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

[ultiarjan 1,223]

You know these days where anything you do goes wrong ... today is one of those days Was planning to test a nice long dual color print, but did not get anything done.

First i had shifting layers due a a loose pulley, but took me some time to recognize it, thought I messed up the dual extrusion S3D settings.

Then my calibration was off, and I had serious under extrusion in head1, so I aborted and instead of homing the head it crashed into a dock (mystery to me why it did not home straight away..) I then noticed head 1 was of by about 5 mm (nozzle to high). I took apart the whole head again and learned the PLA was seriously warped (pic) and the Hot end isolator was seriously stuck, had to replace it, luckily the olsson block was undamaged and I had a spare isolator.

So now the big question is did the warping occur just by the nr. hours printing, or was it due to the 245/250c PET printing... For now I'm just printing a new one in PLA and see how long it will hold...

Edited May 30, 2015 by Guest

[foehnsturm 969]

What was the bed temperature with PET?

Might have a bigger influence on distortion than the hotend temperature.

[neotko 1,417]

I changed my sequence to fix the movements and it works so seeet. It really never tocuhes the hangar and relies on the magnets doing the job. Also when going to pick the heads instead of a straight movemnt now it's a bit in diagonal so it avoid the margin of error of the parked head. Now I get why foehnsturm used a small metal to push the parked head when moving to the hangar.

I'll focus now on making a clamp for a brush (the ones have metal hairs) so the head gets a nice clean before it goes to the priming box. Also I want to test using a 0.7/0.4 metal cable so the head never push a big blop to the printing area.

[neotko 1,417]

- FINAL DESIGN -

(well at least the first working version) There's room for improvement on the software side and it might need accessories bla bla bla

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

There's a PDF with 8 pages of instructions and stuff for starting gcode, toolchange, etc etc. Enjoy !

Link to the PDF https://www.youmagine.com/documents/25718/download

Edited May 30, 2015 by Guest

[ultiarjan 1,223]

@neotko Nice document! love to see some prints.

I'm also back in business, printed a new left head, just calibrated, all working fine again ...

[ultiarjan 1,223]

 

@swordriff I'm printing dual with the Jet 0.4 now (the ones that look like E3D, not the JET RSB) and they are doing nice sofar, less dripping vs the JET RSB, are these inside shaped the same as the E3D nozzles? Am I correct you stopped selling these in 0.4 ?

Edited May 31, 2015 by Guest

[ultiarjan 1,223]

the newly printed head is holding nice, did a +10 hours print today, clean print, satisfied with the result exept for my lousy calibration... I used cubes with 0.6 layer changes, but for good XY calibration it better to use a few bigger cubes..the thing on the right works better for XY calibration.

http://ultimaker.com/en/community/prints/592-dual-extrusion-pencil-box

Edited May 31, 2015 by Guest

[neotko 1,417]

Mine yesterday when I was adjusting the load/unload sequence got a magnet loose on the clamp (exactly the one I forgot to scratch...). Until today I can't touch the printer

But before that I had one really weird error. I was printing just fine for 30 layers +/- and then the extruder 1 X shifted 2-3mm, the weird think it's that even with the homing sequence it did keep shifting the layer, the extruder 2 didn't shifted but the other did. So what I'll try to check it's if introducing the homing between changes adds an error (it shouldn't..) or if it's just better to remove the homing and use a gcode x/y alignment and not the M218. Because even when the X shifted for Tool 0, the movements to clamp the head didn't changed... It was really weird.

I been thinking about how to add a 'wipe' for this setup, but the options are a limited. If the head size it's 5mm smaller then the head could pass the cristal of the hotbed and reach the aluminum 'hole' in the end of the machine, there might be a nice place to insert a wipe and the drips would fall on a non printable area. But I don't think I have strength to change all again. The second option could be to change the fancap so they are 'higher' so a small wipe it's placed on the same area (attached to the wood not to the bed ofc), I'm thinking of this because it's easier to change the fan design than the full head. The third option it's a easier mix of the 2. Place a wipe (with the height adjustable to align it with the bed first layers) and move the Z to make the wipe 'reach-able'. But this means that there should be a small portion of the bed unusable, so to avoid this the best should be to place 2 wipes on the end of the machine, attached to the wood and with a height adjustable mechanism, this way the wipes would be on the area that the 2 don't collide so the 2 extruder print area should be intact... Just brainstorming...

Edited June 1, 2015 by Guest

[neotko 1,417]

E ) Spare weird x displacements during printing.

I've seen that before. Looks like a wire crosstalking issue. No real solution yet.

Long ago in page 10 you listed this X displacements. I have this. Until latter today I can't test new things, but how did you solve it? My theory it's that when getting the heads from a lateral move need's a movement that don't really 'hits' the hangar so it's the head the one that move's to the clamp, and when parking the X/Y need's to be really precise so there's no distortion on the clamp and it don't suffers from alignment. Maybe also a mix of speed to lower this ?

Thanks!

