Discussion - sensorless homing

[jonnybischof 60]

Picking up an idea from here:

https://ultimaker.com/en/community/view/11571-step-by-step-installation-of-silentstepstick-drivers-on-umo?page=5#reply-108156

 

...

Especially the stallGuard2 function - which offers new ways for sensorless homing...

If you didn't follow the linked discussion:

The Trinamic TMC2130 stepper motor driver has a feature called StallGuard2 which detects when the motor stalls, and reports that to the controller.

For now, these drivers aren't available as a ready kit for 3D printers (Note, the Silent StepSticks use the TMC2100 which doesn't have the SG2 feature), and there is no firmware support for the feature.

But that could both be arranged, if there is any actual use for it.

------------------------------------------

Let's say our motor drivers were able to detect and report stalled motors.

Could - and should - this feature be used to simplify the homing process on a 3D printer?

Meaning: Instead of using a limit switch as an endstop, just let the gantry bump into a defined obstacle and have the motor driver report the stall.

--> Any negative effects on the...

motors?

drivers?

gantry mechanics?

homing precision?

...?

------------------------------------------

Motors:

Stalling a stepper motor doesn't have a negative effect on the motor itself. At least not if it's only for a short time.

Drivers:

The Trinamic drivers with SG2 functionality don't have a problem with stalled motors. The feature wouldn't make much sense otherwise...

Gantry mechanics:

A hard "crash" should be avoided. Maybe rubber bumpers can be used?

Homing precision:

There is actually no need for super-precise XY homing, because the position of the printed object on the build plate doesn't need to be accurate or repeatable to 0.001mm. 0.5 to 1mm repeatability would be enough.

The Z axis is another matter entirely, as it does in fact need very precise and repeatable homing.

A hard and precise bumper would be necessary. I doubt that the Ultimaker's leadscrew mechanics will give a precise result, even if there is a hard endstop. Any torsion in the coupling (for those who use a flexible shaft coupling - no go) or maybe even in the leadscrew itself would have a negative impact on precision / repeatability.

Another thing: Homing could sound a little ugly. Wouldn't be a real concern, but if people think there's something wrong, it might be a problem...

------------------------------------------

Why bother at all?

Why not use a driver's feature if it's readily available? It's useful for detecting stalled motors, for example due to things like people's hands obstructing the gantry.

The material savings aren't great, but you can actually save some material costs.

More importantly though, you can save some wiring inside the printer, facilitating the machine's design and shortening assembly time. Now that is a reason to bother!