Project "Groove Scribe"- A DYI stereo 45/45 head.

[EpicenterBryan]

Bryan what do you think?

If you are analyzing a Ortofon stylus playback, you will notice a small magnet at the inner end that fits perfectly between the coils.

I'm with Ciuens on this one. Those cartridges are way to precisely built inside with super small moving magnets. There is no chance we would be able to detect Left and Right movement with one cartridge and one magnet located outside the cartridge. That doesn't say that you couldn't use two, and two magnets... one of each for each push rod. That also means closer proximity to the driver coils - cramped space and much more need for really good magnetic shielding on the driver (kind of like what I've been trying on my test jig with the Ciuens Shield).

The configuration is not taking advantage of most of the flux from the fixed magnet. It would be better if the magnet were standing up so its n/s line was following the spiral of the coil.

Todd has really nailed that, and I know it comes from experience. It really complicates assembly when using these drivers though. But I'm sure there are 1/2 cylinder magnets with pole orientation along the long axis or some other easier to find semi Horseshoe looking magnet pair that could be found. Then they could be added around the push rod / coil after the fact in assembly. That's a research project I'm sure Todd has gone through - So much and so little info on magnets out there! I love those things! Full of mystery!

Bryan

[EpicenterBryan]

Todd, I believe that everyone here (including you) are trying to deploy feedback for several reasons. I concluded that feedback is more difficult to achieve than the dynamic cutting head.

I'm totally with you Ciuens.
My original goal for feedback was to have an automatic (and simple way) to not only fix crappy frequency response for my DIY head, but also remove some of the endless variations and variables that I couldn't get my head (pardon the pun) around...

It seamed like everything changed everything each time I cut frequency response curves and tried to compensate with EQ. And it did. Originally, part of it was my lack of understanding on how it all should work. Part of it related to flaws in gear used (like a card that wasn't sampling at the rate I told it to). But as I went through the process trying to nail down why - and shared flawed results - not understanding why, we all discovered things. We also figured out why they did some things years ago when these heads were designed. All in an effort to make this work on two groove walls on a spinning disc. Amazing!

We also uncovered common errors we were all making messing around on our own trying to make something that worked. In the past many of us would have given up. But now, through this fabulous website each of us messing around in our basement can share what we have discovered and we are all learning from each success and each failure. We have many failures. But we have also had some great successes!

We have many rock stars among us, all chiming in. You all know who you are. I for one thank everyone for all inputs!

And now, one of our contributors has started his own thread - Finally - with some really great results, and a very nice design!
Groove Scribe is not dead, but I want to make sure everyone checks out what Ciuens is up to on his new thread ZEZ Cutter One here:
http://lathetrolls.com/viewtopic.php?f=15&t=5945

Bryan

[EpicenterBryan]

Hey guys,
Groove scribe is alive and well but it is going through many changes yet to be disclosed.
Many parts have been machined and actually fit! Some parts are needed and are due to arrive tomorrow. Some items are delayed a week or more.

So here is the deal... I'm pretty well convinced that using Nylon for the spring is a viable solution and that the Carbon fiber disc for the driver works. The net is reduced mass, and some added dampening is applied. I'm going to let that ride as a done deal until Bennett has time to play with print settings for the nylon and he comes up with some printer settings that work better for this material.

I'm convinced that the drivers have a massive magnetic field that needs to be shielded if feedback is possible internally or externally in any form. I've got that nailed down in my new version - externally. After messing around with a phono cartridge externally as a feedback device and the shielding required to make that happen - I have decided it is a bit more work up front but worth winding and installing my own feedback coils with some special shielding sauce to keep what happens in Vegas to totally stay in Vegas. The feedback is inspired by what Ciuens has been up to with practical inputs from Todd about the magnets that should be used, and my own mashup - being able to have the same polarity magnetic force oppose each other with a ferrous spacer placed in-between. It sounds complicated. It isn't and will be clear to all when I post the update.

So that brings up what I wanted to avoid at all cost - winding coils. Voice coils - feedback coils - any coil for any reason.
I'm going to put together an simple Arduino based coil winder. The first (and I hope only version) will be extremely primitive but only has to wind a few coils... Parts are on order and I may have something working this weekend. I found some great calculators on line that you can input things and get a rough idea on resulting resistance, length of wire needed and such - check this out...
http://www.daycounter.com/Calculators/Coil-Physical-Properties-Calculator.phtml

That's it for tonight. No photos on this update but stay tuned! It will be worth the wait.

