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

[SandervG 1,521]

@foehnsturm different nozzle sizes should be easy as it's already supported by Simplify3D, it's just an easy setting to f.e. use a 0.6 nozzle for infill. I decided to wait for Pink Unicorn before putting effort in dual extrusion with CURA.

I do expect (will for sure try later) more fillament dripping during pauze with the larger nozzle...

 

@Ultiarjan @nallath, please correct me if I am wrong but I believe in the first launch Pink Unicorn does not support dual extrusion yet. I would hate for you to loose time waiting for it when you could already start with CURA.

[nallath 1,120]

@Ultiarjan @nallath, please correct me if I am wrong but I believe in the first launch Pink Unicorn does not support dual extrusion yet. I would hate for you to loose time waiting for it when you could already start with CURA.

Wont have anything with dual extrusion. It is a good idea to wait though, as the old Cura will be abandoned in time (But dont fear, we will still provide downloads to it).

So if you want to put a lot of effort in something (which is required for getting dual extrusion 'right'), waiting for PU release is a good idea.

[neotko 1,417]

Well everything seems to be working now, but for this to trully work I need to put full pen springs or find where to buy this size springs on a harder material. Also I need to change the design of the fan area to make room for spacers. At least the holder for the heads works perfectly after doing some last and hopefully final adjustments.

I don't know how many hours I have put into this, can't wait to have it finally working...

[ultiarjan 1,223]

http://www.alcomex.nl/

If you need springs... think this is also the supplier of the d2110 feeder spring of the um2

[neotko 1,417]

Super. They have the od and 5N force thanks a lot!

[neotko 1,417]

Umm I just think I'll have to modify my design all over. After reading about the degradation on the magnets at high temperatures at this web and got me thinking that my very little and strong magnets won't survive a ABS print because the heat.

I being doing some heat captures on my setup, because I'm testing some titanium nozzle at the same time. and the Upper area goes easily to 50-70C when they are on 0 movement. When the printing goes and the first layer has pass (fan starts) they stay at 45-60C. The problem isn't the magnets that are clamp, that area heats very little since it's almost closed by the magnets itself, the problem it's the exposed magnets.

I saw again the magnets that Foehnsturm uses and they are N45SH so they can go without trouble to 150C, that sure will allow for ABS print etc (but the printed material of the hanger won't survive the enclosure if the enclosure goes higher than 100+ C.

I'm going to think about getting 150C magnets (my room it's full of magnets). And probably I'll use some N44H magnet of 2mm that I can fit on my small heat.

So... let's start over XD

Btw, so far 2 weeks printing with one head on my current design (with small modifications on springs etc) and it's printing rock solid, with good quality. Main problems on ump (not um2 since you can use the full aluminum x2 to have a good material) it's that the heat goes up and up and deforms the pla. I have now one head 2mm different just because small heat deformations while doing tests (head not moving and sometimes I did forgot to put the fans on).

I think on the long therm for this mod to work on ump it might need some laser cut wood or something that might keep the heads protected from heat while on standby. I'm going to take some days to think.

Edited May 12, 2015 by Guest

[macua85 52]

How about printing the parts in Colorfabb XT? It has similar properties to ABS but i've found it much easier to print with. Definitely more heat resistant that PLA.

I'm hoping to start up my attempt at this project again soon. Ive decided to swap to direct drive first though, and order the parts to do this so that i don't have to dismantle my UMO hotend. Its too time consuming when I produce and test an iteration that doesn't work!

[neotko 1,417]

Well the change of magnets I think it's mandatory. PLA could work also since I ordered some wood laser that will be pushing the ptfe etc. To change the magnets on my design it's really easy (I just did in 6mins befofe giving some snack to my 2year boy XD). Also changing the size I stay on the 6cm max tall (to use the umo screws) and I have changed the hangar bay so it's taller to provide better stability while parked. The problem/advantage with my design it's that it uses the space to it's maximun. So the magnets to use can't be bigger than 3mm depth. That combined with the 150C factor of magnets it's hard to get thin magnets of this kind but the web foehnsturm used has some that I have just ordered.

