Pick up winder ideas

[Rj777]

Hello
I'm just getting started in winding pick ups and am looking for ideas for a cheap and simple winder set-up. I have looked on here and the best ones seem to be mini late style ones with a tail stock or at least have a flywheel/face plate to balance and smooth the spinning coil.

I have small lathes out in the garage but ideally I want something small like the Schatten to use in doors and the closest thing that I can see to that is to mod a small bench grinder. (decent horizontal shaft in bearings all aligned and mounted)

Has any one successfully used one like this? I have seen a couple on the web but all look a bit Heath Robinson. I think the main problem will be getting the speed down and making it varible between 0-750rpm. (Is that the correct range for pick up winding?)

I quite like the idea of fitting a hand wheel to the left side to start the first few turns by hand. The counter sensor could also be incorporated in to it to make it nice and neat.

Sorry for so many questions,maybe the sewing macine route would be easier?

Thanks in advance

[Slausonm]

I built this one. It is fairly inexpensive. I'm new to winding and didn't want to spend a huge amount either. I was also not too excited about the electric drill and sewing machine units on the web. This solves the counting and the motor control issue with some fairly simple electronics.

Pickup_Winder.wmv - YouTube

Matt

[Rj777]

Cheers, are you going to post details? I wanted that type of screen read out but thought it would be to expensive.

Anyway I have spent a few hours in the workshop and come up with a working winder. The idea was to keep it cheap and as easy as possible. I used an old metal case as the basis, a 24v servo motor with a 3 ch encoder directly driving the faceplate. The index channel of the encoder sends the pulse to a rewired pedometer. all nice and compact.

I made the face plate out of Al and had to go with a paddle style so i I could drill and tap the holding screws (I dont have a long 3mm drill and tap at the moment) its not my best work but it will do for now.



The only bits i had to buy was the pedometer (£6) and motor speed controller (£15)which although it is only 12v I still get 800rpm with plenty of torque so it wont bog down in use. The only sticking point is that the pedometer is restricted to 5 pulses a sec so it max's out at 300rpm, but that should be fine to get me started.

I just need to sort the wire guides and get some 42 AWG and I should be in bussiness. quite pleased for £21 and an afternoons work




The battery is a temp fix to supply the 5v to the encoder.

[Slausonm]

Hello rj777,
The entire platform is based on using an Arduino microcontroller. It operates the motor control shield and the LCD display. I have a magnetic reed switch that is pulled low with a 10K resistor. When the magnet in the flywheel passes the switch it goes high and increments a variable in the program. The LCD display was purchased on ebay for about $4. I used the type of parallel display recommended for the arduino LCD library. This part was actually very easy. I used the analog pins on the Arduino for the lcd display by converting them in the program to digital outputs. This kept me from having to figure out which digital pins the motor shield was using. The biggest problem I had was learning how to debounce the signal from the switch, but luckily someone had already done this bit of coding. The motor shield is powered by a 12V transformer. It has a jumper on it to enable the power supply to power the 5 volt logic for the microcontroller as well. Since I already had the microcontroller there wasn't a lot of extra costs. The microcontrollers and motor shields are being sold by many vendors. All together I have about $60 US tied up in this winder. I will no doubt have more learning and modifying to do, but for now I have about 20 pickups to wind for the kids project at school. If you want specific info on the program let me know. FYI, I borrowed lots of code from the Arduino site to figure out the program.

[Rj777]

Good work. yes i would like more info please. I think the Arduino stuff is a bit more expensive over here tho. what board do you use and can it be used for full control?

I plan to go full cnc for V2 but I like the fact the that pick up making can be a nice non pc based hobby, I have enough cnc projects that I dont have time to finish at the moment. Theres also something to be said for a simple winder that just needs 240v, speed control and a careful eye

That said by using the encoder and the traverse slide I will be able to add a data logger to the winder and digitally record the winding, so if and when I make a good hand wound one I have the recipe to reproduce it by machine.

[Slausonm]

Hello again,
Yes I think the 110 - 220 difference between the US and Europe probably makes a lot of things more expensive for the UK and the rest of Europe. The Arduino stuff is very cool and fairly easy to use. There is a ton of information to use though.

