Problem with consistency

[kayakerca]

OS is Open Source - you know like SourceForge projects where you can download the code or a compiled version. I was hoping your software was something with a community and documentation and all that.

Not open source as such, even though it is a very fundamental application, I don't want someone taking the source and selling it. That said, anyone can have a complied version for free. It requires that I am given the serial numbers off the 2 Phidgets controllers so I can hard code them into the app then compile it. It also requires that the user gets the sample program up and running on their computer with the Phidgets controllers BEFORE I hand over the compile code as all the related software and dll's from Phidgets needs to be up and running before my app will run. If I just hand over the app and it doesn't run after someone has gone and bought the controllers, motors, etc., they are going to be pretty upset with me when in fact, they just haven't installed the based software/dll's properly.

I know it is a bad thing offering it up as no good deed goes unpunished. Ever.

Anyway, it was as much fun designing and building it as it is winding with it. Fun project if you decide it's the kind of fun project you want to get into. It really isn't a very sophisticated winder. Very much a vanilla version of what could be done by someone with some actual skills. It took about 2 months from not knowing anything about coding or what a stepper motor was, to the first wind (excluding the wait time to get the traverser motor from Haydon). How the physical winding hardware setup and layout takes some thought and could be done a bunch of different ways. It could also be uses as a straight up hand guided unit by adding a guide arm with a couple of collar stop spacers at the front of the unit.

Here's a Youtube video I did a while back of the first wind on the unit.

First CNC Wind - YouTube

Again, I'm just a retired hobbyist type guy. No special skills required, just desire and drive.

[Kindly Killer]

Very cool. That's generous to offer your nice software.

I think what I'm going to do is just add your direct drive stepper idea to my Arduino. I already have it set up to do traverse with a servo, so if I added your plate/motor assembly, that would get me an easy, rudimentary setup that would at least be consistent enough to vary one factor at a time as I poke pickup design with a stick.

[Kindly Killer]

Jim,

What holds humbucker bobbins in place with your faceplate? E.g. the screw bobbin - do you install the screws, or what?

Stupid me, I copied your design, but I copied what I remembered, and I put the groove on the wrong side LOL - the screw side, i.e. I've got the bobbin on my faceplate upside down - I kept clamping down tighter, which of course just mushed the forbon down on the mags... Next I put screws with short heads in the faceplate to keep the bobbin from turning (I made mine from wood scraps).

I came back and looked at your pic but I'm still wondering about how you do a humbucker bobbin.

BTW I did mine with Arduino, so I can input turns, bobbin height, and TPL on the computer - the semi-old laptop that is now my backup Ableton rig - but it can also run without a computer. The downside is that it is tantalizingly slow if you're any hurry to get the finished coil. It takes over 32 minutes to do 7800 turns! I think I'm gonna make this one the best I can and learn from it, maybe build another as cheap as possible with easy to find parts and document it for DIY guitar techs/players who want to make their own pickups (I'm amazed that I couldn't find a documented, shared design, given that so many people have done this exact thing).

Then if my fascination holds I'll move on to a proper CNC setup using the most obvious design and the most fungible parts. I am happy with the good time I had building this one, but I definitely got the bug to build a good one LOL. Speaking of computers and drivers, I'm still curious about your program - is it a Visual Basic interface for another piece of software - e.g. Mach3 - or is your program actually sending instructions to the motor controllers? Why did you do it that way when there are already programs to do all things CNC from the desktop? I think I will do a project like yours and I want to avoid pitfalls.

BTW has anyone ever tried combining a linear actuator (all in one like yours or a cheap one made from components) to move the stock ghetto tensioner - you know, two felt cylinders, springs, washers, wing nut? Seems like an obvious thing to try, so I wonder why nobody is doing it? Other than it's one more stepper driver you would need, and also repeatable motion != repeatable results with that kind of tensioner...

[kayakerca]

Jim,

What holds humbucker bobbins in place with your faceplate? E.g. the screw bobbin - do you install the screws, or what?

Stupid me, I copied your design, but I copied what I remembered, and I put the groove on the wrong side LOL - the screw side, i.e. I've got the bobbin on my faceplate upside down - I kept clamping down tighter, which of course just mushed the forbon down on the mags... Next I put screws with short heads in the faceplate to keep the bobbin from turning (I made mine from wood scraps).

I came back and looked at your pic but I'm still wondering about how you do a humbucker bobbin.