Anyway I think I will release the beta design since the basic idea it's easy to scale, the hangar works soo sweet and I just need to finish very minor changes. The only part that can't be printed it's the 4mm laser cut aluminium. And must be on one piece or the design assembly it's a nightmare.

[neotko 1,417]

Tryed the semi-old fan design (design 14) and seems to work fine. Installed some heat shield (it's like a fabric so it's easy to cut) and attached to the upper part to block the heat from getting on the magnet area.

[/media]

This a photo of 2h into the print, the heat outside the nozzle don't leave with the fabric, but for the first layer it's better to have the fan on minimum to be safe and don't get deformation on the pla. Final versión should have to be abs or xt at least.

EDIT: Uploaded it 2 times and this %$&! forum keeps rotating it...

EDIT: Ok finnaly..

Also good stuff, printing my (hopefully) last design for the fans, easyer to print and the overhangs are getting nice air from the simetric fans (one it's deformed because didn't used the fan on minimum after 15mins trying to fix some leaking with my nozzle because bad assembly).

Today I will have the laser wood parts and next week the new 10x5.5x2mm 150C neodinimum magnets (1kilo force each and I'll use 3).

Edited May 15, 2015 by Guest

[foehnsturm 969]

neotko, ultiarjan

Just had a simple idea: Don't make the same mistake as I did. Don't care about z alignment of extruders! Just measure the offset.

We are not moving both extruders over the print at the same time. So we can easily add a z correction movement when a extruder is selected and deselected. Would be even easier if G10 wasn't misused for retraction, but there should be a workaround using other gcode commands.

[neotko 1,417]

I thought of that! But the weight of the head touching the table if they ain't perfectly stable on the hangar. That's also another reason that makes me think the best area yo place the heads it's at the end of the machine. There's more room and if well places the hangar could be made on pourpose so the head it's slightly angled so the head never touches the bed. But with bowden I don't think I'll be able to doit.

But yeah if the z difference it's very small it should be gcoded. And it should not be a big difference. I wonder if using the end of the machine and make it park on a small angle could help use much wider z (and tools).

[ultiarjan 1,223]

In my case the Z adjustment is relatively easy due to the UM2 setup. I just use the normal level procedure and during the procedure I change between heads, basically leveling with head1, and at 1 point (mid back) change it for head 2 and adjust this one by turning the hot end isolator. I don't have much room to play due to the olsson block, but it working great.

[ultiarjan 1,223]

I'm really starting to get better at this after many hours of testing ... the offset correction is very good now, my main issues with under-extrusion where mostly due to too fast retraction causing many rough edges on the filament. I'm still experimenting with various prime tower shapes, atm I'm still taking off some plastic during printing with pliers. I think I can also reduce the "extra extrusion" to correct for dripping (now at 5mm)

This yellow is colorfabb, the gold is Makerpoint... after using a few colors all Makerpoint home brand I used is rather good, the collorfab is difficult, dripping a lot more...

[neotko 1,417]

I had some trouble the past week with the z alignment initial tests, specially when using the springs, if the angle of the head isn't very 90º the head starts to leave marks on the prints, like small extra blobs filament going to the sides of the print. That's why I bough some good quality springs and I'm waiting for the new wood to arrive (the delivery was going to arrive today but no luck since they printed it on the wrong material). On UMO it's not going to be easy, at least on my mod. The problem for going too small it's that the real tension it's done by 3 screws and one extra that need's to be adjusted by a screw attached on the bottom, because if I used the 4 screws I was going to need to make the head 5-7mm bigger. I did some tests with the 'good' springs and they are quite nice, that's also why I did another fan setup, to leave some room for the m3 washers (around 7mm dia). With the washers and the bottom screw everything seems to be ok but I have leave the design on hold until the new magnets arrive. Also the fans will have a small window of cooling for the peek area, to avoid for the head going up (also should help while printing wood afaik to keep the peek cooler).

