Awesome week
Posted: Mon Apr 06, 2015 3:18 pm
Finally had some time to work on the printer again. The main thing of interest is how Z height compensation can improve the performance of a machine experiencing SEVERE backlash. I mean, every single component out of whack by a huge degree. Belt tension all over the place, cheapskate cams all over the place, plastic arm mounting holes varying in diameter by a bunch (due to jackhammering gap-fill in cura), on top of poor bed leveling due to the above, and somehow, with my multi-directional z-height compensation algorithm, I was still able to create useful stuff. It seems, though, that if you're looking for compensation algorithms, you're probably just trying to avoid reading the manual and following the instructions extremely well. (Or in the case of Orion owners, you don't realize you need to be reading the Rostock manual instead)
I decided to address each issue one by one, so I could get closer to the level of awesome exhibited by mhackney.
First up... Belt tension. I have a really sensitive set of hands, and I adjusted the tension on each belt to be a little bit but not too much, as per the rostock directions, with an emphasis on the same tension on each tower's belt. Unfortunately, the existing hardware on each of the top pulleys/supports were pretty tight from the factory, I suppose. There were substantial indentations in the melamine which necessitated some giant washers to allow tension other than what came with the machine.
Next up.. Cheapskate cams. Similar deal for them. Prior to adjustment, the top bearing cam on the Y tower was WAYYYY tighter than any other, and the bottom one was so loose the bearing didn't even ride in the groove. X tower was loose all around, and had plenty of axial slop. Z tower wasn't so bad, so I used it as a model for the others. They're now pretty close to identical, and have almost zero play in them. The play that is there is from flex in the materials.
Third, the arms. I have another set of the blue plastic arms, but I wanted something I could be really sure about, so I bought some carbon fiber arms from Trick Laser. Unfortunately, Rostock arms got sent to me, and in my haste I installed them on the Orion without realizing that. Pros: They're solid as a rock. Cons: About half the print height is consumed. With adjustments in EEPROM, they worked fine anyway. Brian sent me some replacements and a return envelope that arrived this weekend. Now things are as they should be. Thanks, Brian.
Last, but probably the furthest from least, an E3D-V6 hotend.
mhackney mentioned in the guide he made that if I could print http://www.thingiverse.com/thing:50505 perfectly, then I had "arrived". With the stock hotend, I would always have a screwy X and Y. With the E3D-V6, Except for retraction adjustments still TODO, it came out perfectly the half dozen or so times I printed it. The letters were perfect, the inner and outer circles were perfect, the joining arms were perfect... With the exception of the occasional blob of plastic dropped on some perimeter while traveling, there wasn't anything really wrong about the print.
Anywho, lesson learned... If you want a Delta printer, get a Rostock MAX, and some carbon fiber arms from Trick Laser. The Orion is awesome, but you're going to spend more money on it and have to do the tuning anyway.
Now to get me some Nylon!
I decided to address each issue one by one, so I could get closer to the level of awesome exhibited by mhackney.
First up... Belt tension. I have a really sensitive set of hands, and I adjusted the tension on each belt to be a little bit but not too much, as per the rostock directions, with an emphasis on the same tension on each tower's belt. Unfortunately, the existing hardware on each of the top pulleys/supports were pretty tight from the factory, I suppose. There were substantial indentations in the melamine which necessitated some giant washers to allow tension other than what came with the machine.
Next up.. Cheapskate cams. Similar deal for them. Prior to adjustment, the top bearing cam on the Y tower was WAYYYY tighter than any other, and the bottom one was so loose the bearing didn't even ride in the groove. X tower was loose all around, and had plenty of axial slop. Z tower wasn't so bad, so I used it as a model for the others. They're now pretty close to identical, and have almost zero play in them. The play that is there is from flex in the materials.
Third, the arms. I have another set of the blue plastic arms, but I wanted something I could be really sure about, so I bought some carbon fiber arms from Trick Laser. Unfortunately, Rostock arms got sent to me, and in my haste I installed them on the Orion without realizing that. Pros: They're solid as a rock. Cons: About half the print height is consumed. With adjustments in EEPROM, they worked fine anyway. Brian sent me some replacements and a return envelope that arrived this weekend. Now things are as they should be. Thanks, Brian.
Last, but probably the furthest from least, an E3D-V6 hotend.
mhackney mentioned in the guide he made that if I could print http://www.thingiverse.com/thing:50505 perfectly, then I had "arrived". With the stock hotend, I would always have a screwy X and Y. With the E3D-V6, Except for retraction adjustments still TODO, it came out perfectly the half dozen or so times I printed it. The letters were perfect, the inner and outer circles were perfect, the joining arms were perfect... With the exception of the occasional blob of plastic dropped on some perimeter while traveling, there wasn't anything really wrong about the print.
Anywho, lesson learned... If you want a Delta printer, get a Rostock MAX, and some carbon fiber arms from Trick Laser. The Orion is awesome, but you're going to spend more money on it and have to do the tuning anyway.
Now to get me some Nylon!
