Originally shared by Walter White
http://www.youtube.com/watch?v=kVPK2ylrCC4
Ho. le. shit. What an amazing calibration routine. Even if it’s a bit long, it does all the things! O_O
Yep. The 626 config is great. Not sure why he’s running his so slow (I’m guessing it’s because of his probe), but if you have a conductive bed, it can run the whole thing real quick and its dead nuts.
It’s an amazing firmware. Just wish it could work with network + glcd… Now I have to keep changing the config before and after leveling. Also wish we could have a probe next to the hotend…
Most of all I hope this fork will stay alive and keep evolving… I’ll be happy to test new versions…
@Alex_Skoruppa Unfortunately, we only merge code we fully understand, or we wouldn’t be able to maintain it and make sure it plays nice with the rest of the firmware. Last commit was from March : https://github.com/626Pilot/Smoothieware/commits/edge so the original author doesn’t really maintain it, which means I’m not sure we can count on him to do that in the main firmware if we were to merge. On top of that the general concensus in the Smoothie community is that appart from a small gain in calibration time ( which doesn’t matter much as Smoothie can save the calibration data, you can calibrate one then not worry about it ), that calibration method doesn’t have any realy advantage over what Smoothie now has ( which is pretty great ). And on top of that I think that method is also very RAM-hungry which would definitely be a big no-no.
@Arthur_Wolf I know that the best way to calibrate a printer is to build it square… But in a delta it’s a bit more complicated. This firmware gives us the extra parameters that fixes these little errors (angle between towers, small tilt effects, leaning towers)… I know it’s RAM hungry… But can you implement something similar (even RichCattell marlin’s fork worked on an 8 bit elec)… I must say that I have tried the official firmware - couldn’t get the printer to print right… After calibration with pilot’s firmware… I can print on the entire bed surface.(0.15 mm first layer - no raft)
Please consider implementing something like this… I don’t care if the calibration will take more time… Just need it to fix all the mechanical small misalignments…
Again, thank you for a really great hardware/firmware combo.
@yos_gu The method is inherently RAM-hungry, there is nothing we can do about that.
Please try the latest Smoothie firmware : the combination of calibration and delta-grid should give you perfect bed adhesion and good prints even if there are problems in your printer that calibration doesn’t take into account.
Also note that for RAM-hungry methods like this and others, Smoothie has a way to probe the bed ( G29 ), report the numbers, which you can then input into a web interface ( like Escher3D’s ), and get your parameters that way ( the browser doesn’t have RAM limitations ).
Many users do this to get the same kind of correction the 626 firmware provides and it works fine.
Also Smoothie is getting a new web interface that will do this automatically ( but in the browser, same thing about RAM )
@Arthur_Wolf
Thank you for your reply, I will try again with the default firmware… Maybe I missed something.
Looking forward to that web interface calibration tool… I don’t mind using a computer for calibration… Used to calibrate repetier firmware with the OpenDact pc app… It took a while but also gave great results.