I've gone off the deep end: working on a filament winder

[Scrapmaster87]

The last 2 pictures attached are actually my 1st two tube rolling setups from 2019. The cardboard box one was retired to painting use and got thrown out with the move, I took the picture to show how far I've come.

 

[Scrapmaster87]

And here's the CAD work from V3.

Note: the dashed circle on the headstock represents a maximum 16" mandrel diameter.

 

[REK]

Well done, that looks great. Looking forward to seeing the end result.

I’m planning on getting rid of my X-Winder and making a new one that uses servo motors with closed loop feedback. NEMA 34 size too, should make for a better winder and better results.

By the way eventually the winder did it again where it skipped the positioning, pretty much, it is time to scrap this winder.

 

[Scrapmaster87]

Well done, that looks great. Looking forward to seeing the end result.

I’m planning on getting rid of my X-Winder and making a new one that uses servo motors with closed loop feedback. NEMA 34 size too, should make for a better winder and better results.

By the way eventually the winder did it again where it skipped the positioning, pretty much, it is time to scrap this winder.

Servos are delightful devices, but on the expensive side. I'm only using on on my x-axis because I have an integrated one in my box of industrial stuff (it was a training aid). Otherwise I would have gone with a closed loop stepper. For my mandrel axis, a geared closed loop NEMA 34 would do the job. But I want a massive speed range so I can do roll wrapping and polishing without removing the mandrel. Right now a 1.5kw servo looks to be the path. The only problem is that the inexpensive servo drives need 240V 1ph at minimum. A 120V compatible drive-motor set more than doubles the price!

 

[REK]

Servos are delightful devices, but on the expensive side. I'm only using on on my x-axis because I have an integrated in my box of industrial stuff (it was a training aid). Otherwise I would have gone with a closed loop stepper. For my mandrel axis, a geared closed loop NEMA 34 would do the job. But I want a massive speed range so I can do roll wrapping and polishing without removing the mandrel. Right now a 1.5kw servo looks to be the path. The only problem is that the inexpensive servo drives need 240V 1ph at minimum. A 120V compatible drive-motor set more than doubles the price!

Yes those can be pricey, the servos I am going with are from Teknic’s Clearpath Servo motors. They have one that gets 1.36Kw at 110V 1 phase for $998.

Don’t know if that is pricey for you, but I think I will stick with 2 NEMA 34 stepper motors that will at least handle a 6” mandrel at decent winding speeds for faster winding.

X-Winder is just too slow >_>, especially for larger diameters.

 

[Scrapmaster87]

Yes those can be pricey, the servos I am going with are from Teknic’s Clearpath Servo motors. They have one that gets 1.36Kw at 110V 1 phase for $998.

Don’t know if that is pricey for you, but I think I will stick with 2 NEMA 34 stepper motors that will at least handle a 6” mandrel at decent winding speeds for faster winding.

X-Winder is just too slow >_>, especially for larger diameters.

I'm drooling over the Automation Direct sure servo 2s. That's where I can get 2.0kw off of 120V 1ph for $1,800 (drive, motor, and cables separate) or 1.5kw for $1300.

However being as cheap as I am, stepper online recently came out with their T6 line of 400/750/1000 watt servos and I'm seeing even cheaper systems off ebay that go up to 1500W. Both of these however need 220V 1ph power....which either means a step-up transformer or running a new 220 line in my basement...

 

[G_T]

If you run the speed too great, then the variations in speed of filament feed might cause you tensioning issues. You want to keep the filament tension under control. Ideal is NOT constant tension, due to fiber elasticity layer compression and winding angle, but the best you'll likely be able to do is to approach constant tension.

Gerald

 

[Scrapmaster87]

If you run the speed too great, then the variations in speed of filament feed might cause you tensioning issues. You want to keep the filament tension under control. Ideal is NOT constant tension, due to fiber elasticity layer compression and winding angle, but the best you'll likely be able to do is to approach constant tension.

Gerald

I'm looking at topping out at 1440 mm/sec (56.7 in/sec) as a max traverse speed on my x-axis and keeping my tension-on-carriage architecture. Since the carriage on V4 will be the most complicated, that's what I'm working on right now. From what I've learned on V2, I'll be doing a very light pre-tension off of the "creel".

I'm happy with how I have the carriage drive, but I'm on my 3rd approach for the B-axis pivot... The challenge is that I want something that breaks down fast for cleaning.

 

[Scrapmaster87]

Think I got it, my computer hates me though... Time to start on the combs and tensioners!

 

[Scrapmaster87]

Still hammering away at this. The cable-chain I designed last year and actually have enough printed for V4 (I think). I was really hoping to keep my tension-wheel concept, but I think I'm stuck with straight friction since all (up to 5) tows have to be adjacent. I have the concept already, I just need to start modeling it. There's certainly some fat trimming to occur, but I need to cover my more critical design requirements first.

