Good afternoon one and all i am nearly done building my CNC and waiting on some more EQ to install, openbuilds of course yesterday i received the Black box, IOT and PS, I decided to play with it a little with just motor and i'm pretty excited to start i played with the jogging a little and when I put in 10" on the jogging interface it looked a little off, so i put a tape measure on the rail and did it again, it went about 12-13" instead of 10" I started thinking (I know i shouldn't but i cant help myself), since there is no feedback system for telling the control box how far its moved, then it must rely on the motor position and rotation and the pitch of what ever drive system (screw or belt), then i thought (told you its dangerous) there must be a setup parameter to enter a screw pitch or other info so that the control knows how far to go just based on the movement of the motor and how far 1 revolution or a fraction of it moves the gantry. Are the screw pitches all the same for these ball screws? or could i be enlightened? Thanks in advance Tim H
Its in the grbl settings. If you're using OB Control, go into the Wizards menu and run calibration wizard.
You'll want to do it for all axis once your machine is together. Should be one of the very first things you do.
good! That is exactly what is happening. without feedback it is called an 'open loop' system. GRBL (the firmware in the BlackBox, read all the Wiki, knowledge is power) outputs pulses to drive the motors and has faith that they moved as instructed. The internal math is simple. The settings you have to calibrate are 'steps per millimeter'. We have a motor of say 200 steps per revolution. Multiply that by the microstepping, 8x is standard on the BlackBox. So 200x8 = 1600 steps per revolution. We have a leadscrew, lets say a stock 8x8 OB leadscrew, so 8mm per revolution. So the theory says that steps per mmm would be 1600/8 = 200 steps per mm Now the controller can do some simple math: Gcode says move from X0 to X25, how many steps to do? (25-0) * 200 = 5000 steps for that distance. It does this for each axis and then does a timed series of pulses so the 3 axes all move together to get to the desired end point at the same time. If a motor fails to move as instructed, oh well, bad luck, User can adjust speed and acceleration down and try again. However, even industrial superprecision ground ballscrews have some error! So we use the calibration wizard to fine tune our step/mm values. Details of what the wizard does are here.
on the topic of 'accuracy' we need to carefully separate 'precision' and 'repeatability'. Precision is where instructing a 100mm move results in 100mm of movement. Repeatability means that we get the same movement everytime, it may not be precise, but it is repeatable. Then we consider the odd magnetic effects inside the motors, how that relates to microstepping, and how the cutting forces affect the positioning. In a nutshell, we are assuming that telling a motor to move 127 steps out of those 1600 steps per rev will result in precisely 28.575 degrees of rotation. In an unloaded motor that is probably what we will get, most of the time (depends how well the motor is made). When we load the motor we don't get such precision, it will stop either before (our load is trying to stop the motor) or after (our load is encouraging the motor to move) the desired position. Does it matter? Not on our home hobby machines (-: Not even on many industrial machines. A friend has a 1.5x2.5 meter machine that can chew through hardwood at 30000rpm but it is nothing more than an enlarged LEAD1010 with big stepper motors and rack and pinion drive and a vacuum table. (google 'RF1325 CNC' for pictures). On machines where it does matter they don't use stepper motors, they do use high precision scales for feedback, and they also have tables of correction factors in the software to correct the errors in the scales, and they temperature stabilize everything, even pumping coolant through the leadscrews so they don't get longer with heat. One also pays 7 digit prices for the machine <-: However, I must warn you that delving into these details can distract you from the sheer joy of making things (-: so long as part A fits part B should we care that they are slightly longer or shorter than the drawing stated? I say no.