* Another multi-nozzle tool changing design * Hi guys,

gplus
discussion
(Xiaojun Liu) #1

Another multi-nozzle tool changing design

Hi guys,

At the end of this year, I would like to share my design of a multi-nozzle tool head. You may or may not remember that a year ago I’ve posted some results of dual material 3d printing on my DIY Ultimaker 2 (https://plus.google.com/+XiaojunLiu/posts/4kjzNmNFms). I said I used a different approach which has not been implemented (at least I didn’t find) elsewhere.

Well, the nozzle changing idea I came up with was inspired by the multi-lens microscope (where three or four lenses are distributed around an axis, mounted on a plate. To change from one to another, one can simply rotate the whole plate so that the lens engaging is pointing towards the object).

See [Fig.1]

Took this idea, I replace the lenses with nozzles. For example, given three E3D kraken style heat breaks, heat blocks and nozzles distributes around a slightly leaned axis, at an even angle. The axis is leaned at a certain angle to ensure the nozzle engaged will be pointing perpendicularly towards the heated bed.

See [Fig.2]

Then the heat breaks will be inserted and mounted into a single heatsink.

See [Fig.3]

At the top of the heatsink a large 6807zz bearing is inserted to ensure the rotation with the given axis. Actually we use two bearings for the axis to avoid wobbling when performing the rotation.

See [Fig.4]

But a free rotation is not what we want, for the engaged nozzle will rotates away by fiction. To implement the intermittent rotation, I used the Geneva mechanism, as you can see the Geneva wheel is fixed in the back side of the heatsink. So that the drive gear rotates, drives the Geneva wheel rotates. And when the engaged nozzle is printing, Geneva wheel is constrained so that the rotation freedom is eliminated.

See [Fig.5]

To drive the gear, many approaches can be used. For simplicity, I took the mechanical way (inspired by the Ultimaker 3). I used a rack to drive the gear. So that if the rack is pushed from either sides, the drive gear rotates.

To do that, two rods must be mounted on each sides of the printer, so that if the tool head moves towards the end of the rod, the rack will be pushed.

See [Fig.6]

Then a printed pivot part will connects these components together. The pivot part is also good for holding all the wires and Bowden tubes, which will be connected to the nozzles.

See [Fig.7]

Then comes with the printed enclosure and two fans. The left axial fan is for cooling the heatsink, and the right blower fan is for cooling the printed parts.

See [Fig.8]

Finally, at the bottom a sheet metal piece is mounted to the enclosure to prevent ooz

from un-used nozzle. Also a printed shroud guides the air flow out from the blower fan to the printed parts.

See [Fig.9]

Currently the whole tool head design can be used as a module for UM2. The rack can be pushed by two rods which are fixed on two sides of the wall of the UM2 machine.

See [Fig.10-12]

As you may see here currently only two nozzles have been mounted. The repeat accuracy is impressive, I’ve printed a lot of dual-extrusion models, such as the Gyro, the traffic cone, the hand-drill and also a strandbeest I designed on thingiverse (https://www.thingiverse.com/thing:3263196).

I’m working on the hardware and software to make three nozzle printing possible.

Also, I’m trying to make this design more compact. The heatsink part needs to be simplified, improved, for easy making. And I have not give a good name for this tool head either.

So please help me to improve it! Any ideas, comments, names are appreciated!

Merry Christmas! In advance.
XJ

(Griffin Paquette) #2

Assuming the machining quality is top notch, this could be very viable. Nice design and write up.

Will we see one built in the near future?

(Step Cia) #3

where are your design file and how do we follow any revision you have made? github? might I suggest you look into duetwifi their tool changing method is quite fun to work with. thanks for sharing

(Step Cia) #4

if you want to make this a competition please by all mean do it. I’m merely conveying my interest to help if I can spot areas of improvement as per the OP original request.

(Mike Kelly) #5

I was also staring at a microscope having this though. With high precision machining it would be very nice, but getting repeated motion and position locking will be the challenge of this approach. Ball detents will likely go a long way.

It reminds me a bit of the machine made by diabase: https://www.youtube.com/watch?v=IZ82_L6UmzQ

(Ryan Carlyle) #6

Like it!

(Xiaojun Liu) #7

@Step_Cia I’m thinking make this available on github. It is a matter of time, since over time I have a lot of slightly different files I need find the right one.

(Step Cia) #8

better think up a good name for it :slight_smile:

(Xiaojun Liu) #9

@shauki Remind me of Prusa MMU2 filament changer :grin:

(Xiaojun Liu) #10

@Mike_Kelly_Mike_Make I was looking for the locking mechanism for quite some time. The ball detents was the first reference I could think of. I also considered the magnetic way to lock. The ball detent will need a greater force to unlock. Actually the force to unlock it is the same to fix it. In comparison, the Geneva mechanism needs little push force to the rack end to unlock, and a much greater force acting on the nozzle side to unlock, therefore I think this mechanism is more suitable.

(Xiaojun Liu) #11

@Step_Cia Haha, please do the competition as you wish :slight_smile:

(Xiaojun Liu) #12

@Griffin_Paquette The most difficult part was the machining of heatsink. I’m trying to simplify the design currently.

(Step Cia) #13

I don’t have the bandwith to start new project at this time… :slight_smile: But your design is intriguing. I’d love to play with it virtually

(Ryan Carlyle) #14

I like the Geneva mechanism here for the indexing feature. Helps open up motor/actuator options.

(Jeff DeMaagd) #15

@shauki Sharing the same nozzle and swapping filament works but means a lot of wasted plastic. With a lot of the examples I’ve seen, more plastic ends up in the purge block than the actual parts.

(Xiaojun Liu) #16

@shauki It will produce enough torque to rotate. Actually a 9g servo will be fine. I’ve thought about them, I just had no good place to fit them…

(Xiaojun Liu) #17

@shauki I’m not sure about the holding torque. I think it will be fine to drive the gear and the Geneva though.

(Xiaojun Liu) #18

@shauki Do you know mosquito HotEnd? This one have the similar small form factor, but mosquito is more higher end solution.

(Xiaojun Liu) #19

@shauki Right, the heated blocks should be round shape.

(ThantiK) #20

This design has been done before, by the way. – Just in a slightly different method.