CNC Pickup Winder

[Hugh Evans]

I can understand and appreciate the desire to control every winding variable, and want to incorporate a reliable tensioning system, but for something as simple as a guitar pickup and lacking data on real world variation this discussion is bordering on swatting a fly with a sledgehammer. From a design standpoint, what I'm leaning towards is a purely mechanical system and allowing the tension to be set with a spring because it's simple and elegant. It would also be possible to determine the extent of any tension variation with minimal effort. If it becomes clear that a lot is going on then a more elaborate measurement system can be used. I'll be sure to leave plenty of room for the addition of servos, transducers, etc. to allow for modifications. My entire vision for this project is to make a very capable, flexible winder that will be reasonably accessible for just about anyone wanting to get into coil winding.

Part of my objection to using a cheap digital balance is that it strikes me as something that will end up being more of a pain to actually set up than it appears at first glance. It should also be possible to track down the actual component from a chinese manufacturer and probably get them for $1/ea.

As far as strain gauges are concerned, the properties of the beam will be needed first to determine the required specs. They generally cost less than $10. Also, the smaller ones I'm talking about almost always have pressure sensitive adhesive backing. It's really not a bad way of going, and it's not a CA and duct tape solution. Since I can't tell if you're being sarcastic or genuinely cynical it would be awesome if you wouldn't mind turning it back a notch or two... You have plenty of valuable thoughts and considerations, which I absolutely appreciate, but I can't get a read on you. I don't think there's anything I've said to insult you, and if I did in any way it wasn't intentional. In the design process I question a lot of things in search of the best solutions, and I never claim to be right all the time or assume that I am intellectually superior to anyone.

[Hugh Evans]

If we can find a good price a magnetic particle brake would be excellent, but a sufficient magnetic clutch and brake assembly can be had for ~$20 since we're not talking about a large torque. There are a few options, especially with a closed loop design, where the dancer arm is concerned and salvarsan is absolutely correct in his description. Undesirable oscillations can be avoided with proper design. A position sensor on the dancer arm will allow for exact tension readings. Hall effect will probably be the best approach because it will avoid loading the dancer arm significantly.

Overall, it shouldn't be overly challenging to design and build... But it can almost certainly be expected to end up on the pricey side. Just for all of the parts we're probably looking at $50 to $100... In addition to the rest of the winder. The added complexity makes me much more apprehensive about running all of this from an Arduino. Not to mention enough shields and other electronics will be involved that controlling the whole mess from a computer might be more economical. A Mach3 license runs all of $140 the last time I checked, and it's designed to allow for custom interfaces such as coil winding. Still, a comparable coil winder from a major manufacturer would likely run a few grand. The good news for anyone who already has a coil winder is that this tensioner could function just fine as a stand alone unit.

Thoughts?

[copperheadroads]

One way to get a wire tension measurement is run the wire over a roller attached to a load sensor.

Industrially, the sensor is either a parallel beam or a bending beam load cell.
[ATTACH=CONFIG]20776[/ATTACH][ATTACH=CONFIG]20777[/ATTACH]
In the good ole US of A, they are $120+ from futek.com, rather "F.U. Tek"

Pulling apart a $20 Chinese 100 gram digital scale, I note a parallel beam load cell rated for 250g max.

I'm wondering if a DIY digital wire tension gauge couldn't eventually become the heart of a DIY servo wire tensioner.

-hizself

I am certainly interested in a way of measuring tension ,I'm curious if this cell would do the trick
Electronic Scale 0-300g Range Weighing Sensor Load Cell | eBay

[Hugh Evans]

Yep, something along those lines would do the job. I'll see if I can dig a bit deeper for more specs. The easiest solution would probably be sticking something like this under a pulley in the tensioning system, calibrating it and that should be it. I figured there would be good options available directly from Southeast Asia. Not long ago I got ten limit switches for the price of one (from the same manufacturer) by ordering direct.

[salvarsan]

Hugh Evans wrote:
...but for something as simple as a guitar pickup and lacking data on real world variation this discussion is bordering on swatting a fly with a sledgehammer.

Tension control is a need, not a desire. I received a boutique pickup for repair because the builder's repair price was the same as his ~$200 sale price. Resistance+inductance measurements indicated wire cut-through on the third layer. That wouldn't have happened with effective tension control and I wouldn't have a caustic opinion of a certain Black Guard(tm) tele pickup builder.

Part of my objection to using a cheap digital balance is that it strikes me as something that will end up being more of a pain to actually set up than it appears at first glance.

Possibly, but difficult machinery is poorly designed machinery. Ideally, it should be like a toaster.