In a perfect setup the solution for umo stock head I might be using a spring like the one that um2 uses to keep the ptfe inside the peek, but the size of the ptfe that umo/umo+ uses leaves only 1mm of area to push it and I prefer for now not to go that road unless my x4 m3 springs fail to keep everything fit while adjusting the z, just because that road would need more aluminium to keep the strong force of the spring from breaking the wood/pla.

Also, I think that the Z problem it's something that won't be much problem for others, yesterday after breaking one peek that was almost new I finally saw that it was 1mm different, also the 'melting' of the upper pla added some play from a used head to the other that it's just waiting.

[foehnsturm 969]

You both are correct, with bulky printheads there isn't much benefit from a software z height adjustment.

But given the small footprint of a Merlin, E3D hotend etc. and a crossflow fan or fans attached to the holder it's interesting. With parking positions outside the printbed even 10 oder 15mm height difference wouldn't be problem. So you could use a PTFA hotend like Merlin which is best for PLA and a full-metal type for higher temp materials within the same print even if they are not the same length at all.

[ultiarjan 1,223]

Little demo video ... bit annoying free music ...

**&%$$^^^%$& how the hell do I embed a video here ?? #$$@@$#% the button seems not to work ..

 

Edited May 17, 2015 by Guest

[ultiarjan 1,223]

=== PROJECT SUMMARY ===

Here's a summary of the project for UM2, to make it a bit easier for you if you want to try this yourself. Let me know if I missed anything. I'll put the design files on YouMagine.

Mechanically this hack is simple, for the first head you don't need to touch the electronics, just dismantle the old head remove the black plastic parts and put together on the magnetic head. You'll need to glue magnets in the pickup and head parts, pay attention to the polarity of the magnets. I used superglue (Cyanoacrylate), after cleaning the magnets with alcohol.

I got my magnets at "magnetenspecialist.nl" blockmagnets of 10x5x4mm (quality N50 whatever that means..)

For the 2th head I took the easy option, just connected the PT100, heater and 2th extruder motor to the Ultimaker board and powered the 5 volt hot end fan from an external PSU and the 2x12Volt=24V side fans to an external controllable PSU.

I'm planning on upgrading the "docking stations" to a more "click on - click off" design, but for now the mounting with tape holds up nicely. Glue the 2 parts of a dock together, I used superglue (Cyanoacrylate).

To make some more space on the left of the head you need to change the position of the endstop left-back. You can use this piece for it:

https://www.youmagine.com/designs/adjustable-home-switch-offset-bracket

Firmware (there's probably room for improvement, but this gets it to work):

This is all provided you use the original UM2 components, like the PT100 as sensor etc...

Before you install you need to change a few basics of the UM2 firmware;

- change the startup position from front left to front middle.

- change the head's bed leveling positions.

- activate the 2th extruder.

- I did not alter the material change positions, just don't use this anymore..

You can learn how to change and compile the UM2 firmware here;

http://ultimaker.com/en/community/view/7436-more-information-during-print?page=14#reply-97402

These are the settings I used:

 

Ultilcd2_menu_print.ccp#define HEATUP_POSITION_COMMAND "G1 F12000 X100 Y45"UltiLCD2_menu_first_run.cpp#define BED_CENTER_ADJUST_X (X_MAX_POS/2)#define BED_CENTER_ADJUST_Y (Y_MAX_LENGTH - 10)#define BED_LEFT_ADJUST_X 10#define BED_LEFT_ADJUST_Y 90#define BED_RIGHT_ADJUST_X 180#define BED_RIGHT_ADJUST_Y 90Configuration.h#define EXTRUDERS 2#define TEMP_SENSOR_2 20Configuration_adv.h#define EXTRUDERS 2

 

Leveling:

To level the bed I do this:

- start the UM2 leveling procedure (with the piece of paper) with the 1st head mounted and the 2th docked.

