shaft. It pushes the shaft in and out of the motor by 0.001" per step, and 450 steps per revolution, 1.875" travel.

You'd still have to rig a side-to-side wire guiding mechanism, but adding this stepper lets you put the traverse fully into software.

Not bad for $14.95.

[Peter Naglitsch]

I still use my first winder, but that has soon to be replaced. I have toyed with the idea of using a printer for the traverse movement. I was thinking I could simply just dismantle it, tamper with all the switches that sense the paper and the paper feed and use it as is. Then I would send a document to the printer and that document would hold the right info to create the traverse movement.

[David Schwab]

I'm working on a traverse for my winder, but I'm going with a simple cam operated device. I figure that's the motion I use with my hand. I can change the rate to get more or less scatter.

A CNC device would be cool, and that's the next phase, but at the moment I couldn't see me having a lot of use for all its possibilities.

My goals are a tensioning device, a traverse, and a counter that stops the winder at a preset number. In other words I want to load it up and run it.

[R.G.]

I'm working on a traverse for my winder, but I'm going with a simple cam operated device. I figure that's the motion I use with my hand. I can change the rate to get more or less scatter.

A CNC device would be cool, and that's the next phase, but at the moment I couldn't see me having a lot of use for all its possibilities.

My goals are a tensioning device, a traverse, and a counter that stops the winder at a preset number. In other words I want to load it up and run it.

Given that, a stepper motor tied directly to the bobbin holder might be just what you want. "CNC" does not have to mean "very, very precise and complex". Computers can also act as stone axes, or slightly polished stone axes.

Imagine: a 23- or 34-size stepper motor with the winder stuck on one end. A magnet glued to the shaft on the other end. Motor phases driven from a set of four darlington drivers and a reed switch set so the magnet clicks it. Your computer's parallel port can drive the motor at rates from dead stop to the max step rate for the motor, which is usually a couple of hundred RPM. The computer also reads the reed switch (could be a Hall effect sensor as well, or an optical interruptor, or an encoder wheel if you're REALLY fancy) and not only counts turns, but also ramps up and down speed as well as slows to a stop at the proper count. I believe that's all stuff you can do directly with BASIC from a computer.

Such steppers cost $5 to $15 on the surplus market. It doesn't have to be a fancy program.

If you had a second stepper tied to a linear traverse, you could also set it to "idiot" traversing, nothing but back and forth at a fixed rate like you'd get with an uncoupled cam action. Next step up is to look at the turns count and time the traverses to be X turns long. Again, done in Basic. The hard part of that is getting a good linear traverse mechanism, not writing or using a fancy CNC mechanism. There's enough bits on a parallel port to run two steppers at the same time.

Once you get the computer controlled mechanism, you can make the control as smart - or dumb - as you like. It just simplifies the mechanics a lot to have all of the machining stuff encased in stuff you can buy off the shelf.

[R.G.]

I think I have a lead on a simple(r) linear mechanism.

Small Parts Inc sells nylon pulleys intended for use with steel cable for about $3.00 each. They have pressed in ball bearings. Nothing says that you have to use them only for cable. In fact, any plate of material with a half-round on the edge will work fine.

Imagine that you have two strips of steel, 1/8" thick, that each have a straight edge which has a half-round edge section. Place them parallel , rounded edges facing each other. Two pulleys fit onto one edge. A third pulley fits the second edge. The pulleys are bolted to a plate which has pins on it that let the bearings turn. The third pulley is spring loaded to press against the second edge. The two pulleys guide the plate along the first edge. The third pulley forces them against the first edge. The spring compliance makes perfect parallelism unnecessary. You mount your traverse on the moving plate, mount the two edges on a base plate. Now run a small timing belt past the moving plate clamped on at one spot. The timing belt runs to a stepper motor with a timing belt pulley and to a second idler pulley on the other side of the moving plate. Stepper motor now moves the plate side to side a few thousandths of an inch per step. If you mount an opto interruptor for a home switch on one end of travel, whatever is controlling this thing can sense home and know position by counting steps.

You could use a leadscrew instead of a belt, but a belt is faster moving. Leadscrews have a little too much gear-down in this kind of app I think.

It's not exactly free of having to do machining, but it does sidestep the high cost of linear slides and other linear mechanisms, and is something that could be put together with hand tools in a garage.

[Ampclutz]

For traverse: Heart shaped cam on stepper motor pushing a spring loaded arm with an adjustable center pivot to set the cross feed travel and an adjustable wire feed position. Tensioner on arm.

