Ultimaker Mark2 - The Smart Dual-Extrusion Upgrade

[conny_g 251]

@foehnsturm,

while reprinting and remounting my head for the gap issue (coupling is redone now, head is next) I discovered a possible weak-point of the coupling, depending on the resulting hole size of the coupling.

If the screw is not a perfectly tight fit in the hole (tight fit: you need to screw the screw down, you can't push it down to the alu part) then the coupling can move by up to 0.25mm (whatever the inaccuracy of the hole might be) at the bottom very easily. On the top it less of an issue as the thumb screws press the coupling to the head there, it can't move there.

I fixed it by applying two layers of 5x5mm duct tape to the coupler, above the hole, see red marking in the picture. That covers the 0.2-something mm of distance, now the screw sits tight in the hole.

Possible fixes in the design:
1) duct tape
2) a small screw in the coupler from the front that presses against the head to avoid the movement.
3) some spring-like extension of the coupler
4) a notch of e.g. 0.3mm that needs to be filed down until carefully until the screw goes through tightly

5) a notch in the screw hole that the screw can easily handle that makes sure the hole is tighened down. Or an oval hole with one side tight, but low force to get the screw in.

 

I think the notch in the screw hole or the oval hole could be the easiest if they survive multiple unmounts/remounts.

 

Edited February 9, 2018 by conny_g

[yyh1002 60]

3 minutes ago, conny_g said:

@foehnsturm,

while reprinting and remounting my head for the gap issue (coupling is redone now, head is next) I discovered a possible weak-point of the coupling, depending on the resulting hole size of the coupling.

If the screw is not a perfectly tight fit in the hole (tight fit: you need to screw the screw down, you can't push it down to the alu part) then the coupling can move by up to 0.25mm (whatever the inaccuracy of the hole might be) at the bottom very easily. On the top it less of an issue as the thumb screws press the coupling to the head there, it can't move there.

I fixed it by applying two layers of 5x5mm duct tape to the coupler, above the hole, see red marking in the picture. That covers the 0.2-something mm of distance, now the screw sits tight in the hole.

The only fix in the design (to avoid the duct tape) is a small screw in the coupler from the front that presses against the head to avoid the movement. Or some spring-like extension of the coupler.

 

My solution was two additional nuts below the coupler

[conny_g 251]

Hi all,

 

I keep having some offset between the two heads, see photo. Somewhere mid print the prime tower looks more white front and left than before. Which means silver moved more to the back. White is extruder 1, silver is extruder 2 (the docked one).
I always had that and decided to redo the head and docking to make sure it's not that. Now that I have reprinted and remounted the new head with much care that the magnets sit perfect I still have that offset issue.

Before I had it from the bottom (though calibration was ok every time), now it's starting halfway through the print.

As the head / docking is now perfect it's not inaccuracy of the head, must be something else like step losses or something.

To summarize the facts:

• measuring the prime tower lower half, higher half: it's 14.8mm in Y direction and 14.85 in X direction in the lower half. It's configured to be 15mm.
• In the upper half - after the offset occurs it's 15.2mm in Y direction and 15.1 in X direction.
• So the delta in size is 0.4mm in Y and 0.3mm in X. Difficult to measure, might have an error of some 0.1mm. 
• The delta in size is the the changed offset of the heads
• it starts somewhere mid print. I can't think of any trigger in the STL model, I think it's random.

What could be the reason?

 

Theory: The docking process requires too much force and causes a step loss.

Thoughts: A step loss would change the head position overall, it would not change the offset between the extruders. So it would cause a layer shift once and then E1 and E2 would stay in sync. Seems unlikely.

 

Theory: There was this rounding issue in head offsets. Could special coordinates for the origin or the target before/after the docking process cause the rounding issue and result in an offset?

Thoughts: Then the offset should change over time, it should sometimes appear, sometimes not. Or it should happen more often than just once. Seems unlikely.

 

Theory: The Mark 2 tweak script, could that cause a change of extruder offset? 
Thoughts: it only removes moves, it does not change any ccordinates. Seems unlikely.

Now I am out of ideas.

One thing I could check is the coordinates in the g-code for the prime tower. If they are the same through alle layers, it's hardware or printer calculation. If it's visible in the g-code, it's something else. I wouldn't know a reason why the g-code would change.

Checked it. Coordinates of the prime tower are the same across the file.

 

Edited February 10, 2018 by conny_g

[conny_g 251]

Redid the extruder offset calibration and that was at max 0.05-0.1mm off, but probably just the remainder of the original calibration.
So that seems rather unchanged.

 

[garaha2012 2]

Hi conny_g,

I am having very similar problems. I thought it was down to movements of the head in x-direction.

I keep printing new couplings to avoid the slight movement of the bottom magnets. 

 

But since I read your findings I think I have something similar in addition:

First I calibrate properly using the well known xy scales for adjustment.

Then I print something and the surprise offset is more the .5mm out in both directions. That is right from the start.

Then I used some calibration model (four checkered square) in the center and calibrated from there.

