Sammy Arschavir

Dip Moulding PVC




This week I visited Plastic Mouldings Ltd, they offer a variety of process such as rotational moulding, dip moulding and vacuum forming. A large portion of their factory is dedicated to dip moulding PVC.

To dip form PVC you heat a tool (usually aluminium) to somewhere around 170 °C in an oven then dip it in a vat of liquid PVC. The PVC begins to form around the tool with the heat, so the longer you leave it on, the thicker the coating, it is then cured in another oven.

I think it would be tempting to dip it in one colour of PVC, cure it, then dip it in another colour of PVC, do this many times with many different colours so you’d have all sorts of colourful layers building up but you’d only have one colour showing on the exterior. It’d be satisfying to cut it open and see all the layers. I’ve no idea what I’d use it for. Maybe I’d leave it as lots of layers and use it in an application that would wear through the layers with time, like a cutting mat or a pin board.

It might also be interesting to use various materials as the tool, they’d each have a unique thermal conductivity, so as long as the tool is not left in the oven long enough for it to reach a uniform 170 °C, then each part will be at a different temperature and the volume of the PVC on the different parts of the tool would reflect the various temperatures. The results might not be so straight forward to interpret since each part would cool approximately following Newton’s law of cooling:

Q ˙ = m c p T ˙ = h A Δ T

Where Q dot is thermal power (W), m is the mass (kg), cp is specific heat capacity (kJ/kgK), h is the heat transfer coefficient (W/m2 K), A is surface area and ∆T is change in temperature (K). It can be re-arranged to show how temperature varies as a function of time:

T ( t ) = T + Δ T e h A m c p t

The temperature would decay exponentially, you could use this to try predict the outcome however it is tricky to speculate without subbing in some actual values. It would have to be an experimental process and could just yield crap results.