Coil Winder

[The Kid]

The more I think about this, it wouldn't matter if it was a rectangle, either. You'd just have a wider circle (arc travel). It wouldn't be an ellipse. Well, the shape around the rectangle would be, but since it rotates with the rectangle, it's a rectangle in an ellipse in a bigger circle! And to clarify above, when I said 11.3 ° minimum, that's from the top of the bobbin corner parallel with the deck. You're actually going from 0° to 13.3° total.
(Or wherever your wire stop starts. ) :grin: :sam:

 

[CJ]

I found a surplus shop for machine shop stuff, if you can dig it.
Been here for years, did not even know about it.

Picked up this cool stepper setup, perfect for a precision wire traverse. I was going to buy a lead screw, end bearings, bearing blocks, stepper motor, and then would have to machine together a track for the wire guide platform, build motor brackets, but all that hard and time consuming labor and parts sourcing has been done already, plus it's way cheaper.
75 clams OTD.

This guy has about a 100:1 screw, so I will have ten times the accuracy of a 10:1 screw, if the precision is half way decent, that is.

There is zero backlash in the table and the screw.

http://vacuumbrain.com/The_Lab/Winder/stepper_1.jpg

It even has proximity switches on the back, might be able to input that into the mix for back and forth winding.

http://vacuumbrain.com/The_Lab/Winder/stepper_2.jpg

Now all I need is a Gheko drive for the stepper, and an interface of some type and I'm done.

 

[The Kid]

Cool! That's a nice surprise, about the machine shop. I gotta get to work on windin' some coils myself, one of these days! :roll: Never enough time! :grin: :sam:

 

[CJ]

OK, got the Gecko drive for the stepper/traverse motor!
About $114, but it's state of the art for this stuff.

And, they even send you a free Gecko!

You set up the stepper current with an external resistor.
It automatically lowers curent by 30% so the motor stays cooler during hold mode.

Has fwd andf rev, disable, so I can spin the traverse motor without it binding.
Also has manual strp port for jogging the traverse with a momentary.
Everything I need and none of the garbage. Pretty small too.
Ordered a decoder plug for cabling the dc motor encoder.
It takes a werid DS 11 type 2mm spacing thing.
Digi key.

Getting closer, cant wait to hook this thing up.



http://vacuumbrain.com/The_Lab/Winder/gecko.jpg

 

[Gus]

Cool.

 

[gyraf]

Very cool indeed..

 

[mikka]

Now that is a beautiful thing...... my chops are dripping just looking! GO CJ! :thumb:

 

[Carson Day]

[quote author="CJ"]
Now all I need is a Gheko drive for the stepper, and an interface of some type and I'm done.[/quote]

I just bought one of these fun toys...

http://www.makingthings.com/makecontrollerkit/index.htm

It might be a bit overboard but i think you could use it to write a simple program that would count the number of windings and run the gecko.

cant wait to see it when its done :thumb:
Carson

 

[Marik]

[quote author="CJ"]OK, got the Gecko drive for the stepper/traverse motor!
About $114, but it's state of the art for this stuff...

...Getting closer, cant wait to hook this thing up.

[/quote]

Yours looks like G201. I was drooling over G210 for awhile to pair with my Frog controller system--cool thing. For rotary table makes any angle, or calculates and automates any number of teeth for gear cutting. On a lathe makes any thread, metric or imperial. Basically, can do any wire, as well--good wor winding. Just have other priorities, first :sad: .

 

[CJ]

I went down to the machine shop surplus place and bought three more 5 pound spools of wire, and picked up a piece of angle iron, drilled tapped, painted, ready to go.
So the traverse is mounted along with the Gecko.
I had to open it up to set the jumpers.
The big switching tramnsisotors are under the main board:

http://vacuumbrain.com/The_Lab/Winder/gecko_d.jpg

I had to move the motor to set up the clearance between the traverese and the motor.
Stuck the largest bobbin in the chuck to make sure I can clear it.
Then I just bolted everybody up.


http://vacuumbrain.com/The_Lab/Winder/gecko_a.jpg

http://vacuumbrain.com/The_Lab/Winder/gecko_b.jpg

http://vacuumbrain.com/The_Lab/Winder/gecko_c.jpg
The Gecko needs heat sinking for anything over 3 amps.
I am only drawing 1.4, but decided to sink it for max life time.

http://vacuumbrain.com/The_Lab/Winder/gecko_e.jpg

Getting close!
Maybe a simple 555 divider with a pot and I am done.
If no good, have to PIC it.

