As I am getting more heavily into guitar/bass pickup theory, testing and design, I have found it necessary to get more consistent results in terms of winding. Until now, my ‘winder’ has simply been a drill clamped to my workbench, with the wire guided and tensioned by hand. This is ‘good enough’ ™ for general prototyping and testing but not good for consistency from one coil to another: the best results I have been able to accomplish with a simple drill winder is within 5%, which is not TOO bad, but it would be nice to have the option to get tighter tolerances and more control over tension and winding pattern.
Enter this beast:
I have a distinct DIY advantage where I live: there are several excellent thrift stores and pawn shops within walking distance, as well as several nice mom-and-pop hardware stores with extensive nut and bolt bins. So I set out to gather the materials, and ended up with:
-monitor/TV wall mount
-misc bicycle parts
-screen door rollers
-misc metal braces and straps
-misc nuts, washers, grommets and bolts
-roll of cork
-self-adhesive felt pad
-‘Rock-band’ game controller foot pedal
-Wall-wart power supply
(The total cost was around $30-40)
I would LOVE to make a CNC version, maybe with an Arduino and some stepper motors, but that is a bit out the price range of this humble guitarist/mad-scientist at the moment.. but maybe in the future.
The actual platform and body are made from the TV stand, with the bed being formed by the Ikea shelf. I started with an A-frame, and then mounted side arms for the bed, and adjustable ‘feet’ rails underneath. It is built like tank, and weighs about 7 pounds by itself. The whole thing is resting on a neoprene mouse pad and cork to cut down on vibration noise (gotta keep the downstairs neighbors happy). The bed is mounted to the rails with wing nuts, so it can be slid back to provide clearance for larger bobbins.
The wire des-reeling portion in the back is made from a single mounting bolt with a large washer to secure the spool, plastic funnel, a wisker disk (made from 2 washers, duct tape, and the bristles of a paintbrush), and a nylon bushing on a metal brace arm, held in place with a rubber grommet.
The roller-guide is made from several screen door rollers, which are nylon wheels with ball-bearings. These are mounted to a long steel mending brace, and bolted to the a-frame with several square braces. I made sure to sand all the rollers and plastic parts to make sure there were no molding marks to snag the wire.
The tensioner is another pair of square braces with felt mounted to each side, and thumb-screws to set the tension.
The ‘dancer-arm’ is made from another screen-door part, mounted to the main brace, and bent into form. I have several variations on this for different coil shapes. For more oblong coils, i have a bicycle-spoke dancer arm, which has a longer range of travel.
I took the motor and chuck from the cordless drill and mounted them on more braces from the TV mount. In the picture it looks like they are only secured with straps, but there is a mounting bolt underneath, so they both have several points of contact for stability. The mounting rails are bolted to the bed through the neoprene mouse pad to cut down on vibration noise (which works pretty well). They are secured by bolts and pipe-clamps/straps. I used a ‘wall-wart’ adapter to power the motor. I just made an educated guesstimate in terms of amperage based on the specs of the drill’s old battery pack (I figured a few amps would be sufficient). As a safety feature, I ran the power adapter through a 10w ceramic resistor (about 0.5 ohms) to cut down on voltage spikes, arcing, and the risk of motor burnout. The voltage of the power supply controls the speed of the motor, so I just tried several until I was happy with the result. The neutral lead from the wall-wart is interrupted by a foot pedal game controller from the ‘Rock Band’ game. I mounted copper plates that make contact to engage the winder motor when I step on the pedal.
The auto-traverse is made from a disassembled fishing reel that I got for a few bucks at the thrift store. I had to try a few until I found one with a good enough mechanism to not have the windings bunch up on the ends of the bobbin. The ratio of the mechanism in it’s raw form is about 6 winds per traverse. I wanted to gear for 25 winds per traverse, so I repurposed the reel’s crank gears to change the ratio. I used rubber plumbing caskets to drive the traverse from the main winder shaft. I had some belt-slippage initially, but I replaced all the bushings with ball bearings (repurposed from bicycle and drill parts) and now it works perfectly. The gears are mounted through ball bearings, with housings/mountings made from more bicycle parts (seat post clamps and water-bottle mounts). They are secured to the bed by single bolts, which double as fulcrums to keep the gears tensioned against each-other via the drive-belt (because there is always slight play in bearings).
Here is a pic of the winding pattern on a drinking straw. The ‘fill’ is about 60%, and the pattern ends up in a criss-cross shape. It simulates loose (low capacitance) scatter winding well, but allows for more consistent results than hand-guiding does from coil-to-coil. My current project setup is for 0.5”-0.75” round solenoid bobbins using #38awg double-insulated wire, but the traverse-length can be changed by moving the traverse arm/cam and readjusting the tension and winding speed. (I will improve this in the future to have the adjustments easier to do quickly.)
The counter is a repurposed calculator. There are many videos online about how to do this, so I won’t go into details other than to say that I make contact via a flat copper wires and small neodymium magnets. The calculator can only count about 250 rpm accurately (around 4.2Hz), so I mount it to one of the slave shafts and count in multiples of 10x. I have recently ordered a 20Hz counter from eBay (which is around 1200 rpm), so I can more accurately count individual winds, but that has not arrived yet. I will give more details when the unit arrives and i have a chance to put it through it’s paces.
PERFORMANCE, CAVEATS, AND AREAS FOR IMPROVEMENT:
The winder works better than expected, but it is far from perfect. Adjustments are possible but difficult, and they require moving mounts, changing belts/gears, and rewiring the power supply. Areas for improvement are:
-ability to wind 2 coils at once for even closer matching
-phase-reversal switch to reverse winding direction without having to rewire the power supply
-rheostat to vary speed without having to switch power supply voltage
-more accurate and control-able traverse length
-easier traverse fill-density adjustments
-weighted flywheel for more consistent rpm on the main shaft
-better mechanism to secure bobbins
-ability to wind directly onto solid cored rather than bobbins
-tension gauge (fairly simple to rig up… I just haven’t bothered yet)
-hand-crank traverse for manually setting winding pattern
-likely eventual CNC control of the whole thing (I am currently messing with a stepper motor form a CDR drive, so I may change to an electronic traverse system rather than a mechanical one if I can get that working properly)
So there it is! A functional coil-winder with a decent set of features for a fraction of the price of a retail unit.
UPDATE 3/10/2016: here is the unit updated with the electronic traverse from this post