position feedback

[Greg von Plettenberg]

I have recently ordered a max v2, and while waiting for shipment have been doing some research into the controller. As I understand it , the max v2 uses a rambo controller board with three of the outputs driving the 1, 2 and 3 delta steppers, another driving the extruder stepper, and 2 mosfets driving the bed heater and extruder heater. The only inputs from the machine are the limit switches, and the bed and extruder temperature sensors. My question is , how is the accuracy of the 3 delta stepper motor positions maintained, especially on a long print? If one of the steppers slip, is this error not then maintained for the rest of the print, and, if the stepper continually slips, does the error not build up over time ? The way I understand it, the only position feedback the controller gets (besides the limit switches) is when the printer z height is zeroed. Perhaps the system is self governing with the other two steppers ganging up on the erring stepper, and forcing it to slip back to satisfy the mathematical equations governing print head position. There obviously is a system that works, having seen the print quality, I am just puzzled by how well the system works without position feedback.
.......any ideas ?

[Captain Starfish]

You had it right first off. If one of the steppers slips, it stays slipped and the print is off from then. Multiple slips and she will keep walking off the initial zero. There's no self governing or correction.

If you've come from a background of stepper driven robotics or CNC where the motors are under some load then this state of affairs is pretty rubbish and any result at all is surprising.

but

it turns out there's very, very little load on the column steppers due to a combination of the angles and leverage with the fairly lightweight effector platform. So as long as the belts aren't as loose as, well, insert an appropriately inappropriate simile here, then there is no stepper skip and we're all happy (and accurately tracking) campers.

[Jimustanguitar]

There aren't any home caliber 3D printers that use true feedback like this. At least not that I'm aware of.

It takes a lot more force to skip a stepper than you think it does. Since there isn't a cutting tool, the forces on the print head are pretty minimal and this works out well.

[Greg von Plettenberg]

Thanks for your insight. Given the millions of individual stepping instructions necessary to build a complex part, I find it pleasantly astonishing that 14 + hour prints dont walk off the build platform, out the door and down the street. I will just file this whole topic under ' dont scratch where it does not itch' .

[mhackney]

Having built and using 2 CNC milling machines on a daily basis for my fly fishing reel kits I can say that stepper motor control without feedback works very well. My larger mill can hog aluminum and steel all day long without skipping a step. These things are very reliable, especially for small production service.

[Captain Starfish]

Interesting: I gave up on my hm10 Cnc project when it skipped all over the place. Bigger motors next time maybe.

[mhackney]

Maybe! I use 470 ounce steppers on my larger mill with ball screws (5:1). The only time I've had skipped steps is when I crash the spindle doing something boneheaded. Fortunately, that is mush less frequently than it was 5 years ago! I was at a sign shop a few months ago and watched a big 5' x 10' CNC router with a high speed spindle that was completely stepper driven. This thing was scary cool!

[Captain Starfish]

I'm kinda thinking that if I did something stupid and ploughed the thing through the bed I'd prefer the steppers to skip than for anything to really break under the strain.

Mine was super budget. Steppers scabbed from dead printers at the local shop, L298 stepper drivers, an Atmel micro, my own code for bashing out the step counting based on the G-code coming in through a serial port, can't remember what slicer type app I was using - this was ten years ago.

Managed to engrave my name on a few bits of plastic but anything more than that and she'd just bog and die. Still have the mill, back to handwheels now.

I'd kinda parked it as a "One day I'll have another crack with position sensors and synchronous geared motors or something".

[Polygonhell]

There is a misconception with Steppers vs Servos with position feedback.

Ignoring overheating electronics for now, the ONLY time stepper will miss a step is if something is providing more resistance than the available stepping torque, even if you had positional feedback there would be no way for you to overcome that resistance to correctly position the head. You MIGHT be able to recover more gracefully.
Servos on the other had require positional feedback, and increase power to the motor based on their inability to reach a position, but they have the same fundamental limit in terms of the power they can use to reach a position.

Tormach will happily sell you a $10K CNC Mill that uses Nema35 steppers with no position feedback.