 

[Black Dog]

Very cool :thumb:

Lars

 

[mikka]

Very sweet.... leeetle steps now.... :thumb:

 

[CJ]

OK, I recieved the connectors for the encoder, wired it up, stuck 5 volts on it and it works!

So now all I need is something that changes multiplies or divides the pulses.

This encoder puts out 5000 pulses per rev, I cranked the motor wide open, and read the frequency, 214,000 hertz.

So 214,000/5000 = 40 revs per second times 60 seconds = about 2400 RPM max.

There are two channels available from the encoder, A and B.
Each channel has a plus or minus peak available.

Channel A is offset from channel B so that you can sense direction.

Here is a scope shot with the motor running.
The encoder is that round thing hanging off the back of the motor with the ribbon connected to it:

http://vacuumbrain.com/The_Lab/Winder/encoder_a.jpg

 

[12afael]

I have an old printer waiting to be transformed to a computer controled winder. the print head can control the position of the wire and paper feeder make the rotational work.

as always I have so many ideas and no time.

 

[gyraf]

What about a simple 4059 "divide by n counter" - http://www.datasheet4u.com/html/C/D/4/CD4059_TexasInstruments.pdf.html

BCD-code inputs, so it should be possible to set it with simple thumbwheel switches.

Jakob E.

 

[CJ]

Sucess! ....
.
.

.
Well, sort of.

This is my first automated winder layer.
It might not look that good to you guys, but for me, it was a happy momment.
The motor and the stepper are at least communicating.

First layer:

http://vacuumbrain.com/The_Lab/Winder/first_a.jpg

Added a little tension and adjust the stepper for #48 and I get a little closer!
http://vacuumbrain.com/The_Lab/Winder/first_b.jpg

I am wrestling with a dual 555 timer (556) and it does not look like it is going to work out.

It is fine for constant speeds, but it burps and coughs up stray pulses when I accelerate or slow down the winding motor.

And, it has limited bandwidth. I played and played with different RC values for the time constant, but no luck.
I think Radio Shack might be exagerating the freq response at 200 K CPS, as I only got about half of that.

So the stepper quits at high RPM.

I took the encoder pulse and shoved it into a monostable 555 circuit to generate a triggered pulse of more even amplitude and wave shape.
Then I went into the 555 wired as a divider. The freq out is controlled with a time constant pot.
I put a decade resistor box in there instaed of the pot and got some really good low speed resolution on the stepper.

This thing now has four power supplies

1) bobbin motor-variac-bridge-cap-motor
2) counter supply ( i wanted the counter on it's own supply so I could leave it plugged in all the time while shutting down the other systems.
A litle redundancy to keep from losing turns info.
3) + 5 for the control board
4) + 37 VDC for the stepper motor.

The stepper motor is rated for 3.4 volts, but the Gecko limits current to 1.4 amps, so the motor works better at the higher voltage.

Stepper motor voltage should be between 4 and 25 times the stepper motor voltage for best results.
Also, it likes an unregulated supply, just an xfmr, a bridge and a big cap. Thats it. I made this by "stripping" down a Power One supply.
Regulated suplies do not sink the crap coming back from the stepper very well.