 

[Scrapmaster87]

Just about done with the carriage, I'm working out the final fastener details. I have to add a support bar before the alignment pulleys and some final touches on the final pulley mount.



Part are also starting to trickle in. The big pivot bearings and the alignment pulleys just came in. Thursday the cart and misc hardware comes in. The servo, gearbox, stepper, and both brackets I already have. I also just got a new spool of red PETG and gave the printer a hot-end fan upgrade since the old one was starting to die.

 

[Scrapmaster87]

So things are coming together! Carriage parts are getting a final design polish before going off to print, which so far is going exceptionally smoothly.



In other news, I scored an actual electrical enclosure from the scrap pile at work (with permission of course!). I've got a lot of cleaning to do, but going to 220V means getting a zap is beyond unpleasant. I will be doing a proper installation with the AC voltage stuff.


While I was touring the boneyard, I saw the biggest sonotube I've ever seen! I think winder V5 (or 6?) has its first design parameters.


Back onto V4 here, I have the servo drive set on order. It's a 1.5kW Chinese set, but ordered from a us-based seller on ebay. Keep in mind, 1.5kW is 2 horsepower, this is some serious horsepower in the house! This particular motor set is most often used by people converting their lathes and vertical mills to servo spindles. I'm going to try to only have a single speed belt reduction since the RPM range allows this. My only concern is keeping sufficient torque authority over larger mandrels, testing will show this. My 4" aluminum pipe mandrel should be a good torture test. I'll certainly need the horsepower if i'm going to try to polish out the lumps that prevent it from being usable. If I don't have enough authority, I'm dusting off V3's belt sprocket concept.

I may design up some toothed rail end caps to secure the stationary x-axis belt. The open rail system looks like they typically mash the belt into one of the t-slots with a screw. I want something much more robust however.

The next major component is the head-stock/stand. I may or may not reuse V3's stand, the wood electrical enclosure obviously has to go. The x-axis rail needs to quick detach since I don't have a huge shop (yet!). Since everything should break down nicely for storage, the whole winder should be easily transported via a small SUV with the exception of the 7'10" rail I plan on using!

 

[G_T]

If you are working with that much HP, you might want to design in a slip point so the mechanism disengages before breaking the mechanism, or seriously harming the operator, should something go wrong. The filament won't be breaking. Getting wrapped up in it would not be good. Unlikely, yes, but very bad.

Gerald

 

[Scrapmaster87]

If you are working with that much HP, you might want to design in a slip point so the mechanism disengages before breaking the mechanism, or seriously harming the operator, should something go wrong. The filament won't be breaking. Getting wrapped up in it would not be good. Unlikely, yes, but very bad.

Gerald

Yup, I've been thinking about safety for a bit now. My first thought was a dead-man hand switch, but I need 2 hands when I'm mixing additional batches of epoxy. I'm absolutely putting an e-stop on this to kill it when they'res a problem. Maybe a treadmill key would work nicely?

On my lathe I had a foot pedal to run it. I only ever ran it hand and foot free when I was curing skim-coats. V4 will be getting a pedal too, but only for running as a lathe.

I do plan on keeping the timing belt drive. I'll have to see what torque it cams-out at with only gravity tensioning like on V3. I can see building a detent-clutch into my disk couplings.

I absolutely understand what can happen with the filament. V2 had a little issue with glass fibers breaking and getting wrapped up in the guide eye, hence why v4 has all 1/2 lor larger rollers and bars for the filament to pass over.

 

[Scrapmaster87]

Stinking later shift.


At least with this one I can pretend it came out ok.

 

[rocket_troy]

On my lathe I had a foot pedal to run it. I only ever ran it hand and foot free when I was curing skim-coats. V4 will be getting a pedal too, but only for running as a lathe.

Our industrial scale centre lathe at work has a foot pedal, but it's the Estop. Runs the entire length of the bed and it's by far the best EStop I've ever used on a machine.
It's probably just me, but I really struggle to find the EStop button on a machine quickly when something goes wrong. My mind kinda goes into a momentary freeze for what's happening and which button to push.
With the foot pedal, it was just a natural instinctive reaction (like hitting the anchors in a car).

TP

 

[Scrapmaster87]

Safety-wise I think I've settled on a pull-cord that I'll suspend over the main rail....

I have the get the control architecture settled before I can get much further as the LCD screen and arduino will most likely be mounted to the main rail.

