A different multi-extrusion approach - UM tool / printhead changer

[neotko 1,417]

Using a design like this

http://www.thingiverse.com/thing:340557

Everything would be much easy to clean and wipe and heads. And it could be assembled just above the bed cap (the dust cover).

Problem?

Not enough free pwm connections?

I been reading as much as I can understand. And so far the only free pwm connection from marlin it's on the EXP3 PB7

I suppose with that it could be posible to command a simple small motor pwm to keep it 'cleaning' or zero power. Also I suppose that since this would not need pwm a empty port to just send on/off should work. But designing boards it's above the stuff I know. So if anyone can jump in to fill the gap to control an additional motor/servo to just be on/off would be quite nice.

Ofc I mean doing this without losing other ports. With dual heads all the common normal ports are used, and I won't even like to loose led power and I suppose also any um2 user won't like that either.

Anyone have ever try to control any other pwm exit that isn't the conventional ones (all used on dual) ? I suppose it would need pwm board (cheap and easy to do) connected to PB7 and controlled by gcode pin control.

[neotko 1,417]

Also another idea for my 'manual no power' cleaner. I'll start soon designing it but the main idea it's using this 90º linear motion, adding a spring so it return to the start position everytime.

http://www.thingiverse.com/thing:342301/#files

[neotko 1,417]

Well the basic idea and distances are ok. So the idea should work. Lots of more work to do on this... I hope it works

[KevinMakes 22]

This is my idea 'basic' with some legos. A mechanical return than when pushed brings the wipe. To save area. Ofc this it's just the idea :)I not even sure if I will be able to make it work since I never done something like this (except with lego technics).

 

you may be able to use this method with a high gear ration to a spinning brush to clean the head as it "wipes"

[neotko 1,417]

@KevinMakes sounds nice, any design you can do (even a doodle on paper) would help.

I been thinking all day how to approach this on the most easy way and I might go back to my original basic idea for a wiper (so many pages ago).

And @ultiarjan this might be interesting to you.

I remember that you said that you needed the Excel to calculate the z. I suppose it's because when you move the head to an Z position it's hard to know when it will need to go up, or when it doesn't need to move the Z (for your wiper).

I been digging on Simplify3D file to get all the place holders (Simplify3D guys really suck big time on the documenting area).

And I did this (read all don't skip). Check the Bold part

{IF NEWTOOL=0}G1 F8000 X5.8 Y50 ; Prepare to Park 1 on Left Hangar

{IF NEWTOOL=0}G91

{IF NEWTOOL=0}G1 Z+10 ; Down bed for Wipe clearence

{IF NEWTOOL=0}G90

{IF NEWTOOL=0}G1 F8000 X5.8 Y0 ; Park 0 on Left hangar

{IF NEWTOOL=0}G1 F8000 X176.8 Y5 ; Go Kiss Extruder 0 on Right Hangar

{IF NEWTOOL=0}G1 F8000 X176.8 Y50 ; Extruder 0 ready

{IF NEWTOOL=0}G1 Z[current_position_z]

And the same on the other tool. You get the idea. I change to relative, then at the end the G1 Z[current_position_z] (case sensitive) move's the head again to the real Z position.

This... On paper works and the think, it's that it really works.

BUT... This it's what the Simplify3D Gcode viewers shows (First layer)

The Really really good news. After a good 3hours trying to find out 'what tha hell' or how to change the script to make it work, I was able to find the correct 'code' (not documented, not well documented and very bad documented) and the last 30minutes when I finally got the placeholder right, I just did check the Gcode of that 'first' layer. The Blue part of the image. This it's small piece of the Gcode generated:

G1 Z0.240 F9000

; layer 1, Z = 0.24

M106 S221

G1 F8000 X5.8 Y50 ; Prepare to Park 1 on Left Hangar

G91

G1 Z+10 ; Down bed for Wipe clearence

G90

G1 F8000 X5.8 Y0 ; Park 0 on Left hangar

G1 F8000 X176.8 Y5 ; Go Kiss Extruder 0 on Right Hangar

G1 F8000 X176.8 Y50 ; Extruder 0 ready

G1 Z0.240

T0

As you can see, the Script works, it make's the Z move +10, and at the end of the script the placeholder generated the real 'current z'. The big confusing problem it's that the S3D Gcode viewers goes bananas.

So... @ultiarjan about your excel

If you do this method, and split the process from the normal 2 process, to 4 and you make that the first 2 process print from 0 to Xmm (let's say you need to move the head for the first 10mm) you can make that first processes to have one script that moves the head and the other process to have other script without the Z movement.

I tried, and the gcode works. On my robot-ice test (gcode, not printing yet until I finish the wiper). This allows me to make the Z move down 10mm, then the head will always be 'not touching the bed' and will cross an area full of wipers, and then it will resume printing after returning to the correct Z. I think this could help to solve the problem of the 'first layers' where the head can make a blob of filament that can hit the print, or even make the head loose equilibrium).

