Measurement of Melt Temperature in Extrusion

By Alan Griff

Related links: Extrusion Melt Temperature | Plastics Extrusion Books

Measure in the melt stream after the breaker plate. That's where it's done.

Set the wall controller at/near melt temp so it doesn't cause center-wall gradient.

Variable Depth better, but fixed OK as long as you don't believe it too much. It will do a good job telling you linear variation, which is serious.

Shear Heating is faking the thermocouple because of motion of melt. Easy to measure, most people don't need to bother, may be as much as 10C.

Educating people to understand and observe temp readings is as important as getting them. Sometimes I think the biggest value to temp measurement is not the reading but the sensitizing of people to what's happening inside.

Reactive materials, such as foaming agents and crosslinkers, are the most trouble but have the greatest need to know. For all others, the reading is representative and can be used with some experience. When you know what good is, you'll know what fishy is.

Probophobes are people who run PVC and are afraid a probe will be a degradation site. If the rest of the system is so streamlined that this is indeed a weak point. For the rest of us, get the melt temp, even if you get a flush-infrared to do it.

Flush sensors should be infrared, although a combined flush temp-pressure instrument is sold. It reminds me of the soggy-cookie principle: to a kid, a soggy cookie is better than no cookie, and in an extrusion line, a flush thermocouple reading is better than no reading, once you get used to it, avoid excess wall effect and relate it to production behavior.

IR Guns and Needles are supplements, need special skills, good to use, hard for certain products (e.g., blown flim during operation)

Insulate. The best heater is a sweater, I was told in England not long after The War. Insulation also reduces unequal radiation losses that make for different thicknesses.

Feedback Control is almost never done from melt temp values. Time lag too great, other ways of managing variation.

July 24, 2007