multiple filament bowden extruder drive

by fritzgutten Jan 23, 2012
Download All Files

Thing Apps Enabled

Please Login to Comment

when i open the files the part are massive?????

Hi, great work on the extruder! I was wondering if you cold explain, in a little more detail, how your print file controls the motor to select the new filament. Are you hand coding the motor rotation to select the filament or are you using a "T" (tool change) gcode to select the filament?

Hi fritzgutten, Great concept and design. I've got a dual extruder setup, but it's not nearly this elegant. Any updates on the design or how it's been performing to date?

finally got all 3 running! see the horrible results here:



still much tuning left to do, but i'm calling it functional. 8-)

Finally got more than 1 hot end functional at a time, so to check out the filament drive in action, see: http://youtu.be/CVaYHlLkvWMyoutu.be/CVaYHlLkvWM

and http://youtu.be/yqAGEqaoPg4youtu.be/yqAGEqaoPg4

only 2 colors at the moment, need to finish another hotend.

Could you share your original design file(s) too? Stl-meshes are much harder to handle in most cad programs than the native format (or imported format).

I really like the design, but I think the separate stepper to stand in for a multi way swicht is not a good idea. If you start your firmware up, you have no way to know what state the machine is in.
Replace the switch stepper with a few cheap 3$ servos (or 5$ ones if the cheapies are too weak) an
d drive the servos with PWM from the arduino and a separate 5V line.

thanks for the input.

i'd be glad to share my files, what would be most universal? I'm running solidworks 2005.

as for the stepper drive on the filament selector, i will be installing a home switch on that axis. I'm unsure how current electronics will handle the stepper or home switch, but figure its an option as dual extruders are coming out.

many printers are running or have run by manually setting the home position, this could be an option for the filament selector motor as well.

hobby servos may indeed be an option, but control is less understood for myself. driving multiple servos would simplify the mechanical setup, but complicate the programming and control scheme.

I am using EMC2 and the use of a stepper is not much of a concern for my use.

Ah, EMC is indeed less hackable (IMHO), but then you will be putting all logic to change tools further upstream, into the gcode?

A floating home location doe not matter that much for additive manufacturing, it does not matter that much where exactly the object is built, as long as it is built correctly. With your design here that is not the case, it matters a lot wich filament is being extruded. So you will need that home switch.

But running servos from EMC should be possible as well, it sure can do PWM. I wonder if you have enough IO ports though.

Actually, I think you will lose much more time hacking EMC to do your bidding, than if you were working with a RAMPS board and marlin firmware.

I believe have enough IO ports, the only thing i havent configured is the home switch in question. As far as hacking EMC, there is zero "hacking" to do, just configuration of XYZAB axis. I am not currently using EMC to control tempuratures, just the standard gcode moves.

at any rate, its all possible whether using reprap electronics or driving from a parallel port.

as for the home switch.... the home position of the filament can be set , by hand, to a know location, all moves from there are known locations....the home switch isnt absolutely necessary, but it will
make life much easier.

SW2005 would be awesome! If you output iges as well that is a bit more editable and less lossy on Nurbs packages like Rhino.

In future maybe I suggest STEP format. Its comes in clean to proe, and have found it quite good going back and forth with solidworks and rhino.

Step files should always be given, platform agnostic!

Is the idea to have N hot ends for N filaments, or to have the filaments join into one hot end?

N hotends for N filaments

So... it looks like you are using one drive to "select" the feed and apply pressure on the selected filiment and the other is the actual drive, which drives all three, but only if they have pressure on them. Correct?

wow! i was dreaming about something like that! Can you post some more info about it's inner working? i'm no engineer... :)

I'll see if i can get around to it. busy constructing my other hot ends to test it out.

but basics are : the drive shaft uses hobbed grooves like most, except it is about 2 inches long or so and has 3 grooves equally spaced. the idler arms press the filament into the grooves independant of one another. the arms are vey similar to others out there. instead of using springs or a lever to engage the filiament, there is a 3 lobe cam running along the back of them that can be rotated to select which idler is engaged. filament is selected by isuing a gcode command to the cam motor to rotate to the proper orientation to engage the correct cam.


now, if we'll use two materials (example - PLA and ABS) - we will need to consider the fact that each filament act differently (for example - 10mm of pla extrusion gives different amount of plastic then 10mm extrusion if ABS).

That can be manipulated in software though 8-)