[ Closed ] Filament Extruder - Convert pellets to filament (Page 49 of 58)

I'll be double checking all the numbers and redoing the experiments at some point.  This was all quick and dirty. so take them with a grain of salt.  Problem with this motor is that it is not going to regulate speed very well with delta load because of the Rs.  Good news from the data it looks like RPM doesn't matter.

What it does say though is that the take up reel I was messing with, using a capstan, won't work in direct pull mode unless I also regulate the auger RPM.  Getting too complicated.  May try putting the first copper guide (nearest the nozzle) on an arm driven by a hobby servo to regulate the long term thickness, and use the capstan after a big take up loop.

Sorry about the socket - that one's on me.

The motor lugging is it slicing a pellet, as far as I can tell. There's pellet bits in the hopper when I run it completely out, like they're getting sheared. I think that's why the "T" approach jams - no sharp edge to shear against.

In my opinion, a 7.7% variation in heater capacity is not a big deal - there's plenty of thermal mass. I am surprised your mains voltage varies that much, though.

Yes, I'm at 200 hours on the present PID. The new PID controllers have relays rated at 1M cycles, so 200 straight days of extrusion @ 3 cycles/min. That's 200kg. I think that's reasonable. A DC SSR is not ideal at this price point.

DC gearmotor, latex tubing, nylon bushing, 1/4" bolt, electrical tape. The stuff dreams are made of.

I gave up on the direct draw down when I realized that the feed rate from the extruder was so variable.   I don't think there is really a way to ensure constant pressure in a system this simple.  Instead I'm focusing on letting the filament drop and loop back up to the spooler.  I made a switch out of a LED and Photoresistor to keep the winder and extruder in sync.  I tried hanging a flag from the filament which would shade the photoresistor when the motor has been off too long but it was difficult to get the filament loop just right.  My next attempt will be a seesaw arrangment that was suggested here awhile ago.

I drilled a nozzle opening at 90 degrees in one of the hex faces so the filament can extrude straight down, and that is showing promise in preventing the kinks that happen when you try to manhandle the filament path. 

I've been trying to simplify things a bit and replaced the stepper I was using for the puller with a DC motor again, but I can't seem to shake the Arduino and LCD.  I want to be able to control the speeds of the spool motor, puller motor and fan, and use thermistors to read the temps at the melt zone and nozzle.  It would also be cool to count the length of filament wound.   At the same time I want to come up with a design that isn't too hard or expensive to reproduce.

Hi all,

Just bought in to the kickstarter - excellent project! Out of interest, has anyone tried feeding finer pellets or powder yet? I'm intrigued with the possibilities from recycling, and was looking at using a very coarse burr coffee grinder - which would result in ~1-2mm (depending on grinder) particles - to grind and mix virgin and recycled ABS.

The main issues I could see would be the auger sticking a little, and lack of expansion space at the hot end (due to reduced air-gaps between particles).

No worries if not - I'll probably try it when mine arrives in August ( ) and will report back. I could see my having 2 grades of filament - virgin for final prints, and recycled for prints where quality/consistency is less important.

More than 1/2" in any direction seems to double or triple the chances of hot ramen action. Maybe it will work for you though!

Cheers for letting me know before I tried it! Were you using a burr grinder, or a cheaper grinder (or even a chopper - as a lot of cheap coffee 'grinders' are)?

Looks like some kind of cross-cut shredder may be the next best option then. NB: Recycling is more of a long-term aim - ABS pellets is definitely the plan for near term.

justnick,  how do you plan on using the calipers?

I'm not sure if the filament can shade the photoresistor enough to trigger it.  It would have to keep track of whether it went up or down, or use two sensors to detect if it is rising or falling.  I'm working on having a flag of some kind move in front of the sensor.  The other idea which may be simpler is a limit switch with a stick glued to it.  The pressure needed to engage the switch should be light enough at the end of the stick.  Once the loop is running stable, put the stick above the loop, so when the loop rises high enough it pushes the stick and switches off the motor. 

What makes all this complicated is the fact that the filament is rigid.  It may drop into a loop and go back up, but it is not following a path like a snake or a string.  It is more like a semi rigid coil that is rotating.  If the filament is a "U" shape and you hold the end of the left leg and then let go of the right leg, the bottom curve of the U doesn't just drop.  It stays in place and supports the right leg by itself.  In this case the top of the right leg is the extruder nozzle, and the filament bends because it is being held up by the bottom of the loop.

You can pull filament around like a string, but the way it reacts is complicated by the fact that it has a shape that it tries to maintain at the same time.  The key to predictability would be cooling it enough that it is below 100C before it takes a turn so that its natural shape is straight and not bent.  Also if you pull it around a roller or other guide that is fixed, or at least too hard to move, it will act like a rope around a pulley and that pulling force will stretch it from the nozzle.   That kind of force has to be isolated from the nozzle by being on the other side of a pinch roller (like the puller) or a free hanging loop.  You can constrain it gently with bumpers, but if you want it to change direction it has to bend naturally.  Any solution that involves the filament moving something will require resistance small enough that it doesn't "pulley" around it rather than moving it.  An optical-only solution requires the filament going where you need it and staying in that area, and guiding it such a way that it "wants" to go there.

Winding with an industrial extruder is fairly simple thanks to the complexity of the extruder and the ability to draw down.  With a simple extruder, the complexity is transferred to the winding.

I made a nozzle with a downward opening, and tried taking out the puller and using only the spooler, switched on and off.  It doesn't work, you have to have some kind of tension or you get a big mess.  I think it will be awhile before you see Makibox Ramen, because that napkin sketch you can buy for $150 won't work as shown.

On a side note, I just bought 10lbs of SABIC MG94-NA1000 from Ebay which is considered higher quality plastic for the same shipped price as 2.2 lbs of Jet filament from Amazon (the cheap chinese stuff).

- check with François Lahey at Voxel Factory. He said this about pellets:
"I have order a "/$"/ load ;-)"

so I'd guess he has some.

- nylon spacers (on auger and main bolts) solved the thrust bearing issues, yes. I've run for 100s of hours, no noticeable wear to the eye. Previously I was absolutely thrashing bearings, in under 10 hours. Shattering the bearing faces.

- 4 hours seem to do it. I haven't had any issues with clogging, I've run 10+ lbs through my Solidoodle.

I'm in on the Kickstarter, kit + 1 extra nozzle in case I mess up drilling that!

François did not even know you launched a kickstarter:
"He made a kickstarter without telling me? Could you please send me the link?

Regards,
François"

I sent him the link

Ben

Weird he is usually very responsive!  Maybe the cold winter/snow in Quebec this year froze his brain .. lol

Ben

I told him he was the Steve Jobs of home extruders.  He didn't invent it, but he refined it and made it available to the masses.  He needs to start working on the persona he's going to present to the press.  Turtlenecks and hipster glasses are taken.  Maybe a cravat?