Sunday, November 30, 2008


Granule Extruder - the sequel

I've rabbited on about granule extruders before. I think it's important to have a design for one as:

Here's my latest attempt held in a vice under test. It's a PTFE tube with a 10 mm hole down the middle containing the granules. At the left end a brass plate is (very securely) attached by being screwed in and then held with extra screws, which you can see projecting to the left. (Also note the the solid-state force transducer near the top of the picture...)

A length of M10 threaded rod is pushing against the granules in the PTFE tube and being dragged to the right by the string tied to it.

At the right end is a heated brass chamber where the granules melt and are then extruded. This has a nichrome heater embedded in fire cement with a copper tube round that for mechanical strength. The nichrome is 6 ohms, which is a bit high - it'll just about get to full temperature at 12v, but there's no leeway; the next one I do I'll make 4 ohms. The green and white wires are a K-type thermocouple.

Here it is extruding:

This is running at 243 oC. To get the same flow rate as on my conventional RepRap Extruder at the same temperature I had to hang 17 kg on the string - a force of 167 N. The flow seemed very smooth and even under constant load and - equally importantly - seemed to stop pretty quickly when I lifted the load off. As you can see, I've run the extrude nozzle off sideways so that the main axis of the device is horizontal. This reduces heat convection from the hot end to the cold end, thus keeping the granules at the cold end from melting.

Here's a close up of the device with some of the extrudate it made. The test tube (10 mm internal diameter) contains the granules that didn't melt ahead of the screw. I tipped these out at the end of the test. They form a buffer between the melt and the screw, and mean that it should be easy to withdraw the screw and recharge the device with more granules when the screw has moved inwards to the end of its travel.

Now to design a hopper, a motor drive for the screw, and a force transducer for it to push on so we can control the flow rate...

Labels: , ,


This is great news! Is the filament free of voids, or are there some bubbles in it?

Side suggestion: if it can be kept reasonably cool, an FSR (force-sensing resistor, e.g. from interlink or tekscan) may be an alternative to strain-gages for a force sensor. They're not as linear or repeatable as strain gages, but the larger signal sometimes compensates for those sins.

-- Larry Pfeffer
Not one bubble, even when it first started!

Thanks for the FSR tip. That's just what's needed, and I wasn't aware of them.
A granule extruder would be such a great feature. We will get access to many more plastics and indeed, we will finally be able to start recycling plastics (from objects we printed (design iterations) but also from many of the scrap plastics from packaging, etc.).

It is even more worthwhile to go this route since the current extruder is still one of the tricky things to get right.
Adrian, et al,

I'm pleasantly puzzled as to why a "simple" heated plunger produces a filament without voids -- especially since injection molding machines go to high pressures (and venting) to get the air out of the pellet feedstock. My best guess is that going slowly (compared to injection molders) is letting the air coalesce above the melted plastic.

Now I'm wondering if a slow-moving screw mechanism could achieve the same thing, and not need periodic re-filling.

-- Larry

Side suggestion: to give a constant force to the piston, you could give the motor a constant current.
Using Quantum Tunnelling Composite is another way of measuring pressure.
Larry -

Yes - I think you're right. If you imagine pushing solid lumps into a viscous medium that is on the verge of setting solid, it's obvious that...

(...goes and thinks for 5 minutes...) IS obvious that the air between the lumps will just get excluded at the interface.

An auger feed (which I abandoned a while ago as it's too hard to make a high-temperature strong auger) might display the same characteristics. But I think it's less obvious :-)

Enrique -

Yup - constant current might be a good bet for simple open-loop control. However, Larry's FSRs only cost a few dollars each, so we might as well burn another analogue input pin on the controller and have closed loop...
An auger feed is fine as long as it is behind the heatsink stage which you showed in your previous post on granule extruders.

Check out

for the intermeshed screw pump. We'd only use one side of it. That should nearly be fabable with reprap. It will definitely be fabable once we have a support material. The heatsink or PTFE barrier will keep the screw pump safely away from the hot tip.
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to
Posts [Atom]