Couple of pics.
Thats the controller perf board proto at the bottom.
I made a quick wire guide out of teflon.
Just drilled a 1/16 " hole it, wanted instant gratification, been working on this dang thing so long!
Visible on top of the traverse.
Will build wire guides next.

http://vacuumbrain.com/The_Lab/Winder/controller_a.jpg
http://vacuumbrain.com/The_Lab/Winder/controller_b.jpg

Here is a schematic showing my next plan.
First a x2 buffer to get a little more encoder pulse voltage at the high Revs.
Then, something that will generate equal sq wave halves, independent of frequency.
I noticed that the 555 chip pulse width varies with frequency.
Idf the cap value is too high cpompared to the resistor, the cap does not have time to discharge all the way, and you get missing pulses, which is not good for winding accuracy. Thus, the need for a better pulse generator, which can be a descrete multivibrator (total DIY) or a 74M series logic chip of some kind.

Then, now that I have my accurate sq wave that does not drop out at the high end, I can divide the pulses with a PIC chip, or try Jakob's idea with the N divider ( thanks Jake!)

One good thing about the flip flop is that it auto divides the frequency in half, at least according to the guy hear at work.
This means I do not need as much frequency response from the chips down the line.
I really don't need 5,000 steps per rev out of the encoder, 2,500 will be fine! :grin:

I figued out the accuracy with the 5,000 pulses.
It orks out to 0.00000312" inches per pulse. :thumb:
80:1 for the lead screw.

Here is my scratch pad sheet for the next rev.
I found out that the direction port on the Gecko drive does not need a fancy pulse to switch it as indicated in the instructions.
It's just needs an open collector of a transistor, and a switch and LED's to indicate direction. High or low. Thats it. No pulse required.
The circuit for that is at the bottom and is final.

I think a sig gen into a the base of a 4124 could drive the Gecko to just make sure everything on down the line is in order.
When you get missed steps, you do not know who the culprit is, therfore, running the driver, stepper motor and lead screw with the generator will rule out trouble form any of these components, thus, the aux drawing at the top right.

A endpoint sensing scheme using already existing infrared LED send and recieve is in place.
I need a jog button, so when I want to move the traverse without the motor winding, I can.
I will use the defunct 555 chip as a permanent dpeed pulse generator which will be switche into the gecko while switching out the encoder circuit.


http://vacuumbrain.com/The_Lab/Winder/controller_c.jpg

 

[mikka]

It winds!!!!!!!!!!!!!!:thumb:

 

[gyraf]

Nice! :thumb:

Now you need a second down-scaling divider, looking at the pulses for the stepper drive, to reverse the stepper direction after a set amount of steps, so it'll wind the next layer automatically when it reaches the end of the first.



Question: to control the winding pitch in a predictable manner, shouldn't the wire guide be very-close to the bobbin - maybe even touching this?

Jakob E.

 

[CJ]

There is a magic spot for the wire guide depending on the bobbin size.
So with a rectangular bobbin, placement is always a compromise.
If the guide is too close, then your max RPM will be limited since you have those accelerations we talked about. And to get an even layer, you sometimes really need to lay it down fast. Ollies says it takes him 3 hours to wind a choke, so maybe he has a close guide on a precision machine turning at low REV's. But your right, if you want maximum accuracy, the wire guide should probably be close.
I have never looked at a professional precision machine, the guy down the street has one, maybe I should drop in.

As far as the reverse thing, there are a couple of stops already built into the traveres. These are used so the stepper does not run all the way out on either side and damage the screw. So all I have to do is mount them on a slide so that they can be adjusted, and I will automatically have my margins set. A simple flip flop will work, it gets to one end, you keep the dir input positive, you pulse the flip flop again when you get to the other side, and the dir port goes negative. I kind of have a sketch of this at the top of that late night scratch pad I posted.

Been thinking about a BFO.
Remember when the radio guys bounced two frequencies togetrher to get a lower fundamental?
Build an osc and see what I can do maybe?

There are a mililn ways to do this, thats what the control industry is all about.
DC motors and servos, phase locked loops, all kinds of stuff.

The neat thing is those controllers are $1000 - $2000 bucks a piece, so I am really saving a bundle with the two dollar chips.

 

[Gus]

http://web1.automationdirect.com/adc/Shopping/Catalog/PLC_Hardware/DirectLogic_05/PLC_Units

look around the site at the PLC stuff
also check the AC drive stuff
AC drives are fun.