Filament winder one

A filament winder project.

23'rd December 2008,

Composite pipe and pressure vessels are usually made by winding a continuous fibre, tape or fibre set around a rotating mandril.
This project has two or three main sections. There is a spindle motor, a carriage assembly and possibly a kreal assembly to supply the fibre for winding.
My aim is to make a winder which can make a vessel several metres long and a metre in diameter. After testing I only plan to use it once. The winder can be quite slow. It could take a week to do the job.

Initial tests will be on a small scale but the winder should be able to scale up easily.
It is not clear at this stage whether the carriage motion needs to be synchronised to the spindle rotation. Ideally it would be synchronised but this may be overkill for the task at hand.

The tests will also tell me whether to wind wet or dry (and vacuum infuse). It will also give me a better idea of the number of wraps requiresd and rate of resin usage.

The spindle.

I need to be able to rotate something like 50Kg at five to ten RPM. I'm guessing the job will need 20K winds so it will take at least 2K minutes (about 33 hours). If all goes well this only happens once so it isn't worth wasting a lot of time making it go faster.
A slow spindle speed also allows a slower carriage traverse.

I need a high ratio gearbox so I bought a cheap (au$70) electric boat winch for its 153:1 planetary gears. The motor shaft(top right) goes through the bearing in the housing (bit leftish) and becomes the "sun" gear which drives the three planet gears (bot right). The planet gears are engaged in both the gear teeth moulded into the housing and the large gear they are in for the photo shoot. These gears will have a slightly different number of teeth (different by 3?). Not visible is the spline on the main gear (hidden by the planets) which take the spline which drives the winch spool. The spline in the spool retracts. The shaft,spline and planets are steel but the large gear and housing appear to be an aluminium alloy (not sure) - wear could be a problem.

I was planning on using a 100 Watt 24V wheel-chair motor but the large matched toothed gear was badly cracked. I attempted to cast a duplicate but failed. A 200 Watt 24V electric bike motor directly coupled using some garden hose seems to work well at 12V. I may add some extra gearing when building the larger winder.

Before attempting any automation I manually wound a test piece. I'm using PET soft drink bottles as mandrels for my small tests. I had a lot of trouble using fibre off the outside of the roll. Pulling it from the centre is easy but you then have fibres with a twist - I'm not sure how important it is. I believe my fibre is intended for a chopper gun and is slightly different to the rolls made for filament winding.

Most winders use a wet fibre but I'm winding dry and epoxy infusing later.

The infusion went fairly well but there were some issues with the transfer layer (shade cloth) and peal-ply wrinkling. This is mostly cosmetic (later - wrong! - there is a serious problem here see below) and the tube seems quite strong. I haven't done many vacuum infused tubes so I have some things to learn.

It is quite possible I will use a wet fibre technique in the end but for now at least I will wind dry.

The carriage.


My prefernce for driving the carriage would be a DC servo with optical feedback but this is a lot of work for a one off winding job. Instead I use a 200 step/rev stepper motor I bought some 20 years ago. To drive the motor I build a circuit based on a L6208 controller driver. The circuit is pretty much straight out of the data sheet.
This IC works just fine. It features a chopper drive, I'm running the 5 volt motor off a chopped 25V supply.
The drive is bi-polar so I have the choice of using all or half of the centre taped unipolar winding. Using 1/2 the wind does allow the motor to run a little faster.
The motor would run OK at 1000 steps/second but the torque was low.

To manage the driver board I use a SAM7 based MT-256 which is also use in other projects of this site. The code is written in SAM7 FORTH and accepts commands for a serial link and supplies to magic pulses to the controller. The code is interrupt driven so the high level program can issue a position seek command and do something else while the motor is stepping. Because I'm only stepping at slow rates the software does not need to ramp up or ramp down the step rate at the start and end of a movement.

My prefence for moving the carriage would be a ball chain but gears to drive ball chains usually need to be hand made so I tried using a kite string instead. I machined a pully with two groves from some old set epoxy and despite the groves being a bit too shallow it works just fine.
It appear the string will last long enough to do the job. Replacing it is only a ten minute job so it isn't a big deal to change in once in a while.

