interest in pellet Bowden extruder for deltas?

There hasn't been too much recent research in direct pellet Bowden drive and this may be a missed opportunity. Specifically for delta printers, I feel like we aren't maximizing the vertical space that deltas have which consumer Cartesian machines do not have.

Cartesian machines typically use direct drive extruders while 99% of deltas use a Bowden system. Research in direct filament extruders fr on pellets have proven to be far too large, heavy and unwieldy to use with retraction issues.

To install a filament extruder to a Bowden system in a machine such as an ultimaker takes much modification above the system and is seemingly impossible to mount on the frame itself staying within frame dimensions.

However, I believe we can utilize the space above the effector and stay within reasonable dimensional constraints for larger delta machines.

The most common filament extruder today is the Lyman filament extruder which the designs can be found on thingiverse. It's large, heavy, and a horizontal extruder.

There was research done with makibox ramen pellet extruder last seen Q4 of 2012 loading pellets in a linear fashion, but nothing has come to fruition.

The current hot end design is well established and wide spread and thus keeping that in place for the Bowden Is essential for ease of adoption at this time.

With that being said, the focus of current filament extruders is speed to be spooled up. This mean very long auger screws are used to transport as many pellets to the heated orifice as possible. This requires very high torque from a windshield wiper usually.

I hope to replace this long auger screw with a much shorter one pushing less pellets through as we do not need filament delivered that quickly ( only a theory, research and experimentation required) or just a standard screw. This would imply less necessary torque, so that we can possibly use a geared stepper to drive the rotation. This is an important point using this motor.

Using a stepper motor allows us to directly integrate control to already established controller boards as a second extruder. However with this current design theory, retraction is not possible as pellets are continually fed down to the hotend. This means our current standard extruder will probably be necessary to control retraction and extrusion to the printing hot end. With the filament extruder motor being seen as a second extruder. I theorize that we may be able to synchronize both extruders. One to stop filament production and the other for retraction.

Hopper -> screw driven by stepper motor -> Heating element -> 1.75mm diameter die -> Extrusion (cooled?) -> filament detector via mechanical endstop switch -> Main extruder -> bowden tube -> hotend

More research and experimentation is necessary in:

-If stepper motor is a viable replacement to the windshield motor
-Temp of plastic extrusion and how soft/ solid the plastic immediately after pellet to filament extrusion point
-Whether or not cooling is necessary prior to main extruder to retain rigid shape, if so, how much cooling? Direct fan cooling, tube directed blower fan across the the extruded filament
-Temp of hot end required to print the fresh filament as the filament will be soft/warm/hot?


This could very well lead to developments in directing masterbatch pellets to allow for full color printing. As I continue my research and development, this may very well fail or not, but I would really like just to facilitate discussion about this topic.

Edited 2 time(s). Last edit at 05/08/2014 02:12PM by 3DRapidClone.

Hmm... not to rain on your parade, but I don't think this is a viable option, not with pellets at least. I've built and operated a lyman-esque extruder for the past few months. There are two main reasons:

1) The pellets are 3-5 mm in diameter each. A smaller screw would only feed them a few at a time, and it would be difficult to control the amount of extrusion.
2) Even with a gigantic motor with 12-15 Nm of torque (~10 times what a geared nema17 could put out), a filament extruder can only push out 14" / min of 1.75mm filament. A well calibrated printer doing infill will use more than this rate.

The first problem can be alleviated if you switched to a fine powder. The filastruder folks have used nylon powder from SLS printers with success. With powdered plastic, it's possible to drive a much smaller screw and still have even extrusion. Alternatively, some variety of closed-loop feedback using an optical sensor to measure the amount of plastic pushed out may also work. Also, will powdered plastic be any cheaper than filament?

The second problem is a little less tractable. You can switch to a smaller screw with a higher speed, but the total amount of energy required to form the plastic is fairly large. The resin pellets come full of air pockets and bubbles. I believe the high pressure within the tube is what helps you get nice and even extrusion. In other words, it might be more economical and easier to deal with if you formed the plastic into an even filament with a larger machine, then printed with it. Hence the status quo today with a separate filament extruder

BTW, you might want to check this out: [reprapdelft.files.wordpress.com]
They have a drawing on page 23 of a putative direct extruder that you could potentially try building!

Edited 1 time(s). Last edit at 05/10/2014 03:57AM by jamesshuang.

No need to apologize, just trying to formulate ideas. I was discussing with someone on the Google groups about an idea.what do you think about using a heated chamber to melt the plastic into a vat which is channeled down to a screw drive to extrude. The speed of extrusion would be consistent, and when the nozzle has to leave the print, it would just divert the nozzle away from the print, extrude the excess, return to the print. It should help print quality. And because we are melting the entire chamber, the excess plastic can just be placed back in again

Mmm unfortunately melted plastic doesn't really act as a liquid. It's more of a gooey paste. It wouldn't "flow" like water from a vat. That's why an auger and a massive motor is necessary -- to push the paste through the die. The variety outlined in that recyclebot document actually looks like it might work. I've never tried extruding such a thin filament before so I don't really know the behavior. Might be worth a try?

Edited 1 time(s). Last edit at 05/10/2014 12:01PM by jamesshuang.

From the funnel down the tube, there is no die, its just the plastic itself going with gravity, then the auger bit will drive it to extrude.

Quote
3DRapidClone
From the funnel down the tube, there is no die, its just the plastic itself going with gravity, then the auger bit will drive it to extrude.

You are going to gravity feed molten plastic? Or are you going to gravity feed pellets? First option wont work at all, second option will work intermittently if that.

Why not just extrude into a cooling section, then straight into the bowden tube.

I think you are going to run into problem when you retract. The filament will have no where to go like it does when it is on a spool.

---

greghoge.com

HUGE 3D PRINTER PARTS SALE!!!

Could you elaborate on a cooling section?


On the first idea, yes, there would be difficulty upon retraction. However, Im looking into the similar methods from magnetic tapes, which allows one to continually spool and the other tape to stop and go.

On the second idea, there would be no retraction, it would constant feed, the toolpath would be redesigned to divert the plastic instead of a jump between gaps. Since the hopper to the heated section would be heated all around, the diverted plastic would then be able to thrown back in with virgin pellets.

As for gravity feed, Pellets go down from the funnel through a tube to the auger bit. The entire pathway would be heated with multiple thermistors to maintain even temp. The auger drive would extrude plastic.