Control board with the best Delta kinematics smoothieboard or duet ??
I’m not sure smoothie has mesh bed leveling just yet. Dunno about duet
Duet has RepRapFirmware’s awesome auto-calibration routine. Very highly recommended by lots of people over on the Deltabots Google group. Some people say the segmentation-free delta kinematics in RRF helps print quality too.
Both are worlds better than Marlin.
@ThantiK Smoothie delta just got grid levelling, so we now have both grid levelling and calibration for deltas.
I have yet to see a side-by-side comparison of segment-free/with-segment that shows any difference : the math says there shouldn’t be any.
Yes, I agree there shouldn’t be any visible quality difference from segmentation. You can imagine some very high-speed edge cases, but I suspect the ARM Cortex-M3 in the Duet wouldn’t be able to keep up with kinematics at those speeds anyway.
Does segmentation cause the printer to slow down at all?
It sounds like smoothie is perhaps under more active development? Does both do a aceleration calculations on the effector?
I wouldn’t expect segmentation to cause slow-downs. It’s generally doing less work in interrupts, so all in all it’s probably less limiting your maximum speed.
Smoothie does acceleration on the effector, and I would think reprapfirmware does too. I think only Marlin is messing that part up.
Smoothie is under very active development.
If you want to know which is “more” under active development, you could be getting into “subjective” territory : it depends what you value, what is important to you, what to you “counts” as development.
Smoothie likely has the most work going on on documentation ( by far ), there is a lot of work going on on the current version of the firmware ( and it’s been going on for a long while ), and we are working on both a new firmware and a new board
As far as I know reprapfirmware really only has one dev ( but he is very active and very talented ).
Also reprapfirmware has been exploring things Smoothie hasn’t been ( and vice versa ), like the segmentation-free stuff, the fancy calibration routines, etc.
Their web interface is better too : Smoothie has huge amounts of work being done on the webif side of things, but it’s so ambitious nothing “usable” has come out of it yet.
Getting much closer to that though.
Agreed with everything @Arthur_Wolf just said. Both projects are active, doing interesting things, and vastly better than Marlin. It’s mostly a matter of very technical execution differences that most people will never see, and specific feature development priorities.
David Crocker (the main RRF guy Arthur is referring to) is getting some help now from Dan Newman, the author of Sailfish firmware, which I think is helping round out the features and broaden support. For example dc42-RRF now works on Due/RADDS in addition to Duet, and can be auto-updated via uploading a .bin file on the SD card like Smoothie. (One of my favorite Smoothie features tbh.)
Likewise, Smoothie has been working on incorporating some of the nicer delta auto-cal techniques and PanelDue support (which is David Crocker’s really nice touchscreen for Duet/RRF).
Gotta give respect to both projects for continuous improvement without letting the codebase decay like Marlin’s did.
If I am understanding the technical side correctly, I think the main benefit to segmentation-free kinematics is in memory management for the path planner. Repetier, for example, can end up with a massive motion queue with all the delta subsegments stored for the next 32 gcode commands. RRF saves a bunch of memory (at the cost of interrupt-level computation time) because it only needs one entry in the planner queue per gcode command. They’re running similar acceleration code on the same processor, but loading the MCU in different ways.