I only wind Stat style single coils. But when I get around to playing with other p/u's, I'll do up mounting plates like the ones from CNCDudez.

http://www.cncdudez.co.uk/b-platev3-0.jpg

BTW I did mine with Arduino, so I can input turns, bobbin height, and TPL on the computer - the semi-old laptop that is now my backup Ableton rig - but it can also run without a computer. The downside is that it is tantalizingly slow if you're any hurry to get the finished coil. It takes over 32 minutes to do 7800 turns! I think I'm gonna make this one the best I can and learn from it, maybe build another as cheap as possible with easy to find parts and document it for DIY guitar techs/players who want to make their own pickups (I'm amazed that I couldn't find a documented, shared design, given that so many people have done this exact thing).

Time doesn't really matter as a hobbyist, but I'd up the ante a bit if I were waiting 32 minutes for 7,800 turns. I wind a hottish bridge to 10,000 turns, so I guess that would be ~ 41 minutes on your winder. That's a bit of ouch! ;-)

Everybody does something a little different. And probably some who have designed their own and are pros may believe their design has some "special mojo sauce" secrets that they likely want to keep proprietary. Or, if they are like me, they are not real coders and their stuff would need a "Sit user cavete" sticker on it.

Then if my fascination holds I'll move on to a proper CNC setup using the most obvious design and the most fungible parts. I am happy with the good time I had building this one, but I definitely got the bug to build a good one LOL. Speaking of computers and drivers, I'm still curious about your program - is it a Visual Basic interface for another piece of software - e.g. Mach3 - or is your program actually sending instructions to the motor controllers? Why did you do it that way when there are already programs to do all things CNC from the desktop? I think I will do a project like yours and I want to avoid pitfalls.

The application I builT is based on the application Phidgets puts out so you can see how their controllers work. Mine is coded in Visual Basic .NET and it interfaces directly with the controllers and nothing else. I don't know what Mach3 is. What I did was not difficult by any definition. I believe any winder would be able to acquire enough skills to do an even better job than I did in very short order. They'd just have to have the desire and time and be willing to invest a couple hundred in the controllers. It's a journey. When I built my fist winder (mechanical Lollar style winder with a rheostat to control speed of the traverser), I had not even seen a CNC winder on the internet. Then I saw one and went. . . cool! Then I learned what I had to learn to design and build one. Again, it's not rocket science or I couldn't have done it. You've already done it.

BTW has anyone ever tried combining a linear actuator (all in one like yours or a cheap one made from components) to move the stock ghetto tensioner - you know, two felt cylinders, springs, washers, wing nut? Seems like an obvious thing to try, so I wonder why nobody is doing it? Other than it's one more stepper driver you would need, and also repeatable motion != repeatable results with that kind of tensioner...

If you went mechanical as you are asking, you'd need to have some way to measure the tension digitally so you could build the parameter into your application. WAY above my skill set so I just stick with the mechanical felt tensioner with a tension gauge to set up tension of a wind.

[David Schwab]

It takes over 32 minutes to do 7800 turns!

Yikes! I do about 5,000 turns in about 8 minutes.

I guess it's not so bad if you aren't sitting there guiding the wire all that time, but that's pretty slow.

[big_teee]

Yikes! I do about 5,000 turns in about 8 minutes.

I guess it's not so bad if you aren't sitting there guiding the wire all that time, but that's pretty slow.

David is about twice as fast as me.
I start and stop a lot, to look things over.
I can wind a hand guided strat coil in about 10-15 minutes.
T

[David Schwab]

David is about twice as fast as me.
I start and stop a lot, to look things over.
I can wind a hand guided strat coil in about 10-15 minutes.
T

I stop once or twice to see how the coil is loading up, but what I try and do is get on a roll where I wind one, and then while that's in the wax for 10 minutes, I wind another. So I get get a coil done every ten minutes, including the hookup wires.

I'm so looking forward to a CNC winder.

[Kindly Killer]

Yeah it is game-changing slow. I made 3 strat pickups, and it took from before lunch until the middle of the afternoon. I'd just check in, pull the old one off, put the next one on, click GO, flip the light off and leave the next one to wind. I wrapped and wired 'em, tested (they are within 8mH of each other, which is a first for me). Plus I kinda forgot about the project between each pickup - I'd just go upstairs and work on something else. I haven't had a chance to charge the mags and try the pickups yet, and I won't until morning. I am getting ready for a matinee show I have to leave for in a few minutes - still gotta work in between tinkering sessions!

