Separate names with a comma.
Some features disabled for guests. Register Today.
Discussion in '3D printers' started by Keith Davis, Nov 12, 2014.
A compact, enclosed Core XY design, based on off-the-shelf parts.
Keith Davis published a new build:
Read more about this build...
Interesting Z axis. You should build it! I found the extra tuning requirements for a coreXY setup are negligible compared with the overall build process.
Bowden setups do have their downside. I've only every printed ABS and PLA though. Solution? Two printers!
Changing from stiff filament to flex, or even elastic, does require changing nozzle and feed tension against the drive gear. I sell my BldrBot on ebay and each time I've listed it I know of at least one customer I've lost due to inability to print flex because I've gotten queries about it from them. I've spent the last month developing a direct drive extruder that can easily be used for stiff or flex.
Using a Recreus extruder, I'm running PLA at 150 m/sec and ABS up to 200 m/sec with no problems using a direct drive. Bowdens were originally developed for RepRaps with eye-ball tolerance and sketchy frame stability. My question is, does a bowden actually make that much difference on a solid built machine due to the removal of the motor mass. I haven't seen any one claiming blazing speeds because of a bowden, but maybe I just haven't researched it enough.
Likewise, I haven't seen any claims for blazing speed for a CoreXY system. You've been on to CoreXY longer than I have, have you seen any speed or accuracy claims?
What do you recommend for a extruder for flexible filament.
I've never attempted to maximize print speed, I've always gone for quality and worked on making incremental improvements on that front. My longest print was 42hrs and a typical print ~6-10hrs. It would be nice to shave that down, but I have no problem leaving my printer(s) running overnight or while I am at work. House has not burnt down yet...
One of the fastest popular printers out there is the Ultimaker which uses a bowden setup and claims fast print speeds:
"Ultimaker 2 is unmatched with its print speed of up to 300mm/s and 0.02mm layer resolution."
It probably isn't the best for flex filament, but it can be done:
Huge retractions? Not sure.
I would say that the average user has no need for a bowden setup unless you are going for max speeds. I am one of those people that shouldn't have one!
Though if you are, the justification is simple: F=ma (Force = mass x acceleration)
Faster print speeds require higher acceleration, otherwise you won't actually be reaching top speed. Higher acceleration means greater forces proportional to the moving mass. With greater force there is greater deflection in your system and a reduction in print quality. You can combat that with a more rigid system, but that may mean even more mass or higher construction costs. Eventually you may even need to upgrade your motors to handle the increased loading and/or add cooling to them.
This becomes twice the factor if you move to dual hotends.
Why CoreXY? No moving motors, plus it is a cool and interesting movement system. I have seen some very fast CoreXY you tube videos, but none of then were actually printing anything. The latest makerbot is an H-Bot, basically a Core XY that is rigid enough to counter the racking forces so that the belt cross over isn't necessary.
I just submitted a new build, BldrBot Pro, with info on the Recreus extruder I'm using. My design for mounting a Recreus on a Mini-Wheel Plate is also on Thingiverse.
Printing any flex filament is REAL slow. 30mm/s is average w/.5mm nozzle or bigger. But, that's just for flex. If the flex is also elastic, the drive mechanism has to compress the filament beyond it's own elasticity to push it. Recreus FilaFlex is the only truly elastic filament I'm aware of (you can print rubber bands or condoms with it) and they could not print it with any extruder/printer on the market (beginning with an Ultimaker) - they had to build their own direct drive extruder for FilaFlex. Notice the flex bicycle handle bar grip in the video bends and twists, but it cannot stretch.
And that I think would be a major selling point! And, I think your ingenious way of recessing the belts in the V-Slots would also be a great selling point! You might have gathered, I'm not at all interested in building 3D printers that I can't sell. I'm trying to reach a happy medium between speed, accuracy and versatility .
Never heard of that filament, sounds interesting.
Yeah, as a consumer you want everything: fast, accurate, capable of printing all materials, good looking, and cheap!
Hey, did you ever build this? I've been looking for a CoreXY build for a while and think I am going to start from this design. I've mocked up the frame and XY system in Onshape, and will probably assemble that while I brainstorm the Z stage more thoroughly. I'm curious if your belt & rail-based Z worked out... Is the resolution adequate? With what degree steppers?
The model I'm working on is public in Onshape, though their search function seems to be pretty rough so far. It is called "OTS CoreXY v1.0". If anyone wants to collaborate rather than fork, let me know and I will share those permissions to your account.
I'm actually building the next iteration this next month. The Z axis needed a complete rework. But I still plan on using a belt for the Z axis. On a system where the build raises and lowers a belt can be used because the build will drop when the printer is powered off (as opposed to raising the extruder where the extruder will fall when powered off).
Notice in the lower picture that the two 20x20 guides for the Z axis wheels do not extent to the top bar, leaving room for the Y axis mini-V-wheels to pass above them. The three v-wheels on each Z axis are mounted on a standard t-plate with the two back wheel holes drilled out to 7mm for eccentric spacers.
As for Z axis resolution using a belt - it's the same resolution as the X & Y axis. If all three use 20T pulley and GT2 belts, 80 steps = 1mm. So, 0.1 layer height would be 8 steps. I may go with 16T pulleys which are 100 steps per mm (0.1 layer= 10 steps).
This design is quite inexpensive to build (under $400) but should be able to equal an ultimaker for speed and accuracy.
Do you have a link to your project at Onshape?
Frustratingly, despite all of their narrative of the files being 'public', Onshape does not seem to allow direct linking... and their search function seems pretty broken. What is your account? I will share it with you.
Thanks for the update. I will take these changes into consideration. Was there a reason you changed from 20x20 to 20x40 on the gantry? Just a hunch, or some testing/simulation?
Actually, clicking this link while logged in to Onshape should take you there: https://cad.onshape.com/documents/6fe81cf35e2740b3bbde9aa6/w/4391b881b84645609f2f9a87
This design has evolved into a new project openbuilds.com/builds/printair-corexy.2718. There will be no further development of this design.