Load cell value oscilating, not stable

General Discussion
1

Basic Information:

Printer Model: UNDIFINED
MCU / Printerboard: BTT OCTOPUS V1.1
Host / SBC DEBIAN VIA PC
klippy.log
klippy (2).log (1.7 MB)

Describe your issue:

Hi ,im implementing spool weight scales in klipper and already conected the 10 kg load cell, amplifier(hx711) , followed the instructions from klipper website and did the diagnostic, seems alright
but then when i command it to tare, it gives me various values, without any change in the weight like 6%, 10%, 4%, i dont know why is oscilating soo much, can anyone give me any tips?

2 Likes
2

—Can we see what the config looks like? How many counts per gram did the calibration procedure report?—

Log file has it. Counts/gram is not stable, bu looks like it’s less than 200. But the tare value differed by many times more than that. Essentially the noise floor is too high to see grams.

You might want to try connecting it to the debugging tool and seeing if the graph is wiggling around like crazy. We have had a lot of problems with grounding and ground loops on these sensors. Being that they are analog they tend to be very sensitive to noise.

3

so i was giving the amplifier 5 volts, tried 3.3v and now its running without problems ,thanks for the help. another problem that i run in to is that i have to plug 4 of this load cells in to klipper, in my MMU i have 4 slots, each one with his own scale. How can i add 3 more load cells?

1 Like
4

you can name them:

[load_cell mmu_1]
[load_cell mmu_2]
[load_cell mmu_3]
[load_cell mmu_4]
5

I’ve seen someone else report that issue. The HX711/HX717 do support 5V, however, if the 5V supply isn’t stable enough this could be an issue. It could be that the 5V rail is supplied by a switching regulator and that’s not good enough for this sensitive application. This really requires a linear regulator (LDO). It could be that the 3.3V rail is cleaner because its used to supply the CPU and it just had to be better controlled.