Wiring will be fun since I have:
220VAC split-phase coming in for the mandrel servo
120VAC single-phase going to a 48VDC power supply for the rail servo
120VAC single-phase going to a 12VDC power supply for the arduino and LCD screen
120VAC single-phase going to a 24VDC power supply for a PLC running the manual-mode logic and e-stop circuit
120VAC single-phase POSSIBLY going to another 24VDC power supply for a dedicated stepper drive for the filament head (I might just use a 12V step-stick on the arduino's shield though)

In other news:
I've ordered the 2.4M rail for the linear axis
I've mounted the 1.5KW servo motor with new timing belt pulleys (we'll see how well mt PETG motor bracket holds up)
And I was gifted some high-tensile polyester and kevlar bobbins from work. All 1k filaments so the multi-filament design is a must to use these.
Demolition of winder V3 is also underway to reconfigure to the V4 design

 

[Scrapmaster87]

Some more pictures of "stuff".

 

[bzaidi]

Oops: www.propulsionlabs.com.au/filament_winder/Winder_Sim.zip

Too many forward slashes. Copy & paste into your browser's address bar if it doesn't work now.

It's actually not designed around small diameter mandrels or very thin tow which is something I didn't think about when mentioning the link but I suppose you could scale everything up by 10 if need be.

TP

Hi rocket_troy, i have downloaded above file, unzipped and tried to run the Setup.exe in Windows 11 (release 22H2). It does not get installed. Which OS has it been tested to be installed and working.

 

[Dcsilvias13]

Awesome build! Wondering if your gcode generator is open source?

 

[Scrapmaster87]

Awesome build! Wondering if your gcode generator is open source?

My code generator is not open source, it's just a spreadsheet of configurable repeating patterns that comments itself out when the sequence is complete. I'm not a programmer, so my generator is brutish to say the least. Also, my way may not be the right way, but I got it it to work for me. I don't want to pass something out may not work for someone else

If there's enough of us fiber-slingers wanting to support a code generator (or other things winding-related), then I suggest starting another thread dedicated to it. I'll certainly contribute and give ideas! Otherwise PM me and I'll show you how to calculate your own!

This thread has resulted in a small number of people coming to me with specific questions/ problems. I'm happy to supply the information for someone to build their own generator or hand code a sequence. So far I gather there are maybe a dozen of us doing this, half doing so through a university?



Its been a while since my last update, the holidays have certainly taken up some of my spare time. I'm also still dealing with the fallout from moving back in May. Finally, the servo drive won't fit in the panel I have. Since I need to source a bigger panel, I'm taking the time to actually engineer the controls vs. hacking them together. I've been hashing the value-engineering and safety in the controls since I'm out of freebies and actually have to buy some industrial components.

The x-axis is about 90% complete, just needing flex-cables, connectors, and safety-cord mounts finished. The head-stock needs the wooden panels sides trimmed down and the control panel mounted, as well as a vertically-adjustable mount for the x-axis. The tail-stock is just a pile of parts and wood leaning up against the wall.

 

[hslggreen]

Did my first real wind yesterday! Surface is just a little lumpy, the tow is not perfectly even coming out of the spool.

The tube here is for a 38mm motor mount and is 3 layers thick. My mandrel was a 1.5" aluminum tube and the OD is measuring 1.558"-1.569". The wind was finished with a 50% overlap of Dunstone shrink tape. I have to sand off the excess resin on the surface yet, but I'm very happy with the result and ease of the wind.

Tensioning out of the center pull spool still needs some work and I could use a length of filament calculation to estimate resin usage.

Next up is a 29mm motor mount and some 5/8" guide tubes for a 2-stage!View attachment 462508View attachment 462509

Can you make the program for that winder available, I am in a Rocketry team and having a rough time coding the program for our winder

 

[Scrapmaster87]

Update time!

First off, apologies for such slow progress. Last May we moved to a bigger house and I took till June getting the old house ready. From there there's been quite a bit of settling in and taking care of things around the house. There's also been another development which I'll get to the end of this.

V4 is alive as of today! The headstock have been mechanically equal to V3 for some time, but V4 is now electrically and developmentally further along!

I'm not done wiring the controls or installing the Marlin LCD yet. But I did install the fuses, confirm voltages, and put the servo drive in test. It just hums along at 300 RPM with a gentile bobble from the printed pulleys. Max RPM is 2,500, which I can only describe as industrially intense. Max speed is for hand finishing in-place and cleaning mandrels. Being a 1,500W servo, torque is effectively unlimited. I set the speed to 1 RPM and I cannot stop the big pulley by hand at all! This thing maintains RPM/position with authority! There will be a proper e-stop setup on this, I have a pull-cord stop that will be working in conjunction with the mushroom button on the panel door.


Anyways, here's the pictures!































Oh and for that other reason I've been taking so long: my wife and I have been getting ready for an "assistant" that will be joining us at the end of this month. "She" of course will need several years of training, but I'll do my best to get her indoctrinated.

 

[G_T]

From my observations, I think the training goes in the other direction!

Gerald

 

[tfish]

@Scrapmaster87
Assistant....funny
My old assistants moved along awhile back..they are letting me train the next batch..8 now.


Keep up the good work!

Tony