Anyone have experience filament winding?

General Composites Discussions Composites Talk
#1

Hi All,

I was recently put onto a program at my work that is in very bad shape (high scrap rates, etc) as the process engineer. The old filament winding machine that we primarily use to make about 20 different parts has been giving the company trouble for the past 6 months. The lead technician who ran the machine for 20 years left the company a few months before I came on to the program and essentially left no knowledge base behind. The engineer who was assigned to program also recently left (before I was assigned to the program), leaving some documentation but not really enough for me to learn the ins-and-outs of the machine or process. So basically I’ve been reverse engineering the program for a while, but I’m having trouble really finding the nuances fast enough to save the program.

I’ve been trying my best to problem solve (and document) the process but am hitting a point where we have a large backlog of parts, a faulty machine, and really no one who has experience with winding at the company. My troubleshooting hasn’t been effective and I really can’t find any good resources or consultants in the area to at least give me a baseline set of experience.

Does anyone have any tips for diagnosing filament winding machines, or have any good resources that I can turn to? It seems like payout design, speeds, feeds, gapping solutions, etc are tricks of the trade sort of items. We only wind tow-pregs/prepreg roving and generally small diameter parts.

Any help would be greatly appreciated!

Matt

#2

Hi Matt,
Yes you’re right they are normally tricks of the trade. Is it a consultant you’re looking to (cant help you there sorry) for more extensive work or is there a specific problem that’s really troubling you? Maybe we could start looking at the worst problem here?
I can imagine how you feel when all the years tricks are locked in someone’s head who has left - Im actually just starting some placement students in writing SOPs this morning to avoid just this problem in another machine?

#3

I’ve found that you can often lean on your suppliers for tech support. I deal with Composites One for a lot of materials and they have always been able to find someone in their organization or among their vendors who can answer questions or provide tech support. Also, you may get help from the equipment mfg. (if you haven’t already tried that).

#4

Fastest and probably cheapest option is for the company to hire the technician or engineer back as consultants at ~3x rate of what they were making.

And then they need to learn to carefully document things so this doesn’t happen again. They’ll most likely find that this tough love creates a better product at a lower cost.

#5

Thanks for the replies. It’s a tough situation losing someone vital to a process and I’m not really sure how they let him walk. The old tech moved away so it’s not very feasible to bring him in as consulting help, but may be the end solution; I’ve been told to pursue all other avenues before attempting that one - although I agree that’d be the best way. I’ve queried the engineer for help a lot to fix a few problems, but he had limited experience himself (~a year).

I’ve been talking to the machine supplier but they’re an English company and have no real service techs in the US so an on site visit would be costly. We may have to go that route, but again, I’ve been asked to pursue other avenues first.

Our main issue is gapping between bandwidths or tows. The way that the machine is laying the material onto the mandrel, whether it be the program, the payout, or the actual machine is causing the material to gap. We have to hold what I’ve found may be a tight tolerance for winding on these gaps (.020"-.030" width max) or else spur rejection. We lost an NC program and have had numerous troubles regenerating it to what it once was (no backup - I know, a lot of bad managing going on here). We are forced to run the machine quite slow for a winding process and with lots of operator involvement moving the plies. I am trying to get away from that and make it a more fail-proof, automated process the way I picture it. Any ideas from Vets? I’ve narrowed it down to a poorly maintained machine (expensive to fix), bad NC program (though it was spit out by the trusted NC generator, but there are tweaks to be made apparently that I can’t figure out), or a bad payout design. I’d like to tackle the list in reverse order, but again, resources are lacking and the company is reluctant to bring in consultants (although I’ve asked).