- do the first rough leveling (till 1mm from bed)

- level the 1st head only at the back position, with the piece of paper.

- mount the 2th extruder, level with a piece of paper.... but DO NOT LEVEL using the printer knob but instead turn the "hot end

isolator" do this only while the head is warm... this thing is easy to damage.. I'll advice to keep a spare one in stock...

- mount back the 1st extruder and finish the leveling on the first 2 positions.

**In case you use 2x olsson block (www.3dsolex.com) make sure the block  does not touch the metal fan shroud, if it does, go 1 step back and change the height of the 1st extruder head and start the leveling again)

Slicing:

Now your ready to go, but first You need to calibrate the G-code movements to precisely move the head on the docking station. I use the direct PC printer controls of "Simplify3D" to find the correct coordinates, but there are other options.

Here's my Simplify3D tool change G-code;

 

;*;***************************************{IF NEWTOOL=0};change to Tool Zero-Left{IF NEWTOOL=1};change to Tool One-Right;***************************************;{IF NEWTOOL=1}G0 F9000 X8.6 Y160 ; back to start T-zero - left{IF NEWTOOL=1}G0 F9000 X8.6 Y49 ; move front park T-zero - left{IF NEWTOOL=1}G0 F9000 X182.2 Y49 ; move to right  pick T-one{IF NEWTOOL=1}G0 F9000 X182.2 Y160 ;move T-one to the back - right{IF NEWTOOL=0}G0 F9000 X182.2 Y160 ;move T-one to the back - right{IF NEWTOOL=0}G0 F9000 X182.2 Y49 ; move to right  park T-one{IF NEWTOOL=0}G0 F9000 X8.6 Y49 ; move front pick T-zero - left{IF NEWTOOL=0}G0 F9000 X8.6 Y160 ; back to start T-zero - left;***************************************;*

 

And here my startup code;

 

G28 ; home all axesG1 X110 Y45 F3000 ; bring extruder to the frontG92 E0 ; zero the extruded lengthG1 Z30 ; lowerG1 E50 F225 ; purge nozzle with 50mm of filamentG1 X80 Y60 F3000 ; tikkie opzij om niet in de prut te landenG92 E0 ; zero the extruded length again

 

You'll have to figure out Simplify3D yourself if you want to use it, but some further tips:

- in "Layer" use "Choose start position point closed to specific location" to start from the back (prime towers)

- in "Layer" flag the tab "Print islands sequentially without optimization, to make sure it's starts with the custom made prime tower.

-set the tool change retraction in the advance tab, I use 20 mm and no correction for dripping (This is up to the prime tower)

Obviously you can use another slicer and "find-replace" the tool change code in the G-code with a text editor.

Calibration;

You need to set the XY offset in the slicer you use. Start with a rough estimate, then print calibration pieces to refine.

I started with the D shape at 0.3 layers, and then went to the other one at 0.15 layers.

My curent ofset is: X 44,3  Y-0,9

You'll find the calibration pieces I used here:

https://www.youmagine.com/designs/dual-extrusion-calibration-tests

The pieces have layers of 0.6 mm, so you can use them with layer hights of f.e. 0.3 / 0.15 or 0.1

And as extra info:

My feeder design is here, to print the 2th one mirror the body in your slicer;

https://www.youmagine.com/designs/feeder-for-ultimaker-2-rail-system

The extruder motor fan bracket I use on the back is here;

https://www.youmagine.com/designs/fanbracket-for-u2rs

Or just search "U2RS"to find all related parts;

https://www.youmagine.com/search?utf8=%E2%9C%93&search_objects%5B%5D=designs&search_objects%5B%5D=collections&search=U2RS

If you want to use original UM2 parts for this project, contact UM support to order parts. You'll find the partnumbers of the stuff you need here ("download zip" button)

https://github.com/Ultimaker/Ultimaker2

Edited May 25, 2015 by Guest

[neotko 1,417]

Superb! I can't wait to upload mine. Congrats!