B

[opcode66]

No machine will ever give you this. I recently graduated from Coil University.

4 layers. 120 turns.

[EpicenterBryan]

I recently graduated from Coil University.

With a PHD I'll add. Great looking coil there Todd.
So you hand wind wind those? Do you have to use a microscope? Have you got some kind of turn counter on the shaft?
Man, I really wanted to avoid this....

[Bahndahn]

Nice coils, Todd!

A computer controlled coil winding mechanism is totally do-able and a great option. It will be possible to get great results with it. My atlas/craftsman 618 metal lathe has coil winding possibilities that are described in the user manual, using the threading gears setup with pitch=wire diameter. Direction has to be manually changed.

If anyone wants to wind flat-wire coils but doesn't want to flatten the wire themselves, various distributers carry stock of that sort. To name a few:

http://www.alphacoredirect.com/contents/en-us/d4_flat_magnetic_wire_specialty_wire.html

http://www.hmwire.com/flatmagnet.html

http://www.thelitzpeople.com/flat_magnet_wire.php

• • •

I'm excited to get rolling on the experiments with materials Bryan has sent me! Midterm exams will hold me off a little bit, but more as soon as possible, I will be sharing my findings!

[EpicenterBryan]

So what is the real advantage of using flat coil wire? Flat means more layers in the same space, but fewer wraps per layer. But I guess it increases density overall by eliminating dead space from around normal round wire and reduces mass as a result? I assume this is really only needed for for drive coils, since feedback wire can be so much smaller due to the super low currents involved. Right?

b

[opcode66]

So you hand wind wind those? Do you have to use a microscope? Have you got some kind of turn counter on the shaft?

I have a manual coil winding machine that you can purchase on ebay. http://www.ebay.com/itm/Manual-Automatic-Hand-Coil-Winder-Winding-Machine-NZ-2-One-Year-Warranty-/321714228013?hash=item4ae7a98b2d

I don't use the automatic traverse mechanism. There is just no way to use that feature when you are trying to wind a coil to this level of precision. Otherwise, if winding coils for speakers or transformers or other uses, the automatic traverse mechanism works great and lets you make coils of the size and number of layers you desire. I would use it to make mono coils sometimes. But, honestly, for small coils, it is unnecessary. Coils want to form themselves naturally.

As you can see, there is a turn counter. And, yes I use magnification. At first, I was just using a Jeweler's Loop. That was horrible. It gave me headaches squinting into a loop and it was extremely hard to get things working like that... I made a system of magnification. Can't explain it because it is proprietary. Needless to say, I can comfortably wind coils without straining my neck or squinting into anything.

I turn the coil by hand. I align each turn as it is formed. I can do up to 5 layers with the same precision you see in the image above. Mine would look even better if I were rolling onto machined bobbins. Since the plastic isn't perfectly flat or round and there are no true right angles, that makes doing what I'm doing challenging. When I move to metal parts, I will be able to whip out coils like nobody's business. Maybe 10 minutes per. Currently, I can achieve what you see here in about 20 minutes.

While winding I use a tool in my left hand. I've done so many that I can effortlessly operate chop sticks left handed now. My left hand is now about as dexterous as my right. Nice little perk.

[opcode66]

Nice coils, Todd!

Thanks! I've worked very very very hard to get here.

A computer controlled coil winding mechanism is totally do-able and a great option. It will be possible to get great results with it. My atlas/craftsman 618 metal lathe has coil winding possibilities that are described in the user manual, using the threading gears setup with pitch=wire diameter. Direction has to be manually changed.

Something like this is really nice if you have $1500. You can find older ones not computer controlled for $500. But, again, all is unnecessary. For small coils, it is total overkill. Also, for someone who isn't looking to make zillions of coils, a manual winder like the one I posted above is just fine.

http://www.ebay.com/itm/CNC-Automatic-Coil-Winder-Winding-Machine-/261827968489?hash=item3cf629bde9

s-l500.jpg

If anyone wants to wind flat-wire coils but doesn't want to flatten the wire themselves, various distributers carry stock of that sort.

While that is true, if you want to do precision coils like mine, you have to have wire that is flat and a consistent precision width. If not, then you can't pack the same number of turns for a given space for each layer. For instance, in an SX, you need to be able to make extremely accurate measurements to dial in a mill to create wire that is precisely dimensioned. No one doing SX repairs are buying off the shelf flat wire. It is too inconsistent. And, has other issues like little kinks.

I can certainly mill some for anyone who may want for a nominal fee.