For spindle: Direct drive off the end of the stepper motor.

Butt simple... It does everything you have asked for yet! Plus a bunch of features you haven't asked for yet. What's not to love? The one I built cost me about $5 for hardware not counting the motors and control or the scrap wood. Shareware TurboCNC for the software... Other software works too like EMC.

I doubt you can come up with a winding scheme I can't program this thing to do. It counts to one 400th of a revolution. Want 8672.45 winds? No problem! Pick any scatter. No problem. Want to change the scatter several times during a wind? No problem. Want the coil a little fatter in the middle? It's doable but not as easy. The machine will just run slower...

It can run on a 486 PC. Read free or nearly free computer.

Load the bobbin, Press a few keys to start the machine, Come back when the coil is wound and unload it.

That my story..

AC

[David Schwab]

No, this is butt simple. You guys are making it too complicated!

Very good ideas though... and I will do that eventually... right now I want a traverse to use with my current winder.

If I'm going to use a computer, I want it to control the entire winder. But for a semi manual winder, I just want something to replace my hand, so I don't have to sit and feed wire.

So something like this would work. I'd use a geared motor, with speed control. I have to come up with a tensioning device, and a counter I can program with a preset to stop at a certain turn count.

Here's a very rough concept (I just drew this now... it's not well thought out).

[Ampclutz]

David,

That is very simple! In fact it is almost identical to mine. Very close to the same parts count too. But the way it is drawn it will pile the windings up on the edges of the bobbin and starve the center. A heart shaped cam will solve that piling the windings on the edges thing and eliminate the connecting rod between the motor cam and the wire guide pivot arm. A spring is needed to keep the roller bearing (roller skate wheel bearing) against the cam. I adjust the center pivot forward and back in slots to adjust the proper throw for whatever bobbin I'm winding. With a cam and roller bearing has to stay pretty much in the same place relative to the cam. Mine could use a motor that is not controlled by a computer like your system does. Our two systems are more the same than different.

Cams are no big deal to calculate and draw BTW... And larger diameter ones are easier than little ones.

How do you plan to solve the piling the windings on the edge of the bobbin thing?

AC

[R.G.]

Ampclutz is right - the velocity of the wire guide from side to side is a sine function. That is, the side-to-side wire speed is fastest in the middle and slowest on the ends. The wire will pile up more on the outer edges because the traverse dwells longer there.

As to the heart shaped cam -
I think that's an attempt to get constant velocity per unit rotation on the traverse stepper motor, right?

The rack/pulley traverse was an attempt to get linear motion per unit step. If you're not going to have that, David's crank mechanism is simpler, but you do have to program the step rate of the stepper running the crank. By doing that, you can re-linearize the traverse rate to be uniform and not get edge-of-bobbin pileups. I was after linear from the start.

The only issue with steppers driving a traverse is that you have to have a step rate low enough that the motor torque can cause a step-to-step reverse in direction. That's probably feasible for a direct stepper driven traverse because the motor rpm on the traverse driver is slow enough. That's why I didn't like a leadscrew for the traverse. The leadscrew gears the motor down so much that the motor's own inertia would likely prevent an instant reverse.

[David Schwab]

Our two systems are more the same than different.

That's interesting! It just made sense. I was sitting thinking about a traverse and that idea popped in my head. I mentioned it to my friend that I build guitars with, and he's an industrial designer.. so he drew out this elaborate thing with a sliding gizmo and stuff.

How do you plan to solve the piling the windings on the edge of the bobbin thing?

I hadn't thought about it yet, but I was aware of the problem. I've seen winders with heart shaped cams and stuff, so I figured something would need to be done. I think once I sat down and actually started to plan the thing out, I'd get a better idea what needed to be done. I can conceptualize in my head, but I need to have parts in my hand to see them move.

This would be a god job for a computer simulation!

I'd love to see your winder. I figured if I came up with something that worked, I'd just go ahead and share the plans. I see people trying to find winder plans, and then trying to find oscillating fan motors and stuff. So the thirst for knowledge is out there.

That's what's so great about web sites like this, and R.G.'s too.

I'm using a Schatten winder now, and it works fine, but I'm planning on selling it and making my own. This was a good solution at the time, when I needed a winder right away, and didn't have a work shop. Now, a year later, I have a work shop with a milling machine.

I just want to load the winder up and press a button. Go have something to drink and come back! As I said, eventually I want a computerized winder. (and a dulicarver, and a pony... no, I don't want a pony...)