[ Closed ] Filament Extruder - Convert pellets to filament (Page 33 of 58)

Voltage - set to 20vdc on your meter, probes in the V and COM terminals, pull the red cap off the motor, touch probes to motor terminals where the wires are soldered. Should be around 12v.

Current - set to current on your meter, probably "10A". Wires in terminals marked "A" and "COM" or similar. The motor will pull around half an amp or less. Put the terminals inline with the motor circuit - you'll have to cut or disconnect one of the wires. Then, i'll look something like this:

black wire from power supply --- meter --- motor --- white wire from power supply

You can install it on either side of the motor. It just needs to be in the loop, in series with the motor. Keep in mind that when the leads are in the A and COM terminals on your meter, the meter is a dead short - you wouldn't want to accidentally try to measure voltage with the leads in the A and COM terminals, you'll short the power supply. Theres no danger to being and V and COM and trying to measure the current, the motor just wont run as in voltmeter mode, the meter's resistance is nearly infinite.

RPM - I just put a sharpie mark on the shaft and time how long it takes to make 10 rotations.

Why I ask all of this: With a 1.6mm die and no puller, the motor is somewhere in the middle of the torque curve. I'm worried that with a 3mm die and a puller, the motor is very much unloaded, and starting to touch its no load RPM. On a DC motor, current is proportional to torque, and applied voltage is proportional to speed. Of course, these things are linked. Here's an equation to represent these variables:

V = I * (Rstall + Rbackemf)

Rstall is the resistance the motor has when the shaft is locked. At 0RPM, there is no backemf, applied voltage is roughly constant (~12v in our case), so current is at its highest at stall. Hence, a DC motor produces maximum torque at stall. As the motor spins faster, it creates back-emf, a result of Lentz' law. It increases the effective resistance of the motor. Thus, the no-load RPM is finite.

Anyway, I'm curious if you're starting to hit the no-load region of the curve, which would result in inconsistencies - the motor can't speed up any more, even though there's nearly no load. Since we want pressure constant, we need torque to be constant, not speed, since pellet density is not uniform - it depends how the pellets are sitting. Sometimes they pack more densely into the pipe nipple than others, so pure speed control of the extruder motor isn't enough. If you're nearing the no-load portion of the curve, then extrusion backpressure will drop because the motor can't feed fast enough.

Disclaimer - I've made some simplifications here for the sake of brevity.

Motor RPM is about 1.11 for the 1.54 nozzle, and 1.16 for the 2mm.  12v in both cases, I haven't measured current because I'm too lazy to cut the wires at the moment.

I measured feed rate coming out of the nozzle.  I tilted the extruder so that the filament hung straight from the nozzle past the edge of the table so it didn't touch anything.  I clipped the filament at the beginning of the timer, and marked it at one minute so there would always be the same weight hanging from the nozzle.  For the 1.54 plug I was getting 2.5"/min at 180 celsius, which increased to 6-8 when I bumped the temperature to 205.   For 1 minute of extrusion at 205 I was getting a range of lengths with as much as 1" or more in variation.  Diameter tolerance was still in the area of .05-.07mm or so.

I think what is happening is that there is a lot of variation in pressure at the die.  The diameter stays within tolerances while the variation in pressure translates into variations in extrusion speed.  With the puller, that speed is forced to remain constant.  The filament is stretched into a straight line from the die to the puller, becoming a rigid rod once it passes the fan.  If pressure increases, it can't speed up extrusion because it runs into the end of that rod of cooled plastic so it widens instead.

I was thinking that it would be good to have a spring between the collar and bearing so I can see how much the pressure varies.  It occurs to me that it might actually be a fix, if the pressure variation can compress the spring rather than squirt the plastic faster.  The trick is getting a spring that is strong enough, or perhaps something else that compresses that isn't a spring.

One method of coupling that Tim tried was a socket with a washer jammed into the end that normally goes on the wrench.  The washer slid into the slot on the motor shaft (of the original motor, I think he mentioned the current one has a flat).  The socket is able to fit the auger shaft and turn it while still allowing it to slide in and out.  You can use this kind of coupling, and leave some room for the auger to push back into the socket when the spring is compressed, rather than needing to move both the auger and the motor.  Something on the auger could be set up to then hit a limit switch, or maybe a rack and pinion arrangement to turn a potentiometer.

If the whole thing gets controlled by an arduino, then you might be able to sense the compression and tune the motor speed directly.

It is likely to require a large stiff spring, but that is something that might fit around the outside of the coupling.  With everything rotating, it makes it harder to do a microswitch, but a photo diode would be something that would fit nicely with an Arduino controller.  It would be better to have a bit more movement of the spring without changing the pressure too much (longer spring), since the slice and dice action of the auger might create additional movement in the coupling as it rotates.  I suppose a controller could average the position over about a revolution to keep from over reacting.

Yes, but mechanical pots wear out if in a production machine.  Light beams have a very low wear factor.  ;-)

1.16 for the 2mm nozzle? Weren't you using a 3mm at one point? Unsure if that was a typo or not.

I think there's quite a bit of variation in pressure at the die. The motor is fed a constant voltage, not a constant current. That means torque (and pressure) can be all over the place.

One issue with the spring idea is that some of the thrust load comes from the pressurized melt zone, and some comes from the sliding of pellets in the barrel. I'm not sure what the split is.

Even though contaminates have reduced quite a bit for me I am curious as to how well the Filastruder would work with a screen.

I only have extensive experience (>1lb printed) with white and gray. Their MFIs are here:

https://docs.google.com/spreadsheet/ccc … iSmc#gid=0

They are 12 and 4.5, ASTM D1238, respectively.

I could print the white at 150mm/s, the gray at 90mm/s.... so I'd say somewhere between 4 and 12, using the ASTM D1238 standard.

OSPrinting,

Any word on new ABS masterbatch colors?

cool. I am very happy with the green you currently sell.  Just getting a bit bored with it.

How much have you extruded? How much have you printed?