That give me less offset in the printed model but still and again .3 to .5mm offset.

 

As I said I blamed my mechanical execution of the coupling print but the offset I find is more than the mechanical looseness I have at the bottom magnets.

I cannot offer a solution to you other than saying if there would be some software glitch it would also explain what I find.

[conny_g 251]

19 minutes ago, garaha2012 said:

Hi conny_g,

I am having very similar problems. I thought it was down to movements of the head in x-direction.

[...]

As I said I blamed my mechanical execution of the coupling print but the offset I find is more than the mechanical looseness I have at the bottom magnets.

I cannot offer a solution to you other than saying if there would be some software glitch it would also explain what I find.

Good to know there is others that have this issue!

Currently trying to print the identical object with identical settings to see if the effect is identical. If it is one could see that as a hint that it's more a software issue than a mechanical issue. For mechanical I wouldn't know why it should be identical if the effect occurs due to some "random" event during the print.

[foehnsturm 969]

It's a sudden change, right? Then I would either expect a slicer issue (unlikely) or some kind of of mechanical impact. As you already reasoned, a step loss or something similar would cause both extruders to shift. Weirdly enough it looks very much like the coupling issue you reported above. I'd push some kind of tiny wedges in the gap between the original head and the coupling to make it bombproof and then try again.

[RudydG 62]

I also had this issue again some weeks ago, after it was gone for a couple of months.  I analysed my G-code, everything was perfectly straight! Did the print several times,  slanted!

Next time i encounter this, i think i will try to calibrate the offset thrue the slices settings.

 

 

[conny_g 251]

Interesting.

This time it's coming and going across the print.
The only thing I changed was the X/Y extruder offsets, I tuned the calibration by Y 0.1 and X 0.05.

I'll reprint this with same settings and same g-code now to see if it mechanical or 100% repeated and thus software.

Next time I will tune the offsets by 0.05 different, maybe on just one axis.

I have the feeling that this 0.05mm rounding issue is not gone, maybe it's more complex than just setting an offset of multiples of 0.05.

 

Edited February 11, 2018 by conny_g

[conny_g 251]

Zoom photos of both sides, showing the X-direction effect. In the first marked where it begins and where it ends.

The prime tower is positioned (direction-wise) as it was during the print.
Interesting in this print that only part of the print is affected.

[conny_g 251]

Identical settings, offset different. So it's pretty random. This drives me crazy.

Any ideas what I could do or where to search for the cause?

 

[morfeushav 3]

Hello to all.

 

Has saw someone this ?? It´s the part that we need to set for the Mark2 project to control the fans for the second extruder??. these chinese copy all very fast.

 

 

https://es.aliexpress.com/store/product/Ultimaker-Mark2-dual-print-upgrade-kit/1470565_32827637039.html?spm=a219c.search0104.3.190.5745634bkHFjs2&ws_ab_test=searchweb0_0,searchweb201602_5_10152_10151_10065_10344_10068_10342_10547_10343_10340_10548_10341_10084_10083_10618_10307_10301_10303_5711211_5722315_10313_10059_10184_10534_100031_10103_10626_10625_10624_10623_10622_10621_10620_10142,searchweb201603_25,ppcSwitch_5&algo_expid=3966118b-43ce-4ea2-abd0-aecc204885e1-29&algo_pvid=3966118b-43ce-4ea2-abd0-aecc204885e1&priceBeautifyAB=0

Edited February 13, 2018 by morfeushav

[morfeushav 3]

Hello. I am working to set the Mark2 on my printer. I still having some doubts about its benefits comparing it with the use of the normal dual extruder that come with the ultimaker2 printer. Could you someone let me know which are its improvements and benefits?. As a disavantages, the print area decrease and are more parts for move. Sorry if I do not explain very well. I do not want to offend anyone and nobody gets upset, I only want to have a clear idea about the Mark2 project and its benefits.

 

Another thing 

 

Have someone see this? I do not understand german. In this utimaker2 mod, appart the fact the is cooling by water the hotend, during its functionating the nozzles are  going up and down continually

 

 

 

 

 

Regards, forgive me for all and sorry for bad english.

 

[conny_g 251]

4 hours ago, morfeushav said:

Hello. I am working to set the Mark2 on my printer. I still having some doubts about its benefits comparing it with the use of the normal dual extruder that come with the ultimaker2 printer. Could you someone let me know which are its improvements and benefits?. As a disavantages, the print area decrease and are more parts for move. Sorry if I do not explain very well. I do not want to offend anyone and nobody gets upset, I only want to have a clear idea about the Mark2 project and its benefits.

 

The key is: there is no "normal dual extruder" of the UM2. UM2 never supported dual extrusion, that's why the community projects as Mark2 and the watercooled thing (and others) have taken care of it.