I don't have the code for the Arduino on this computer, so if you want to see my code I'll have to get it to you later. The motor shield has the capability of controlling two stepper motors, so it should be possible to build a "CNC" two axis winder. The motor shield came from Adafruit Industries, Unique & fun DIY electronics and kits but I just ordered two from an e-bay site that have not arrived yet. One issue with the Arduino Uno board is that I don't think it has enough pins to do two axis with steppers and control the LCD display. I think you would need an Arduino Mega board to do this, which of course is more expensive. One nice thing about the motor library in arduino is that the DC motor outputs are controlled by PWM. My code sets the speed of the motor at a constant, but when i get better at it I'm going to add a variable input to control speed. I'd like to know more about how your encoder system works with the pedometer because I think it would add more functionality to my system.

I'll send you a PM with code when I get to my other computer. I'm not a good programmer so I borrow a lot from the Arduino site. I have to put together a wiring diagram for someone else so I'll send that too. FYI, I've built four CNC machines as well. I'm more of a hardware guy.

Regards, Matt

[Rj777]

Hi Matt
I think I was looking at the high end boards, a quick search and I can get the Uno for £25, lcd £10 and another £25 for the motor shield so I'm quite tempted to give it a go. I'll have a look on ebay first tho.

Is the standard 12v 2amp motor shield ok or could it do with a bit more power? I was also wondering if its possible to have the display read both turns and rpm. I.e

No. Turn:####
RPM:###

Dose the display and motor card use up all of the I/O on the uno or is their spare capacity to add encoder outputs for data logging? it would be ideal if it could output coordinates of each wind on to a text file. 6000 turns would produce a lot of data but would be a perfect way to be able to reproduce the wind again if desired. I see that you can even get pressure sensors so you could monitor the wire tension if it was necessary.

I can see may possibilities for this.......................

The encoder bit is quite simple. not sure how much you know so I'll explain it all (better get a pillow). the encoder is a 3ch type so it has A & B channels that give 512 pulses/rotation which are out of phase so the controller can see which way the motor is spinning, ie ChA then ChB = clockwise, ChB then ChA = counter clockwise for example. the Index channel just gives 1 pulse/rotation and is used for accurate positioning, normally the home switch is mechanical type that has to much variation so once triggered the motor then turns to the next index pulse as an accurate position.

and now the short answer that you probably wanted in the first place......the index channel has 1 PPR and is wired direct to the pedometer, I removed the mechanical switch and just soldered the leads direct to the contact pads. Despite the pedometer being 1.5v it dosnt seem to mind 5v TTL for the encoder.

I was going to ask about your CNC machines but this probably isn't the place do you have it documented elsewhere.

[Slausonm]

The motor shield has dedicated pins on the arduino uno for the different types of motors it will run. If you dig into the documentation you can probably use the ports not being used to run your motors. The hobby servo ports might be a good candidate for the data logging and inputs you want to do. There are 6 analog pins that can be converted to digital outputs, which is how I ran my display.I think you will also be able to display turns and rpms. My display has just got text on the top line. No reason it couldn't display data from a second variable like rpms. As far as the 2A current draw, there is a hack on the adafruit site for piggybacking another motor control chip to supply more current, but they warn you that the magic smoke could come out. For my application I'm happy with the stock amperage.

Thanks for the encoder info. I've read enough about quadrature encoders to know the concept, but I've never put them in a project. I have to admit I understand what you want to do with the data logging, but I'm not sure how you're going to reverse it and create a duplicate winding from the text file. I have some youtube videos with my cnc and other stuff I've made...some are way too long and boring. Do a search on slausonm and you will find my stuff.

Too bad there's a large body of water separating us, sounds like we have similar interests. My brother spent a semester in college in the UK and loved it.

regards, Matt

[Rj777]

cheers, Looks like I will have to get one and have a play. I was going to wait and have a look a PIC circuits but this looks like a much easier route.

I would appreciate it if you could Pm any info that might speed up the learning curve. Yeah sorry for the encoder rant, wasnt sure which bits you were after.

I will need to get to grips with what the Arduino can do but my logic is that the encoder on the motor (M) gives info on its rotation position say for arguments sake 10 PPR and then the wire feed that although controlled by hand will be linked to another encoder (W)(linear or rotary) also 10 PPR then at any given time you will effectively have coordinates like G-code,


M1,W1
M2,W1
M3,W2
...............etc

although time consuming it would theoretically be possible to make an exact copy with all the natural variations of hand wound providing the CNC winder had the same resolution as the origial one that is.

