RostockMini4tw,
Thanks for the info. It's interesting that your problem went away when you switched to a smaller pulley - that should mean more steps per mm movement, which in turn means bigger numbers of steps overall. My immediate thought is that overflow or rounding errors ought to be more likely. I have tried to use Repetier firmware, but it has a separate problem for me. My drive pulleys are not exactly equal diameters (I have measured and calibrated them all), so I use a slightly different steps/mm setting for each. Marlin is happy with this. Repetier allows me to set the values, but then blithely ignores them and just sets all three to the same number when it actually comes to do its calculations. I haven't had the stamina to go through the whole of the Repetier source to try and fix this. It took me long enough (and several increasingly frustrating calibration cycles before looking at the source code) to realise what was happening.
I think you are right that Marlin is the cause of the problem, though. I guess I'll have to dive in and see if I can spot where it's happening. Ugh. Debugging embedded code is horrible.
Thanks for the info. It's interesting that your problem went away when you switched to a smaller pulley - that should mean more steps per mm movement, which in turn means bigger numbers of steps overall. My immediate thought is that overflow or rounding errors ought to be more likely. I have tried to use Repetier firmware, but it has a separate problem for me. My drive pulleys are not exactly equal diameters (I have measured and calibrated them all), so I use a slightly different steps/mm setting for each. Marlin is happy with this. Repetier allows me to set the values, but then blithely ignores them and just sets all three to the same number when it actually comes to do its calculations. I haven't had the stamina to go through the whole of the Repetier source to try and fix this. It took me long enough (and several increasingly frustrating calibration cycles before looking at the source code) to realise what was happening.
I think you are right that Marlin is the cause of the problem, though. I guess I'll have to dive in and see if I can spot where it's happening. Ugh. Debugging embedded code is horrible.
