# Filament winder algorithm

G

#### Geo Ernsperger

I'm looking for a control algorithm for a fibre glass filament winder, to control mandrel dc drive and carriage motor (brushless servo) using a pc based control and programming in c++.

A

#### Anthony Kerstens

I don't know about glass, but we have uncoilers
for 30-40,000lb steel coils. The controls use PID
where the D implements 1st, 2nd, and 3rd order
derivatives.

If I understand it correctly:
1st derivative is the error rate of change.
2nd derivative is the acceleration of the error.
3rd derivative is the jerk.
Tuning is done in the same way as 1st order PID.

Jerk on a steel coil is easy to imagine with a steel coil, given the rotational momentum of 30,000lbs.
For glass fibre, it might be important to avoid plastic deformation and breakage. If there is slack in the fibre, it could prevent run-away rolling that would cause
damage upon a sudden increase in tension.

That said, I would recommend researching some
advanced control texts. You have to consider all
the parameters of the physical situation.

Anthony Kerstens P.Eng.

M

#### Mark Bayern

> I'm looking for a control algorithm for a fibre glass filament winder, to control mandrel dc drive and carriage motor (brushless servo) using a pc based control and programming in c++.

To simply do some hoop (circumferential) winding or helical winding you can adjust the bandwidth, wind angle, or cylinder length slightly to make the wind fit the given parameters. I have yet to find a good algorithm for winding heads onto cylinders for a two-axis winder. If you (or anyone else) finds some good resources, please post them here!

Mark

T

#### Tim Sevener

In order to wind domes on a cylinder with a 2 axis
winder you need to lie about its length and over shoot the ends. The dwell gets adjusted to suit