[ultiarjan 1,223]

I been thinking about how to add a 'wipe' for this setup, but the options are a limited

@neotko been thinking of it myself but it basically comes down to the need of more space around the bed. The ability to park the heads next to the bed i.s.o. above it would solve not only this but also the point that for the first layers the parked head is sitting in a pool of PLA almost touching the bed...

I like the cleaning solution with a small box + rubber flap used in the Reprap BCN Sigma ...

 

[neotko 1,417]

I think this might be a good idea. Since when using 2 heads the area where they 'print' it's smaller we could place something to raise the bed (easy Z change) and create 2 wipes attached to the sides of the machine (I think this could be easier on UMO+ since the panel it's open.)

This way also could help to use 2 different nozzles, different Z controlled by software (Simplify3D has a easy Z gcode offset) and might open the way for the original idea of using different tools.

[ultiarjan 1,223]

could be ... think I let you test it first

[neotko 1,417]

Now I can trully say that I have this fully working. It might be summer but Winter is coming.

 

Printed at 65mm/s 0.15 layer with supercheap pla. 4h print. Zero errors. Sliced with Cura.

I need to align x/y a bit better but today finally had time to watch it print a long 4h. Totally unattended until the prime tower fells. But even so the print didn't fail.

Printed design:

https://www.youmagine.com/designs/robot-ice-winter-is-coming

Edited June 4, 2015 by Guest

[Dim3nsioneer 557]

Really good print quality... as far as I can judge from the pictures. Congratulations! Dual head seems to be the way to go for proper dual prints.

On the 3D printshow in London I had a look on the BCN sigma printer which works the same way. The guys there could even leave the ooze shield away as they have a rubber blade to strip off oozed material at the start of each extruder sequence. Maybe that's something to make this setup perfect?

[nallath 1,120]

Very nice result! Shame I can only like it once :(

[foehnsturm 969]

Congratulations! From the very beginning I had the feeling that this idea has a lot of potential.

Some z-space at the parking positions to implement a wiping mechanism would really come in handy now.

[ultiarjan 1,223]

For people in the area. I'm planning to demonstrate my setup on the next Ultimaker evening in Utrecht... The 15th this month.

[Dim3nsioneer 557]

Very nice result! Shame I can only like it once :(

 

Ha, I liked it twice...

[pm_dude 27]

Wow great work!

Cant wait to get a 3rd UM and put it together...

Z offset per nozzle could be done in a plugin. Just watch for T1 changes and insert G92 Z0.01 if you want a 0.01 offset. Then on the next T0 insert a G92 Z0.0 and your back to default height. Or at least it should be something similar to this.

If you use different nozzle size you'll need to change the slicer if that is not yet supported.

Edited June 4, 2015 by Guest

[neotko 1,417]

Wow great work!

Cant wait to get a 3rd UM and put it together...

Z offset per nozzle could be done in a plugin. Just watch for T1 changes and insert G92 Z0.01 if you want a 0.01 offset. Then on the next T0 insert a G92 Z0.0 and your back to default height. Or at least it should be something similar to this.

If you use different nozzle size you'll need to change the slicer if that is not yet supported.

 

The problem it's physically the nozzles will hit the borosilicate/template glass.

Ideally a new glass on top could make a higher Z to have 4mm of margin but I have no idea about how good the heat will go through 2 glasses. Easyer would be to place a different glass but the metal clamps of the bed only give 1cm of margin to make it smaller from the sides. That might work but for real use it might need to remove the metal clamps and use the reprap kind with a smaller glass. Even so that's only 4mm z gain.

[pm_dude 27]

How about having some ramp upward or at an angle sideways on the ducking? I have no clue if that would actually work.

We could have a custom cut of glass that doesn't go all the way in front. You leave the front clip open and make side clips.

[neotko 1,417]

Yea on my setup, placing a bit of blue tape on the hangar let's you raise 1-2mm the nozzle so it don't hits the bed. I had problems going 3mm but 2mm works. I'm doing a wipe clip, on this photo I was using a wiper blade from a car (3€ on the store) but it's too bulky. Now I'm changing this to use 0.4 silver metal that I have on a roll (it's used to make earrings). But for now I have 'heat' problems on my UMO+ so I need to stop to make room for a fan on the mainboard :( (or maybe it's other think, but noone answered to my post https://ultimaker.com/en/community/view/16256-weird-z-movements)

Anyhow, this it's the beta (bad print cos failed but enough to make a test).

I will try to use this metal:

[macua85 52]

Well I think this is it for me. I know I've posted some setups i've tried before, but I think this one will work. I've modelled it fully to ensure it all fits. So with mine, the heads are stored at the back of the machine. Now printing the main holder, hopefully I can share some results tomorrow.

[pm_dude 27]

Might be a minor thing but there is more swing on the front of the build plate so having the head use the front of the printer seems like a better idea than the back. Again that is a minor thing.

[ultiarjan 1,223]

@pm_dude Indeed, was my thinking as well, I prefer to lose some print area on the front. It's also easier as you (almost) can keep the normal homing position.

@macua85 Is that a dual E3D-v6 you'r planning to use ?