[opcode66]

So what is the real advantage of using flat coil wire? Flat means more layers in the same space, but fewer wraps per layer. But I guess it increases density overall by eliminating dead space from around normal round wire and reduces mass as a result? I assume this is really only needed for for drive coils, since feedback wire can be so much smaller due to the super low currents involved. Right?

b

Flat wire packs more tightly than round. No wasted space. As has been stated here before, try to pack a bunch pencils together, then try to pack a bunch of cd cases together. Which is more tightly packed??? Squared objects of course. Why does this matter? Well, a flat wire coil produces a much more efficient EM field when compared to its round analog. When you pack the wire tightly and very regularly together, you create a very focused and efficient EM field. When a coils turns are not so regularly aligned the field it creates is more scattered than the same coil made in flat.

Why fewer wraps per layer?? Flattening a round object doesn't necessarily result in something much wider than the original round shape. You are simply compacting vertically. If you dial the pressure up then of course it comes out wider and also unusable since the enamel will be compromised and it will short. But, light pressure on the correct gauge give you a very nice flat wire that packs as the same number of turns for a given space as its round counterpart.

One big point that you don't see yet is that when you wind round wire and try to do multiple layers on a coil with round wire, the second layer will start to look bad, the next worse, etc. Why, well, round doesn't like to perfectly align itself over round... If you are doing multiple layers and precision is a goal, flat is the only way to get to the promise land.

Just like every phono cart ever made, pickup and feedback coils are made from the smallest finest wire you can deal with. SX feedback wire is so small I can't even find it. No one sources it. Good luck using a powered machine to wind those.... It is thinner than hair and snaps very easily.

Hope this all helps. I posted some more information here that is useful to anyone working with coils. Of course, I've kept a few things to myself. But, this is a lot of good experiential wisdom.
http://lathetrolls.com/viewtopic.php?f=1&t=5567#p37920

[JekaTallbot]

opcode66 wrote:
when you buy my eBook on cutterhead design


I would like to buy one as well. Do You prefer Paypal?

Greetings from Berlin.

[opcode66]

I was serious about writing one once I've completed my work.

[EpicenterBryan]

One big point that you don't see yet is that when you wind round wire and try to do multiple layers on a coil with round wire, the second layer will start to look bad, the next worse, etc. Why, well, round doesn't like to perfectly align itself over round...

OK, I get it now that I've tried it. It's harder to do than I thought. Your coils do look awesome.
I have some pretty OK first results tonight with my DIY winder that I want to share with folks...

So here is the mechanical section of my version of a DIY coil winder. Damn sexy if I must say so myself!

IMG_4067.JPG

Mechanically, it's simple enough. A stepper with a chuck attached to hold a rod for winding the coil, two V wheels to guide the wire, and an X stage with lead-screw to move the winding motor and rod (or bobbin) right and left relative to the V guides.

It's being run by an Arduino and a version of Marlin 3D printer Firmware that I modified for use with a Plasma cutter. Basically, it's just being used as a Gcode interpreter to run the stepper motors. And "Repetier-Host" is being used on a netbook PC to send the Gcode to the Arduino.

I only finished building this gizmo last night and this is the first run... But... Obviously Todd is right that the wire wants to go where it wants to go. The V guide for one is too far away in tonight's version for sure. Todd said it would happen, and yes, Todd was right! So I'll be moving the guide closer, and will likely put in a small diameter hollow guide tube (adjustable) to get the exit closer to the coil. I also think if it's a big deal in the next round, I may be able to do a quick Gcode move to force the wire to jump over where it wants to bunch up on the ends.

IMG_4068.JPG

Let me show you the end result of the first test. This was wound on a hollow core with temporary (removable) washers as bobbin ends for winding. When the process and mechanism gets tweaked, I think I would apply glue during winding to keep the coil winding in place. The idea is to have the winding on a core that could slide off the winder and be glued in place on a push rod (link) when assembling...

IMG_4069.JPG

Yes, this is not pretty. But it demonstrates it can be done. Again, my eyes are not as good as they were so doing this by hand is not an option for me.

Now we get to the cool things in this post...

I mentioned a link the other day where you can enter parameters and figure out how many coils are needed and such... But not everyone knows what Gcode is or how to write it to make steppers move even if you have the gear... So here we go.

I found a pretty cool utility called Gcoil. It's part of a set of useful Gcode utilities called "Gsuite_V1.4" and it's free. Shown here on the left is the program, and on the right is the Gcode output. This output is what I send to my winder via the 3D printer host software. Now it also has the ability to isert a machine pause (which you can always do manually if you know Gcode) so glue I could be applied between layers. But I think when I get this nailed down I'll more than likely dispense glue on the wire as it is wound.

