@Ryan_Carlyle homing should make the effector losing position a non-issue. Miniscule amounts of friction or counter-weights or fancy springs should prevent falling. Not like I really personally care about what leadscrew is used though.
@NathanielStenzel That’s true on Deltas but not i3 type Cartesian Z stages. If you have two Z motors wired in parallel, and they get out of square, you have to manually fix that.
@funinthefalls , delta or cartesian? Or other? I just now noticed you never specified.
@Ryan_Carlyle nice point…so I asked for clarification on delta vs cartesian vs other. Hahaha
@Ryan_Carlyle “A 1:4 overdrive gearbox makes the load look 16x heavier to the motor.”
Yes but it if 1:4 makes it move the same speed as the belt solution, then the load was only the same fraction to start with compared to the belt. The math will always work out such that the motor will see the same load if the same RPM on the motor results in the same speed on the carriage. This follows directly from newtons laws. Otherwise you suddenly have power going nowhere.
Remember power equals RPM times torque. Unless you are claiming it suddenly requires much more power to move the carriage, the torque will be the same no matter what solution is used to move the carriage, if the same RPM results in the same speed on the carriage.
I imagine the play in the gears will have a minimal effect as long as there are not to many gears. Just mentioning it for completeness.
@Baldur_Norddahl This is a dynamic inertia-acceleration system, which is a lot less convenient to do energy/work calculations on than a steady-state friction load or gravity load… “power in = power out” is only accurate when you add up ALL the masses being accelerated, not just the carriage load. The main massy bits in this scenario are the stepper rotor, screw, and carriage. You’re proposing to turn the screw into a de facto flywheel, and that has a meaningful impact on how effectively the motor can move the load.
You need to do a “reflected inertia” calculation here to determine the real behavior, not just an in=out power balance. Motors produce maximum load acceleration (ie have the easiest time driving the load) when the reflected inertia of the load+drivetrain equals the rotor inertia.
- Calculate the reflected inertia of the carriage through the screw’s mechanical advantage (this will be very small)
- Add the carriage reflected moment of inertia to the screw moment of inertia
- Calculate the reflected carriage+screw inertia through the gearbox mechanical advantage (this is where it gets huge)
- Compare the screw+carriage+gearbox moment of inertia to the same drivetrain with only the belt/pulley… if the belt drivetrain reflected inertia is closer to the rotor moment of inertia than the S+C+G reflected inertia is, then you’ve made the drivetrain perform worse and the motor is using too much energy accelerating the screw.
The inertia of a screw less than 10 mm diameter is going to be very small.
@Baldur_Norddahl It’s a lot of inertia if you’re trying to jerk a 600 RPM speed delta in a few hundredths of a second at a corner.
@Ryan_Carlyle Try to do the calculation. You have 100 grams of steel 3-4 mm from center of rotation compared to 300 gram of carriage that you claim to be able to stop in a few hundredths of a second. I am claiming that carriage is not stopping that quickly without throwing the printer to the floor. Incidentally stopping the screw will not cause unwanted stress on the frame in the same way as stopping the carriage does.
The firmware will handle the problem by controlling acceleration. It might well be that maximum acceleration is limited by the torque of the motor in a lead screw setup while it is limited by the stiffness (or lack thereof) in the belt in a belt setup. I have neither done the calculations nor build a lead screw printer, as proposed, so I do not know. All I wanted to point out was that the speed of the system is not limited by max RPM of the motor because you can always change the ratio between RPM and torque with gearing.
@Baldur_Norddahl I actually have a simulator spreadsheet for this; I’ll run the numbers when I get a chance soon. And yes, 3d printers routinely put a several g’s of acceleration into belt-driven carriages during corner jerks, and they don’t fling themselves off tables, because the rest of the printer is a lot heavier than the carriage. In comparison, typical screw driven stages can only handle 10-20% of the acceleration and jerk compared to an equivalent belt stage. (Check your printer’s accel/jerk settings…) Which is a strong reason to avoid screws for XY motion: you lose corner quality from blobbing due to corner slowdowns.
The simple fact is, an overdrive gearbox + leadscrew is going to have more mass and lower performance than a belt, along with higher cost and engineering complexity. You can easily put yourself in a corner where changing the gearing to fix one problem (RPM limits) creates a new problem (excessive reflected inertia) and there’s no good combination of parameters with adequate performance.