Under bed load cell based Z-probe.

I’m experimenting with a load cell based Z-probe. It’s somewhat a mixture of JohnSL's and palmer23's.

What’s needed:

- STM32F103C8T6 board. I’m using a “black pill” [wiki.stm32duino.com] (Any Arduino board could be used, but those * pills board are really cheap.)
- STLink v2 or USB to TTL (3.3v) adapter to program the board.
- Arduino IDE + STM32duino
- HX711-multi library [github.com]
- 3 load cells. I’m using 3 x 20 Kgs because I already have it, but 3 or 5 Kgs one should give a better S/N ratio.
- 3 HX711 breakout board, set sampling rate to 80 samples per second (see the link in palmerr23’s thread above)
- Printed parts to mount the load cells & support the bed
- Connectors, wires, screws, springs, etc.



The output from HX711 is so noisy that I have to throw away the 10 LSBs. After that, even without calibration, the load cells + HX711s give reliable outputs. The outputs are quite linear, we can add them together (digitally) so that the hot end can be probed anywhere on the bed. (The first column is the summation, the 2nd to the 4th are each individual output)

YouTube

With proper calibration and a little more calculation, we can use it to calculate the weight and location of an object on the bed. (Planned)

YouTube

I'll put more details here when it's in a usable state.

A little progress. Test run.

Delta auto calibration with Duet Ethernet/dc42 fork of RepRapFirmware. There are some fault triggers. Calibrated 6 factors using 17 points, deviation 0.072. Not sure if it is my build or the Z-probe that is inaccurate. Just ordered 3x5 Kgs load cell, hope to improve the S/N ratio.

YouTube

Good work. There are a good number of methodologies for sensing nozzle contact.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

If you are not already doing it, you may find that adding a Z probe recovery time (R parameter in the M558 command) reduces the false triggering. We use 0.4sec for the strain gauge effector.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
DjDemonD
Good work. There are a good number of methodologies for sensing nozzle contact.

Thank you!

Quote
dc42
If you are not already doing it, you may find that adding a Z probe recovery time (R parameter in the M558 command) reduces the false triggering. We use 0.4sec for the strain gauge effector.

I haven't tried it yet. Thank you for your suggestion!

I finally found the culprit that caused the false triggers!



A missing capacitor (C6) in one of my HX711 board cause random spikes in the readouts. Replaced the board, problem solved.

The 20 Kgs load cells were replaced with 5 Kgs ones too. It is more sensitive but sadly it didn't improve the deviation. Anyway, I think it's usable and I'll move on to integrate Diamond Hotend to my delta.

I have put the firmware and instruction into GitHub

This is a test print 300+ mm diameter using PETG, 0.4mm nozzle, 0.3mm layer height.Calibrated using 6 factors, without grid compensation.

Hello. Can you add a fourth beamWitam.
Czy możesz dodać czwartą belkę?