Just a note N50 magnet, the numbers mean the strength. The number means more powerfull. I mean, the maximum magnet force it's N52. But also the higher the strength the easier it's for them to loose magnet power over time because they support less ºC. The loss of magnet power it's immediate and not recuperable. So your mod might work forever or not, it just depends on what's the real minimum magnet power needed to handle the head.

Standard N grades, without some letters after the number, will start to loose power at 80C, and the higher the heat the more power they loose forever.

Some shops, increase their 'power' rounding the numbers. I used

https://www.kjmagnetics.com/calculator.asp?calcType=block

To get a real calculation.

The grades of the magnets (from that website) The second term it's the Curie temp, that's the point of instant zero magnet power. But everytime the barrier it's cross the magnet won't recover their power.

Grade Max Temp Curie Temp

N 176°F (80°C) 590°F (310°C)

NM 212°F (100°C) 644°F (340°C)

NH 248°F (120°C) 644°F (340°C)

NSH 302°F (150°C) 644°F (340°C)

NUH 356°F (180°C) 662°F (350°C)

NEH 392°F (200°C) 662°F (350°C)

I bough (not the first time and I'm still waiting for them to arrive) N45SH that can support 150C, just in case I wan't to print ABS or XT and don't get any magnet loose. But even with NH it should be more than ok.

[neotko 1,417]

One question, you could connect the other fans to the led pwm, it's the same electronics. And using the 5v minifan in parallel shouldn't be a problem I think.

[ultiarjan 1,223]

 

@neotko do you know if it matters for the magnets durability if you leave the head on or off the pickup while not in use ?

[ultiarjan 1,223]

And another calibration tip, use your phone to magnify the printed piece for precise measurement ...

[neotko 1,417]

@neotko do you know if it matters for the magnets durability if you leave the head on or off the pickup while not in use ?

On my camenra tests the outside magnets that are not inuse go up to 60C (the ones that are on the air from head and main changer). Ofc they go to 50 when fans are on. That why I think it's important. For pla I don't think it will matter but on stuff that needs 100C/+ bed the heat will make aome dmg. And some of that materials don't like fans. In your case it won't matter for the pla since you must have the 5V fan all the time but in theory the heat outaide ahould easyly reach 70-80 ambient. My heat shield fabric (1 page ago I think Inposted it) cut's arround 30C of that and while printing pla the upper area stays at 40C. What did worry me about N50 magnets (I'm using some right now until the new pne arrive) it's that they are too close to the margin of error.

I think in my case with the extra lasercut wood and heat shield I should be able to keep it ok.

So imo your serup should work fine but the magnets exposed to air might get dmg overtime if you print 100C bed materials.

[foehnsturm 969]

@Ultiarjan,

Great summary! So good to have well-organized and conscientious people like you and neotko working on this project.

It looks like you perfectly realized the tool changer idea on the UM2. What remains to be addressed is oozing. While we already benefit from the fact that oozing nozzles cant't do anything harmful to the print, the required amount of priming to fight underextrusion more or less successfully is high. If the standard hotend behaves similar, what I assume, I perfectly understand why UM ditched dual extrusion ...

With a retraction of 20mm no additional filament should melt during idle times. The already liquified one just drains. So far I see no way to prevent this reliably. One could live with that if the following priming process would safeguard consistent extrusion immediately thereafter. But this seems to be difficult.

[pm_dude 27]

Just add a wipe tower for each nozzle. When you attach that print head you then print the layer of a cylinder and retract. That will also help stabilizing the material temp in there before going to the real piece.

[foehnsturm 969]

I did that and it works quite well. However it's surprising to me how much extrusion is needed for certain filaments to stabilize. E.g. I almost failed with dual extrusion and Filaflex which requires a crazy amount of priming after idle times of more than a few seconds.