So I think with this method you could make all by a script, and 2 extra processes.

Also interesting. After digging on the simplify3D I found all this placeholders that do work on the toolchange.

[toolchange_retract_distance]

[toolchange_prime_distance]

[retract_distance] ; doesn't work on toolchange, maybe on other place

[prime_distance] ; same

[retract_speed] ;same

[travel_speed]

[current_position_x]

[current_position_y]

[current_position_z]

[next_position_x]

[next_position_y]

[next_position_z]

[previous_direction_x]

[previous_direction_y]

[previous_direction_z]

[next_direction_x]

[next_direction_y]

[next_direction_z]

[average_direction_x]

[average_direction_y]

Btw with all this stuff, someone could make an easy script to report all the time where the Z it's to the ulticontroller. Or even more funny, the last x/y/z position, to maybe try to resume a print?...

It's been a good day of work =)

[KevinMakes 22]

@KevinMakes sounds nice, any design you can do (even a doodle on paper) would help.

 

I can slap something together.

Ill probably start with the hand drawn.

[KevinMakes 22]

@neotko

Can I please have your ultimaker2 design...?

[KevinMakes 22]

Im working on something, trying to keep it 3d printable as possible.

Edited September 23, 2015 by Guest

[neotko 1,417]

@neotko

Can I please have your ultimaker2 design...?

 

My design it's for Umo+

https://www.youmagine.com/designs/ultimaker-original-plus-dual-extrusion-magnetic-head-changer

For UM2, ultiarjan:

https://www.youmagine.com/designs/dual-head-ultimaker2

[ultiarjan 1,223]

@neotko thats interesting indeed, I'll play with it sometime ...

only thing else is I have a 15 cm wiper and I don't want to wipe at the same place all the time as there will be to many waste plastic in the same place and it will smear over the wiper. I solved that now by generating a random position on the wiper X position in excel...

And about the process splitting, thats interesting if you have a wiper mounted to the printer housing (which is probably better) but my wiper sits on the bed, so I have to do a z move at every change.... so not a real good solution for larger prints...

I need to experiment with different wipe materials, my ultimate goal would be to have a system stable enough to leave a print unattended, but this is impossible atm, still need to clean the wiper once and a while. And due to the z move and the time consuming head change I consider this setup only usable for small prints..

[ultiarjan 1,223]

 

@neotko

Can I please have your ultimaker2 design...?

 

For UM2, ultiarjan:

https://www.youmagine.com/designs/dual-head-ultimaker2

 

I have an update, including mounting brackets on the printer housing but it didn't upload that yet.. try to do it somewhere the coming week...

[Labern 774]

@neotko thats interesting indeed, I'll play with it sometime ...

only thing else is I have a 15 cm wiper and I don't want to wipe at the same place all the time as there will be to many waste plastic in the same place and it will smear over the wiper. I solved that now by generating a random position on the wiper X position in excel...

And about the process splitting, thats interesting if you have a wiper mounted to the printer housing (which is probably better) but my wiper sits on the bed, so I have to do  a z move at every change.... so not a real good solution for larger prints...

I need to experiment with different wipe materials, my ultimate goal would be to have a system stable enough to leave a print unattended, but this is impossible atm, still need to clean the wiper once and a while. And due to the z move  and the time consuming head change I consider this setup only usable for small prints..

 

If there was room you could have a rubber disk that rotates after every wipe, then at the bottom of the disk have a collection bin that the bits of plastic get rubbed off and drop into.

[neotko 1,417]

only thing else is I have a 15 cm wiper and I don't want to wipe at the same place all the time as there will be to many waste plastic in the same place and it will smear over the wiper. I solved that now by generating a random position on the wiper X position in excel...

 

I found also another placeholder inside the simplify3d, ofc isnt documented. I'll play some with it to see how to actually use it because it might be just off internally, so far on simple test it didn't work, but it exist (future versions?). And this should help on that

"{MATH }"

This it's the dump of info from the app hex edit:

 