As far as strain gauges are concerned, the properties of the beam will be needed first to determine the required specs. They generally cost less than $10. Also, the smaller ones I'm talking about almost always have pressure sensitive adhesive backing. It's really not a bad way of going, and it's not a CA and duct tape solution.

You misunderstood my question about strain gauges.

There are several general technologies. It wasn't clear that you meant the "resistance wire on kapton sheet" type.

My intent is one of immediate prototyping and you seem to have some serious industrial experience and a brain worth picking.

Beam properties: If it deflects enough under a 1 oz. calibration load, it's usable. Polycarbonate scrap may be all I need for proof of concept. FYI, a quarter + 4 nickels + a penny weighs within 1% of an ounce.

N.B. I just ordered a couple of 100g beam cells off eBay since they were already temperature compensated and I won't need to kludge up a resistance bridge. If accuracy is better than 5%, I'm happy (joke: beams cells are 1.0% accurate on a bad day). Ringing for this size cell is usually greater than 4kHz so the deflect+restore period will detect tension oscillations on a pickup winder if you run it slower than 24,000 rpm (note: factual but tongue in cheek).

Since I can't tell if you're being sarcastic or genuinely cynical ...it would be awesome if you wouldn't mind turning it back a notch or two...

Most of my stuff should be read with tongue firmly in cheek since I *am* genuinely sarcastic/cynical/ironic, have a low taste for arcane puns and worse, talk too much. When in doubt, refer to the Dorothy Parker quote at the end.

When the questions are short and direct, assume them to be a sincere information request. When statements include design links, patent links, primary reference quotes, computations, or code snippets, you may safely assume that they are exactly what they purport to be, although of varying degrees of pertinence and correctness.

The only thing we can promise around here is that we won't insult your Mother (unless she's an easy target) and we don't insult sheep by suggesting that you go about shagging them.

... but I can't get a read on you. ...I don't think there's anything I've said to insult you,

Just call me on my bullshit or ignore me. Pickup Makers Forum can be like that.

Overall, it shouldn't be overly challenging to design and build... But it can almost certainly be expected to end up on the pricey side. Just for all of the parts we're probably looking at $50 to $100... In addition to the rest of the winder. The added complexity makes me much more apprehensive about running all of this from an Arduino.

Got it.
You want to build a robust simple auto-winder.
A tensioner, while significant, is not germane and should rightly be a separate project.

May I suggest a prototype benchmark of winding correctly at 450 rpm?
You can do 450 rpm <-> 60 Hz interconversions in your head, and a modest 4500 wind pickup finishes in only 10 minutes ... proof of pudding and so forth.

Faster is for later.

Many here have built a winder 'lathe', a few have built an auto traverse, but probably none has synchronized it to the winder speed.

Belatedly, welcome to the group, and I speak for all of us in saying that we would love to see you succeed in making a solid little auto-winder.

-hizself

[Joe Gwinn]

A good solution is usually a mix of methods.
The Tanac patent US457087 shows you the tensioner concept in beautiful simplicity.

A mechanical tensioner sets the initial tension with a brake (felt pad, pinch roller, magnetic brake, something?).
A spring-loaded dancer arm often handles the fast changes.
The time constant of the spring+arm and damping from the brake determine how bad the oscillation gets at resonance.

My first guess would be to apply magnetic or friction braking to control the tension.

Current Tanac tensioners use the hysteresis drag without the feedback, as discussed in http://music-electronics-forum.com/t30765/.

With the spring loaded dancer, it's necessary to ensure that mechanical resonance is well above the pulsation frequency (40 Hz at 1200 rpm). A bit of damping also helps.

[Hugh Evans]

Cool, then we're on the same page essentially.

Your story regarding the fancy overpriced pickup from a maker who refused to stand behind their product genuinely angers me. It should never have been a problem in the first place and I further see your point that people are out there flying blind because off the shelf tensioning solutions are prohibitively expensive. I don't know what kind of equipment you have access to, but as long as we can reach an agreement on the development license, pooling resources to build a quality prototype is fine with me. I have plenty of sheet plastic to cut parts from that will be more than sufficient for prototyping and have no qualms providing CNC machined parts given that you're already investing in components for the build... And you sound like you're ready to start testing ASAP. Sounds pretty close to a win-win to me.

I'll dig in on the design and get some idea of what can be sourced off the shelf and what will require fabrication. Oscillations will be addressed during the design process. It's been a few years since the last time I took a vibration analysis class, but fortunately my fiancé is finishing up her mechanical engineering studies so I can ruin part of her day by getting a double check of my analysis.

Even though you're getting some load cells and they could be a great solution... Let me know if you would like to try some calibrated pieces of polycarbonate. I can measure the modulus, cut some appropriately sized pieces, and tension could then be accurately determined by deflection. Another easy method would be simply engraving known static tension based on dancer arm position (my personal vote if we don't go for a full closed loop system.) As long as the load cell turns out as I expect it to, I can't imagine being successful in making a case against using such a reasonably priced and ideal solution.