It is definitely different from having an idea, sitting down and winding one, listening to it, adjusting, lather rinse repeat... It's more like a project, where I just write down the next step and come back to it when the winding is done. I would definitely like it better if I could combine the consistency of CNC with the speed of hand winding.

They all came out perfect BTW - big thanks to all for the help in this thread.

[Rick Turner]

How about syncing a strobe light to a winder so you can see what's going on? How fast can a strobe be triggered, anyway? Or you could watch through a synchronized spinning wheel with one or more holes.

As for me, well I just love my antique CoWeCo automatic winder.

[Joe Gwinn]

How about syncing a strobe light to a winder so you can see what's going on? How fast can a strobe be triggered, anyway?

No problem syncing to a winder, at up to 250,000 RPM. But units accepting external trigger are more expensive. Genrad (now IET Labs Home Page) is the standard, but there are lots of clones.

With LED units, beware of overly-long flash durations, which cause blur and may make it impossible to see the #42 wire flail about.

One can buy a manual-adjust (non-sync) strobe for about $300. 461830 - Digital StroboTach 115VAC, 60Hz

I ended up building my own xenon stroboscope unit that's synchronized to my homebrew winder, controlled by a simple analog computer that allows one to trigger the flash at a desired bobbin angle, regardless of RPM, by turning a knob. It isn't a hard thing to do, but an electronics background is necessary, plus a bunch of lab equipment, and the winder needs the right kind of encoder on the bobbin axle.

I used xenon because it was easy to get bright-enough flashes 5 or 10 microseconds long; this is hard with LEDs. The brighter, the slower.

Or you could watch through a synchronized spinning wheel with one or more holes.

This is the traditional approach for sure, but it may be hard to get the open period short enough to control blur, and yet have enough light.

[Rick Turner]

Joe, that's great; but I don't think I'll be winding at 250 K p M any time soon! So your knob basically controls the phase angle of the light blip.

Me, I just blink really fast...

But it would be fun to have a strobe on the winder. It would be funny to lash a computer to my ancient and totally gear and cam driven, nearly Steam Punk winder. You should see the original manual for the thing with all the tables of gear and cam setups for different wind patterns. Incredible... The last guy who had my machine used it to wind those amazing self supporting chokes...no bobbins...for RF use. They used a magnet wire with a texture on it so the wraps would hold onto the layer below. I'll try to find one and take a picture.

[Joe Gwinn]

Joe, that's great; but I don't think I'll be winding at 250 K p M any time soon!

But, but, ... think of the productivity increase!

So your knob basically controls the phase angle of the light blip.

Yes, with a resolution of 1/500 of a circle, which is slightly finer than one degree, fine enough that it seems stepless.

Stripped to the essentials, the analog computer integrates change pulses from the shaft encoder on the bobbin axle, and the integrator is reset when the index pulse comes by, yielding a voltage that represents the shaft angle. The knob controls the integrator voltage at which the xenon is fired. This yields a stable display, and variations in speed cause no visual artifacts. (Well, I do get some jumps due to DC motor interference getting into the circuit, but it hasn't yet annoyed me enough to get it fixed. Probably all that's required is some added metal shields isolating the motor from everything else.)

Me, I just blink really fast...

But it would be fun to have a strobe on the winder. It would be funny to lash a computer to my ancient and totally gear and cam driven, nearly Steam Punk winder. You should see the original manual for the thing with all the tables of gear and cam setups for different wind patterns. Incredible... The last guy who had my machine used it to wind those amazing self supporting chokes...no bobbins...for RF use. They used a magnet wire with a texture on it so the wraps would hold onto the layer below. I'll try to find one and take a picture.

Maybe you should use the spinning slotted metal disc approach after all, to remain within the spirit of steampunk.

[Rick Turner]

That and a mechanical TV Mechanical TV Sets of the 20s and 30s and I'm set, so to speak...

[Joe Gwinn]

That and a mechanical TV Mechanical TV Sets of the 20s and 30s and I'm set, so to speak...

Definitely in the spirit.

For a winder, no scanning is needed, just an intermittant aperture through which to look. The spinning disk would be geared to the bobbin axle.

And the winder needs to be steam powered.

[Kindly Killer]

I just figured out the strobe joke: it's not a joke - that's to "freeze" the assembly so you can see it in situ while you wind, right? Like freezing motion with a camera flash?

Is that a common thing to build into a winder? If you're making enough to worry about QC during the wind, wouldn't it be more efficient to just throw each dud in a bin for rebuilding? Still not sure whether I get it...