[toolchange_retract_distance][toolchange_prime_distance]G92 E0[retract_distance][prime_distance][retract_speed][travel_speed][current_position_x][current_position_y][current_position_z][next_position_x][next_position_y][next_position_z][previous_direction_x][previous_direction_y][previous_direction_z][next_direction_x][next_direction_y][next_direction_z][average_direction_x][average_direction_y][average_direction_z]{MATH }0+$\.${IF ==!=>=<= =  !=  >=  <= =>< >  < TOOLFEATURETYPEtravelouter_perimeterinner_perimetergap_fillsolid_layerinfillbridgesupportdense_supportraftskirtprime_pillarooze_shieldM102retractM101primeRectilinearConcentric[old_tool][new_tool][toolchange_retract_speed]OLDTOOLNEWTOOL tool H W; prime pillar; ooze shieldhours minute Build Summary  Build time:    Filament length:  mm ( m)  Plastic volume:  mm^3 ( cc)  Plastic weight:  g ( lb)  Material cost: "[install_dir]/gpx" -p -m xyz -u  "[output_filepath]"Beginning G-Code creation...Slicing 3D models...Finding outlines...Looking for overhangs...Calculating infill...Writing toolpaths...Finished creating G-Code instructions.  Estimated print time: %1 G-Code generated by Simplify3D® Version MMM d, yyyy at h:mm:ss AP Settings Summary  G90M82M83[fan_speed_percentage][fan_speed_pwm][bed0_temperature][bed1_temperature][bed2_temperature][bed3_temperature][bed4_temperature][bed5_temperature]#[extruder0_temperature][extruder1_temperature][extruder2_temperature][extruder3_temperature][extruder4_temperature][extruder5_temperature][progress]\d{1,2}\d{1,3} layer , Z = [previous_Z_position][current_Z_position]

 

[Labern 774]

I guess if you were going to do the spiny disc wiper you wouldn't need a motor, you could just make a mechanism similar to the click counters, where the head can push the plunger in as it wipes and as it releases the plunger the disc will rotate a little depending on the gear tooth size.

[media-thumb=5705][/media-thumb]

[nallath 1,120]

I think just moving over a sponge of some sort would have the same effect. It doesn't matter if the disk or the nozzle moves right? I'd even go as far saying that moving the nozzle is probably better, as you'd have less moving parts.

[Labern 774]

Yeah it is better but as Ultiarjan was saying that even though his is a long strip he still has to keep cleaning it during the print as over time the oozing builds up along it. Unless i understood him wrong.

[neotko 1,417]

I think the main problem it's that with the magnetic heads, the actual movement to clean can move and separate the head. The problem also without cleaning specially on the first 5-10mm or so it's that if the dripping builds up it goes and hits the primming tower/ooze shield and that hit can also separate the head. It never happened to me, but when you see the head doing 'clank' and remagnetising then you see that you need keep the head clean. Also the cleanning imo must be done on the strongest angle of each head (ie the opposite to the one holding the stuff, so the reverse hit it's strongly resisted and the had don't looses it's grip)

On my setup the strongest angle it's from x0+ and y- (for right one) and x+ y- for left one.

Since I fixed the Z stuff on simplify3d I'll try to go back to my original idea for a wiper, using the unload sequence and straight x movements. But this time I'll add 4 wipers with less tension (using parts of a silicon cookies mold).

The old version:

New version:

[ultiarjan 1,223]

Yeah it is better but as Ultiarjan was saying that even though his is a long strip he still has to keep cleaning it during the print as over time the oozing builds up along it. Unless i understood him wrong.

 

you understood correct..

Best would be to have the wipe next to the bed so the waste plastic can fall down.. but it requires a bigger rebuild than I'm willing to do atm... my goal is to only make reversable adjustments to my um2....

[neotko 1,417]

Well I might have found a 'good' clean way to pop a cleaner. Ofc this one counts that there will be some processing to avoid the states where it might be 'not on-not off'. Like on the previous posts I said about s3d that the dual extrusion could be 4 processes (2 extruders x2 processes for each extruder). This way to find the optimum safe position to avoid moving the z for cleaning could be easily implemented by a few tests.

The idea it's to use the bed or the cristal as 'level' to push a wiper that has a 5mm shaft with a plastic printed gear. I have absolutely zero idea about how to make a proper gear, and this might need some spring to make a faster return to 90º position... No idea yet, but looks promising, at least on the size & bed space used.

[KevinMakes 22]

 

@neotko

Can I please have your ultimaker2 design...?

 

My design it's for Umo+

https://www.youmagine.com/designs/ultimaker-original-plus-dual-extrusion-magnetic-head-changer

For UM2, ultiarjan:

https://www.youmagine.com/designs/dual-head-ultimaker2

 

Thanks m8

[KevinMakes 22]

@neotko thats interesting indeed, I'll play with it sometime ...

If there was room you could have a rubber disk that rotates after every wipe, then at the bottom of the disk have a collection bin that the bits of plastic get rubbed off and drop into.

 

you could have the spinning done with the original wipe action hook a drive with the x travel and spin the drive with the y action...

and the rubber high temp spatula material would work perfect, its safe up to the 250 deg c range

[neotko 1,417]

Well early beta. As I said previously I have zero gear knowledge

But the concept idea it's solid.

 

Thinka to do:

- Learn to make a solid gear etc

- Solid clamp with no play

[KevinMakes 22]

I love the ideer of using existion movements to articulate a whole new purpose.

[neotko 1,417]

Beta 2 looking very good.

 

Thinka to do:

- Refine position when bed lower than 18 (but it should still work as a wiper on that position)

- Add a small box to get extruded filament

So happy!

[foehnsturm 969]

The housing is "almost" ready.

So this

will soon go into this ...

[/media]