Does this sound like a good starting point to you?

[salvarsan]

Hugh Evans wrote:
Cool, then we're on the same page essentially.

Right.
Simple design, inexpensive materials, facile implementation, no jibes about sheep shagging.

...as long as we can reach an agreement on the development license,

If a project uses out-of-patent or unpatentable technology and its development details are discussed in public, then an open source license is prudent.

Creative Commons License Creative Commons
CERN Open Hardware License CERN Open Hardware Licence - Cernohl - Open Hardware Repository
TAPR Open Hardware License The TAPR Open Hardware License

Oscillations will be addressed during the design process.
...
It's been a few years since the last time I took a vibration analysis class,

Give me a moment to do a first cut on vibration.
It may be simpler than we think, and I have a suspicion that non-longitudinal vibration in the copper wire may dominate anything a tensioner might contribute.

... my fiancé ...

If you have time to spend with your fiancé, you might not want to share her with an engineering project. :-D

Let me know if you would like to try some calibrated pieces of polycarbonate.

Polycarbonate was thrown out as an example of obtainium.
Any material whose deflect-restore motion produces enough sensor Delta-R should work.
All random materials need calibration, if only at the force of interest.

Term of art: force/tension of interest.
Magnet wire maximum winding tension is derived from the elastic limit for copper 101, or 10,000 psi.
Recommended winding tensions are 50% - 80% of that.
MWS Wire uses cautious numbers in their chart, recommending 17/22 grams min/max tension for AWG 42 without mentioning the 25 gram elastic limit. http://www.mwswire.com/pdf_files/mws...ook/page36.pdf

Note my recurrent mention of 25 grams which is close to 1 oz (28.3 grams) and trivially approximated by a few coins.

Another easy method would be simply engraving known static tension based on dancer arm position (my personal vote if we don't go for a full closed loop system.) As long as the load cell turns out as I expect it to, I can't imagine being successful in making a case against using such a reasonably priced and ideal solution. Does this sound like a good starting point to you?

Cheap is very, very good.

Hooke's Law analysis of the dancer arm+spring(s) is probably enough and the arm material may be the only crucial part.
"Light enough, strong enough, cheap and ubiquitous" usually describes aluminum.

Pragmatically, a partial tension control is all that's necessary. It would allow builders to confidently make coils with controlled microphonics, perhaps avoiding the need to pot coils later.

We don't have numbers on wire tension during the winding process.

Numerical data on tension at different winding speeds may be all we need to fine-tune a bone-head simple mechanical tensioner.

I'd like to see a few empirically-based rules of thumb, such as:
"If your static tension is 17 grams at 1200 rpm, lower it to 12 grams for 2400 rpm", and
"A freshly-minted U.S. nickel weighs 5.000 grams."

So for now, I plan to instrument the winding process.
If I can nail down tension particulars of using a 350 ohm strain sensor with an Arduino without it self-heating, that may encourage others to do better.
Better?
You wouldn't believe the great pickup electrical data posted here once people got comfortable with Extech's LCR meter. Some have moved up to Bode plots from SysComp's USB scope+signal generator.

[salvarsan]

Note to self:
A miswired strain gauge provoked the coining of Murphy's Law.

[Hugh Evans]

Haha, sounds like you're speaking from recent experience. I'll be curious to hear how your initial tests turn out. I haven't had much time to work on the rest of the design recently... My Sister's Father-in-law passed away suddenly and very unexpectedly this week, which included all kinds of emergency travel for my Brother-in-law and Sister from GA to OH.

I have a pile of parts set out to prototype my proprietary active filter/low-z pickup system as soon as I have time. Family comes first, but it still doesn't keep me from chomping at the bit to start working on this project. Not to mention our first endorsing band, Allegaeon (they're difficult to place into a definite genre aside from having a basis in technical and melodic death metal) is scheduled to tour through Columbus in early December and I'm hell bent on showing off what we're up to even though their custom 8-strings won't be ready yet. There's a good chance that I will be able to get the external power supplies for their active pickups up and running by then as well. Words cannot describe how ready I am to scale up into manufacturing mode.

[salvarsan]

It doesn't wirk rite when you mix the excitation leads with the measurement leads. Der.

[David King]

I've been away for a month. Nice to see interesting stuff happening around here.

[David Schwab]

I've been away for a month. Nice to see interesting stuff happening around here.

I know! I had no power for eight days and there are lots of new posts here.

[salvarsan]

I've been away for a month.

...and you are?

[David Schwab]

...and you are?

Can't you read... it's King David!