Hugh Lyman won the Desktop Factory Competition

Could you explain why the filament diameter after extruding, up to solidification is more ambient sensitive if you are going slower/faster?

I can see why the temperature gradient would need to be very controlled for the barrel of the hot end, but the temperature of the molten plastic should be the same for both speeds. After the plastic leaves the nozzle, the temperature gradient of the hot end has nothing to do with the solidification, and the shrinkage that may incur. All that matters is the heat within the plastic filament... Unless I'm missing something here.

... when extruding viscous material through a thinner nozzle, you'll get some 'swelling' after the nozzle, what's depending on speed, nozzle diameter, viscosity, ambient humidity, temperature and several material properties.

When changing speed, pressure and temperature gradients before the nozzle, this dependencies will be of even more 'weight', so small changes in ambient temps or humidity will interfere with this 'swelling' more too.

I had a similar problem with dispensing viscous heat curing glue in contour lines - we had to strictly controll all temperatures and pressures to get it nearly correct, but had variations over the day with maybe 5% to 10% of the dispensed volumes

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

If you are extruding fast, the plastic is going to get further from the die before it solidifies. Thats a larger area where something can mess up the dimensions, cause it to kink, etc. You would have to cool it much faster which is why the industrial extruders have to quench it with water. Also the die would need to be thicker to allow a much more gradual transition from the barrel to the opening.

When the plastic is squeezed down to go through the die, there are elastic forces that get released when it comes out, which is what causes the swelling. The thicker the die, the more time the stress has to sort itself out before emerging. Also the faster the plastic is compressed, the greater the bounce back. When the feed is too fast and the plastic compressed to quickly, the release of those stresses causes melt fracture, which is a rough, broken surface on the filament. That places an upper limit on production that gets reached eventually no matter what you do to minimize it.

I've seen melt fracture running at 18"/min with a 20mm plug drilled straight thru with one diameter. Making more of a funnel going into a die will fix that, but it's possible to have the problem even at low speeds.

The biggest problem with running faster is that there isn't much of a way to do it without quickly increasing cost. Either with a bigger motor, or a more complex die that would need to be machined, more temperature controls, etc. It isn't going to happen for $200 which is how much 1ft/min costs. If you have to pay much more than that, you will have to extrude a lot more spools of filament to make it worth building an extruder vs. buying filamet. So long as you can extrude faster than UPS can drive, you are coming out ahead.

If you want to do 1kg/hour or several KG a day even, you are going to need to sell filament to subsidize the cost of that machine and compete with the pros. You would need to be either cheaper than China or offer things they don't, like specialty colors in small quantities, metallic effects, etc.

Just an idea; how about a multiple hole die?

It may get quite complicated on the spooling end, but could multiply throughput without having to go for very fast extrusion speeds.

I suppose the complexity would not be worth it for someone making filament for personal needs, but for a small time producer or a group of people (hackspace) it may be an interesting path forward.

/Andreas

Ian,

Have you attempted to make a filament like galaxy blue from faberdashery? Something with flecks of material in it?

I'd really like to build one of these things, but if I do I think the wife would have an aneurysm. She already complains that I'm slowly killing her with ABS fumes...

---

- akhlut

Just remember - Iterate, Iterate, Iterate!

[myhomelessmind.blogspot.com]

Clariant, a company that makes colorant masterbatches, has some metallic effect colorants. I would love to get some samples to try out, but I can't get them to respond to me. Small potatoes compared to the typical customer that orders by the ton.

Interesting point about the die swell and melt fractures. Thanks for the info!

How do you guys should think the extrusion bot does it?

I know we want to keep the cost low but I think speed for making a kg of filament should be around an hour to an hour and a half max.

Just increasing the extrusion speed is not a viable solution if you want to maintain filament accuracy at +-0.01.

I think the extrusion bot uses a small constriction before pushing the plastic into the melt chamber. That constriction acts like a gear pump which is used to maintain extrusion quality in bigger machines. Giving feedback to the motor to adjust speed depending on internal pressures is way too much work for a thing as simple as this.