Jump to content


  • Content Count

  • Joined

  • Last visited

  • Days Won


yyh1002 last won the day on April 3 2019

yyh1002 had the most liked content!

Community Reputation

54 Excellent


Personal Information

  • Country

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Hi Anzen, Sorry for the late reply. There is a critical bug for Cura 4.4.1, which is exactly what you found. It cannot generate correct tool number in start code for dual prints. They fixed the bug in 4.5 beta. You can use 4.5 beta now. But I think the first layer path of 4.3-4.5 is not as clean as 4.2.1. I uploaded new profile for Cura 4.5 to Github. And also an updated version of firmware. This new firmware will prevent any tool change when head position is unknown. It avoids the problem of bad start code movement caused by the first "T0" or "T1" generated by Cura which makes the printer to think it's holding a wrong head. It works together with a new start code that automatically select and purge the nozzle used in a print. The machine "DXU" should be used for single nozzle print, either left or right nozzle. The machine "DXU Dual" should be used for dual nozzle print. The difference between the two is start code.
  2. I haven't checked the new embedded Mark2 profile. You will also need to change platform shift and print area. I'm not sure if Mark2 has the material adhesion option enabled. But for me the old DXU profile works directly with Cure 4.3, so I didn't find it necessary to do a new version. Did it not work for you? I use 6.5mm, 25mm/s, same as UM2+ default.
  3. Not really. There needs to be about 1-2mm spacing between the top of heat block and hex flange on heat break. The block is loose before install the nozzle, and you can turn it to proper direction so it doesn't touch anything. When nozzle comes in, it will lock everything tight.
  4. Thanks. Good observation. On the front housing, there are two bores to restrain the hotend for xy precision, one is above the heatsink and spring, the other is around the bowden coupler. Both bores have teeth around them so it's easier for the hotend to slide. I made the tolerance quite tight and use a drill to spin the hotend inside the housing to run-in. It's metal vs plastic so they fit quite well without being all fixed. I agree that it's not as precise as UM3's tapered socket solution, technically. But I think it's good enough for 3D printing purpose if fitted well, just like the brass bushing on XY rails. I would like to go with more precise solution if I have easy access to machines. The PI insulator does go between the nozzle and bowden tube. The bowden tube coming with hotend is intended to be used without PI insulator, so it needs to be trimmed slightly shorter. But the total length of tube+PI should be slightly longer (about 0.5mm) than the shaft to keep the bowden tube in compression so it doesn't get squeezed shorter by hot filament over time. You need to make sure the nozzle is right up against the heat break, metal to metal. Heat the block to at least 160 degrees before install. And hold the block firmly when screw nozzle in. If there is cold plastic, debris or air gap between the nozzle and heat break, it will leak.
  5. For example, in the extruder offset menu, say the detected z offset is 1.90. Feel the flush point on the z calibration print with finger. If the +0.04 section is where the blue is flush with yellow, it means nozzle 2 is 0.04mm lower than ideal. In extruder offset menu, increasing z offset will raise nozzle 2 (this is where it's confusing). Therefore, changing z offset to 1.94 will make the two nozzles level.
  6. There is advantage of doing your three step levelling. Because the standard bed levelling detects corners of the bed with nozzle 1, while the nozzle offset detects with center of the bed, there might be discrepancy due to glass not being perfectly flat. However I think the nozzle 1 z offset levelling will overwrite you three point levelling. But IMO the z offset detection is too coarse anyway, i.e. moving increment by turning dial is too much. So what I do to calibrate the z offset is: 1. Do three point levelling with nozzle 1 2. Swap to nozzle 2, do z offset levelling, get a rough offset 3. Print the z calibration model with 0.02 layer height to fine tune the z offset, and update the value in menu. Once the z offset between 1 and 2 is calibrated, you shouldn't need to redo z offset levelling in the dual print menu unless you did something to throw away the relative position between the two nozzles, such as changing nozzles. The z offset is very valuable, it takes some effort to get it perfect, it shouldn't be affected by bed levelling. In you are only changing beds, you only need to do the standard bed levelling with nozzle 1. Both nozzle 1 and 2 should be good to go.
  7. The hardware will constrain you to have a rough nozzle offset of x19mm, y0mm, and z2mm. It then gets fine tuned by calibration models. The calibrated offset, say x19.03, y-0.06, z1.94, is manually set on the LCD manu to firmware. Basically, the dual print Gcode from Cura gives absolute coordinates, without any offset between the two nozzles. The firmware will step in and shift nozzle 2 relative to nozzle 1 according to the offset values.
  8. That’s not how I find it works. I think the XY offset in cura is only for reference to see available print area. It makes no change to gcode coordinates, at least for Ultimaker 2+ machine type. Only the firmware Mark2 offset will take effect at the end. I set rough XY offset in cura and real offset in firmware and it works without problem. @Whom you can take a look at the cura machine for DXU to see how to get the dynamic print area working. DXU has a X offset of 19mm. I changed the X dim of printer from 220(?) to 239 to be able to shift the area for both nozzle 1 and 2. I think I also changed the buildplate shift in json to re-center the larger print area. DXU is based on Mark2 firmware and the cura offset has been working well for us. https://github.com/yyh1002/DXU
  9. https://github.com/TinkerGnome/Ultimaker2Marlin/tree/Mark2 这里是mark2固件源代码
  10. Sorry I didn't see your comment earlier. You can use 2*4*1.5 PI insulator and the original 2+ feeder with 2*6.35 tube with DXU. In the BOM I try to list a generic link where possible. Only when the parts are very rear, I specifies the Taobao seller. All the parts from the taobao seller are very good quality, you can absolutely by all possible parts from them. I have the same question as burdickjp. Why do you need to replace the original heat break with the all threaded one?
  11. I asked the seller. PI insulator is still available. The stock from now on is 2*4*1.5, which also works with DXU.
  12. Thanks. Software side it works basically the same as Mark 2, hence similar CURA and E3D profiles. Nice post. The 1.75 hotend runs with a custom 6.35 OD 2 ID PTFE bowden tube, sold by the same Taobao seller. The major advantage of this PTFE hotend is that there is pretty much no chance of nozzle jam or clog for PLA even with frequent long retractions. With the PI insulator ring applied, the PTFE last much longer and can also handle high temp materials. As for using E3D heatbreak, the current heatsink probably won't be able to cool it down properly because the surface area is much smaller. You can consider trap a piece of 6.35 OD 3.2 ID or 5 OD 3 ID PTFE tube inside the heatsink. I think the 4*3 PTFE would deteriorate pretty quick due to thin wall thickness, and it is also too soft to hold up to the compression from filament. Making 1.75 and 2.85 hotend components swappable isn't all that necessary. You would need to make compromise on both sides to achieve that. I think it would be better to make two entire printheads with 12pin connector for quick swapping. That would be much less hustle, much more reliable and cost very little.
  13. Misalignment is one factor. IMO it is also because the short bearing is off center and the dock apply quite some force to the head during nozzle change. The force transfers to lateral load at the bearing. Looking at example applications of LM6 bearings, a single short bearing for one axis with such a thin rod really seem to be pushing the limit. Ultimaker did a good job getting it to work with their quality control. For individual tinkerers I would try to avoid this situation.
  14. You can totally try the existing version first. Short bearing should last at least several months. For the blower fan, 3015 works as good as 3510 thanks to more accurate duct aiming the nozzle and better fan shape.
  • Create New...

Important Information

Welcome to the Ultimaker Community of 3D printing experts. Visit the following links to read more about our Terms of Use or our Privacy Policy. Thank you!