0
\$\begingroup\$

I'm making this rig to measure rates of carbonation. It's sensing 6 different variables including mass of sample. The board is Wi-Fi capable (ESP32 WROOM), so I can publish results wirelessly and see what's going on in my browser. Good stuff. HOWEVER, the load cell is killing me.

enter image description here

I've tried a few things, ending at this setup:

enter image description here

I've confirmed the input and output resistances of my load cell are identical (995.3Ω) with an LCR meter. I've used Sparkfun's HX711 PCB because I heard the power supply on it is super noise-free. I've made the load cell rig out of aluminum and steel, so it's super stiff. Yet I still see constant drift. Values going down, down, down. Using basic example of the bogde library (link.)

enter image description here

No idea what to do anymore.

Project is for carbon capture tech.

\$\endgroup\$
15
  • 1
    \$\begingroup\$ Do you see the same drift if you just measure the load cell resistances directly with a good meter? You need to isolate the source of the drift and determine whether it is in the load cell itself or in the electronics. \$\endgroup\$
    – Elliot Alderson
    Commented Jun 30, 2023 at 13:50
  • \$\begingroup\$ You should deliver more information. Manufacturer and model number of the load cell, data sheet, circuit used with load cell? Amplifier and AD converter used for the load cell, voltage source for the load cell? \$\endgroup\$
    – Uwe
    Commented Jun 30, 2023 at 13:51
  • 1
    \$\begingroup\$ Have you quantified the actual drift? Ie converted the raw numbers into the weight value? Realise that most of your 24bit resolution will be noise and you’ll be doing good if you get 14 bits of useful resolution. Have you tried turning the loadcell upside down to see if it sag that is causing the drift. \$\endgroup\$
    – Kartman
    Commented Jun 30, 2023 at 14:23
  • 1
    \$\begingroup\$ "drift" is often how high-frequency oscillations look when you sample them too slowly. You're sampling once per second. I urge you to closely observe the output that you're sampling with an oscilloscope. \$\endgroup\$
    – Scott Seidman
    Commented Jun 30, 2023 at 14:53
  • 1
    \$\begingroup\$ @DrMoishePippik the load cells came without datasheets. I assumed they use Wheatstone bridges. I've ordered a certified load cell as a sanity check. 300g limit. \$\endgroup\$
    – Popeye
    Commented Jun 30, 2023 at 17:51

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.