I suppose the real trick would be to use 200PR encoders then get the Arduino to directly output g-code and then use steppers for the CNC winder

Its was only an idea originally but now your have brought these processors to my attention you have got me thinking .......I do have bit of an over active imagination tho and there's probably lots of flaws in there but has all part of the fun eh

1 more project to add to the list

[Rj777]

This has got me thinking Looking at the programming code examples it seems very simple.

The easiest way is to initially set up the resolution of the encoders or sensors to measure real world values ie degress or mm. that way the g code can then be used on any machine.

Realistically the bobbin only needs to be sampled twice per revolution but at the ends of the coil as they pass top dead centre each time (as the wire is held at the ends and then lays in a straight line to the other end) this removes the need for an encoder and two reed or otpo switches could be used. for count and rpm the software can divide by 2 and the varibles for the arduino labeled 100, 200 (if using 200 step steppers)

using the print and println the X value for 2 revs would look like this

X100
X200
X100
X200

In g code I think this would make the motor constantly spin as it interpolates the rest of the positions. By using the feed rate command that would set the speed

Then if the pulse from the bobbin is used to capture the position of the wire feed traverse that can also be printed next the X value and the N code can also be incremented +1 each line

N60 X100 Y0 (red text is fixed values in the string)
N60+1 X200 Y1
N61+1 X100 Y2
N62+1 X200 Y5
N63+1 X100 Y9
N64+1 X200 Y1

I dont think it matters if the Y values jump as again the cnc machine will calculate the required timing so it arrives at that point after 200 pulses of the X axis

If it is possible to measure wire tension then that could also be added as a Z value.

I agree that the exact resolutions may need to be tweaked and also if I was going to use a stepper for the bobbin on the cnc side I would probably gear it 2:1 so it can wind at 1000 rpm but only run at 500rpm to keep it cool and the noise down. But as a starting concept can any one spot any obvious problems?

[Slausonm]

At 6-8000 windings, collecting that much data might run you out of memory on the microcontroller.

[Rj777]

At 6-8000 windings, collecting that much data might run you out of memory on the microcontroller.

yeah I think that will be a problem. It will have to be done using the serial monitor as it would be simple to copy and paste the text. The other option is to add and SD card reader if it had to be stand alone.

I've just bought one so let the tinkering commence

[Slausonm]

yeah I think that will be a problem. It will have to be done using the serial monitor as it would be simple to copy and paste the text. The other option is to add and SD card reader if it had to be stand alone.

I've just bought one so let the tinkering commence

Yeah I think there might be a project in the works some day to make some kind of effects pedal.

[enkindler]

You may want to look at existing Arduino Software for the reprap 3d printer project.

They already use G code but in an XYZ fashion, it would be trivial to take one of their existing systems and modify it to work with synchronized spindle Gcode.

Or you could write a sketch or just use a LPT port on an old pc and use EMC2, which is linux based or Mach3 as the Gcode Interpreter.

It you look out on github there are already a few gcode generators for coil winding but most do not do scatter gather at they tend to be producing coils that they want perfect winds on.

Also note that one pulse per rpm of the spindle is enough for lots of CNC lathes that produce precision threads, I would see if you can get it working with that first, the AVR Can run out of steam pretty quick if your code gets a little sloppy.

I am winding on a harbor freight 7x10 lathe that I just finished extending to 7x14, it will be a while before I finish the complete CNC conversion it but with just a simple interrupter disk with one pulse per revolution I have EMC2 doing smooth acceleration and stop with counts.

[Rj777]

Thanks for those links, ive looked at a lot of those coil winders and although very good, they all seem to just output pulse and direction data that need separate stepper drivers, which as you mentioned could just be hooked direct to Mach 3 and run from a PC.

I think that winding perfect coils or threading is relatively easy as its just a case of advancing a calculated distance in a smooth motion, monitoring 1 pulse a rev is sufficient as there are no sharp changes in direction. Scatter winding at low speed could cause missed change of direction and give a false recording which is why I think it is necessary to capture at both ends of the bobbin(I'll test the theory when I get that far )

I should start a new thread really as it started out as a simple coil winder project with a digital counter, but thanks to Matt and the Arduino I now want to create a data logger to act as digital duplicator.

It would also be nice to use the AVR as the motor driver and build a true stand alone unit. I'm still very new to the AVR programming and learning what it can and cant do. I'm try to learn the least way of programming so it saves space.


I would defiantly use an optosensor like an encoder as it doesn't need de bouncing in software. I have sorted the count and RPM using short interrupts to use the minimun of code so next is PWM motor control,mosfet H bridge and PID feed back .

Should be finished by christmas..........