Gcoil program and output.jpg

To make this all work you have to tweak your Arduino firmware a bit. In my case Y is being used as the rotary (spindle) coil winding stage. I'm running that stage in 1/16 microstep so it takes 32000 steps to make that motor turn 1 rev. My firmware thinks all units are in mm, so if I want the stage to move 1 turn - I tell it to move 1 unit (normally that would be 1mm). Notice in the output file 50,100,150,200. Those would normally be movements in mm, but the machine is doing those in rotations (coils). Get it?

Also, the X stage is using a 4 start thread lead screw that moves 8mm for each rotation. This is very coarse but that's what I had this weekend and is OK. So using 1/16th micro-step, that again works out to 3200 steps but now for 8mm of travel. So 400 steps represents 1mm of travel. So when the wiresize is 0.1mm, that represents 40 step pulses to the motor drive.

I'm pretty happy with this first test of my coil winder for sure... But Todd was totally right. It's not trivial to wind nice coils. Todd's coils are so nice!

More updates on Groove Scribe in the next week...

Bryan

[EpicenterBryan]

I had some really great results yesterday with the coil winder. Before I get into that, I want to show you guys a few things.

I'm using salvaged wire from AC wall transforms right now while I'm trying things out. Everyone has some of those sitting around and it's pretty easy to get wire for experimenting. Here is how I did it...

Start with small output current ones. They will have the smallest wire in them. The metal core is made of many alternating plates and have some lacquer coating them. Hit the plates with a hammer until they become angled. That's when the lacquer has loosened and you can start taking it apart.

wall transformer.jpg

Removing the first plate is the hardest. Get under the two sides with a box cutter blade, then a screwdriver and bend them up. Then pull on the center end with pliers or a pair of wire cutters. Once the first one is out, pull the second one from the other end. Sometimes the lacquer is still bonding in the center. Just push your box cutter in there to loosen.

take apart.jpg

I'm using 0.1mm wire right now (38 awg) and I found that and many other sizes in the transformers I took apart. Everything from 0.03mm to 0.61mm. I have since ordered and received new 0.1mm wire and will start using that as soon as I'm done experimenting, but having as stash of various sizes for experimenting has been really great!

wire sizes.jpg

On thing to note is that this wire has a coating on it before it's wound, but then they coat the entire transformer in more lacquer so when you unwind the wire it does have some surface irregularities but you can live with it for experimenting I'm sure.

I did take one apart that had a weld bonding all the plates together. If you run across one of those toss it.

Bryan

[EpicenterBryan]

Here is an update on the coil winder.
I made two changes that really impacted the quality of the winding.

First, rather than spool salvaged wire off the transformer and all over the floor, I put it on a spool, added bearings for the spool holder and a brake of sorts on the bearings so the wire stayed in tension. Then I used some spare parts for my 3D printer and added a 0.4mm nozzle, pickup tube and mount block for the wire to go through as it's spooling onto the winder.

IMG_4075.JPG

Here is a 500 turn coil (10 layers - 50 turns per) as the wind completed. Much better - and with salvaged transformer wire. I also applied glue as it was winding.

Image5.jpg

Here is what a 200 turn coil (4 layers) looked like:

200 coils.jpg

And the 500 turn coil:

500 coils.jpg

CONTINUED ON NEXT POST:

[EpicenterBryan]

And here are the DC resistances for the two coils:

Coil resistances.jpg

Bryan

[opcode66]

Tension is critical to making well aligned coils.

I find that gluing the finalized coil works great. Gluing while rolling is a pain and not particularly necessary for small coils.

My guess is that you'll be very pleased with 8 Ohms or less. Most amplifiers only want to drive up to 16 Ohms DC resistance before having issues.

I thought that looked like a .4mm nozzle and a heater block assembly.

You can still hand roll a finer coil. But, much better than the first ones. Keep at it.

[EpicenterBryan]

My guess is that you'll be very pleased with 8 Ohms or less.

Those coils are for feedback. I'm still going to use the stock voice coils, just mount them onto my carbon fiber disc / nylon spring.

[opcode66]

The diameter of the wire is a little big for feedback coils. Just say'n.

[Ciuens]

The wire that I used in my feedback coil is 35.
The impedance is very low, only 2ohms. I need thinnest wire to increase the resistance of the feedback coil for at least 8ohms.

Ciuens