 

4 hours ago, morfeushav said:

Another thing 

 

Have someone see this? I do not understand german. In this utimaker2 mod, appart the fact the is cooling by water the hotend, during its functionating the nozzles are  going up and down continually

 

 

Haven't tried it or followed it more closely on the results in the community. But my guts tell me it work's as well and works fine, but for me the effort of the water cooling is more than it needs, I wouldn't want to deal with water cooling. But the overall effort is probably not too much worse in comparison. So try whatever you like better.

[bob-hepple 59]

that's a lot of kit todo the same job as the Mark two,dont like the idea of water cooling just takes up more desk space

[ultiarjan 1,223]

I can really appreciate the water cooling installation, do wonder about the stability of the nozzle lifting, but the results shared looked very good.

I prefer the Mark2 as it's technically simpler.

[morfeushav 3]

I do not like the water cooling installation either. The mos important for me on the videos is the retraction of the nozzle I would like to understand how it work. 

[garaha2012 2]

Hi morfeushav, 

I understand that the lifting is through the wave spring on each side and the extruded "active" filament presses the relevant nozzle down. 

Once retract the nozzle goes up. That part is quite ingenious in its simplicity.

 

I do not fully understand why this needs water cooling.

 

[yyh1002 60]

6 minutes ago, garaha2012 said:

Hi morfeushav, 

I understand that the lifting is through the wave spring on each side and the extruded "active" filament presses the relevant nozzle down. 

Once retract the nozzle goes up. That part is quite ingenious in its simplicity.

 

I do not fully understand why this needs water cooling.

 

The original UM2 heat sink is probably not efficient enough to cool two nozzles, especially they are connected to the same heat sink. There might be oozing if the inactive nozzle is not properlly cooled. 

[Glomby 0]

I'm currently trying to setup Octoprint for my Ultimaker 2+ with Mark2 Upgrade. The Raspberry Pi is installed and connected to the Ultimaker and the camera is working fine as well. When I print from an sd card I get some temperature data for both print heads and the bed.

 

The only thing that doesn't work is printing with Octoprint itself. Uploading a print job over the network works but when I press print, nothing happens. As far as I understand everything should be setup correctly so I'm wondering if this is a limitation of the Mark2 Firmware. I couldn't find anything regarding this problem in Octorprint related topics, this is why I'm posting here. I wouldn't mind if Octcoprint doesn't work properly with the Mark2 Firmware, since the camera addon is mostly what I need out of it anyway.

 

EDIT: It seems this could be an issue with the gcode flavor setting. Octoprint apparently needs RepRap flavor, but because of the Mark2 FW I'm using Marlin flavor. Does the Mark2 Firmware work with RepRap flavor as well or is this a reason why Octoprint network printing would work with the Mark2 upgrade?

Edited February 19, 2018 by Glomby

[ahoeben 1,955]

What gcode flavor are you using? If you want to print over USB (and OctoPrint connects to your printer over USB), you need to use the "Marlin" gcode flavor.

[Glomby 0]

4 minutes ago, ahoeben said:

What gcode flavor are you using? If you want to print over USB (and OctoPrint connects to your printer over USB), you need to use the "Marlin" gcode flavor.

I'm currently using Marlin. I can also see the print head routes in the Octoprint webinterface under gcode viewer so the gcode seems to be readable for octoprint. I can also move the print head and the bed manually over the octoprint interface, so it also has control over the printer. It basically doesn't start the print after pressing the button.

 

I did also add the Mark2 Cura startscript to the ocotprint start script section, but I think that should only be used if I would use octoprint to slice and not Cura on my PC. Or maybe this shouldn't be added there?

[tinkergnome 927]

@Glomby I can state that i never used something else than Octoprint (resp. OctoPi) with the Mark2. There's no limitation, it works just fine.

I would start with completely empty "GCODE Scripts", you can change it later if you have a need.

 

Are you able to send commands to the printer via the terminal?

Octoprint writes log-files, perhaps you can find some hints there?

[Glomby 0]

12 hours ago, tinkergnome said:

@Glomby I can state that i never used something else than Octoprint (resp. OctoPi) with the Mark2. There's no limitation, it works just fine.

I would start with completely empty "GCODE Scripts", you can change it later if you have a need.

 

Are you able to send commands to the printer via the terminal?

Octoprint writes log-files, perhaps you can find some hints there?

Thank you very much, just knowing it should work helped a lot understanding Octoprint. So it turns out Octoprint indeed always executes the gcode scripts and deleting them did fix the problem. I can send the print directly from Cura to Octoprint now and everything executes as expected. I'm currently testing if it will successfully finish a print job but so far everything looks great. It's not a dual extrusion print though but I don't see a reason why that wouldn't work.

 

I guess the combined Mark2 + Octoprint Upgrade was a success. Thanks everyone!!!

[tinkergnome 927]

6 hours ago, Glomby said:

So it turns out Octoprint indeed always executes the gcode scripts and deleting them did fix the problem.

 

I added two things to the "After print is finished"-script, retract the filament a bit (additional 11 mm as "end-of-print-retraction")
and home all axis.

I think Octoprint does this not automatically  - in opposite to the firmware if you print from sdcard.

Edited February 20, 2018 by tinkergnome