The carriage is simply hand moulded polymorph around a 25mm aluminium square tube. The carriage has a hole on the under to pass fibre. Obviously you want it on the underside so you don't drip resin on the tube if you use wet fibre.

One unexpected problem was the string would vibrate with a deflection of several centimetres - adding some (green) tack to the string calmed it down.

Before doing the first wind I held a felt pen on the carriage and ran the winder so I could see the wind angle I could expect and adjust the step rate to suit. The step rate for this small job is only 40 steps/sec.

For the first automated dry wind I had no tensioning mechanism. It became clear I needed something so I grabbed a phonebook and ran the fibre through the pages to apply some tension - this worked to some extent.
I vacuum infused the glass and had a serious problem with a large crease forming. It seemed clear the fibre was too loose and when vacuum was applied the the reduction in diameter meant the execess fibre had nowhere to go and buckled.


OK, I needed a tensioner.

I made tensioner from a mouse trap. Once again good old CAPA (polymorph) came the the rescue to make a fibre guide.

The improvement in winding tension was obvious.

I put the wound piece under vacuum and it seemed to be good. Unfortunately when I set it up for infusion it creased. It didn't crease as badly as the first one but enough that I abandoned it.
If think with enough tension that infusion would work but my stepper isn't strong enough to increase it much more and with too much tension the mandril may crush - probably no problem when using a drink bottle but the mandril for the big job may not be strong enough.


So how hard is it to make a resin bath? It turns out to be quite simple.

While on holidays at Woodford I had the idea that a 2 litre milk bottle had potential for a quick and dirty, cheap disposible resin bath. I bought 2 litres of apple juice from the festival shop and once emptied literally hacked it with my gerber supertool/knife.
This was similar to the photo above but minus the syringe (added to mack2).

The fibre passes through the handle which is the lowest point in the bath. It is slight inclined so excess resin runs back to this area.

Version one of the bath used the folded bottle itself to squeege the fibre of excess resin. I concidered using this for tensioning as well but initially kept the mousetrap contraption in place - this works.

The first wet wind worked pretty well but it was too wet. For V2 I repleced the clothspeg idea with a wiper made by poking a hole in the rubber bung off a syringe plunger - this works perfectly. You can see in this (above) photo the fibre drags a lot of resin with it and the is quite a stream running back down the hill into the pool at the bottom. The cured result looks very good to my eye.

Here the fibre reaches it final destination and is wound into place. I've choosen to wind at a little over 45 degrees and the big job maybe wound at higher angle - say 60 deg.

The first two wet wound tube look pretty good - there are a few voids in the walls but these don't really matter. There will be a milimeter or so of hand laid fibre-glass under the winding so it will be airtight. The voids are so small they will not blow out at the pressure used.
My feeling is some small defects actually improve safety because should we ever experience severe over pressure these will pop and prevent the tube from bursting.


It is time to move up from 1/10 scale to 3/8 scale - now for a bigger bottle.


Image Autonomous resin dispenser in action.

Winder one has completed its mission and will soon be disassembled and recycled into winder2.


A short clip of the winder.
wet1.mpg (4.51 MB)
Invalid YouTube URL provided
A longer clip of the whole setup.
longwind.mpg (8.99 MB)
Invalid YouTube URL provided
view the gallery


Created by eddie. Last Modification: Sunday 08 of January, 2012 16:07:59 AEDT by eddie.

Main Index

Switch Theme


eddie, 11:29 AEST, Sat 22 of Jun, 2024: Offline wiki repair is mostly done. It is running on tiki-27-alpha
eddie, 18:14 AEST, Sun 12 of May, 2024: Phase one of wiki repair is to upgrade to V22 (the last version with image galleries). All 150 images in the wiki-up directory are broken and need manual wiki page repairs.
admin2, 19:45 AEST, Sun 28 of Apr, 2024: Tested tiki 26.2 on a raspberry pi. Tiki 26 does not support image galleries so I have to try again after migrating images to file galleries.
eddie, 13:04 AEST, Thu 10 of Aug, 2023: Offline tiki 26 upgrade went badly. Waiting for 26.1. Will limp on.
System Administrator, 18:45 AEST, Wed 26 of Jul, 2023: Recovered from lockout but unable to upgrade to V24

Last-Visited Pages

Online Users

13 online users