0
\$\begingroup\$

I am trying to simulate car pedals with load cells. For that I wired up 2x20kg load cells and 1x 100kg load cell each wired to a INA122P which join up at a stm32f103c8 blue pill board.

The circuit works but I have two issues. The first problem is that I get a lot of noise and the second is that when I press one pedal I get a bit of input on the other pedals, what was maximum load at one point will then suddenly not be 100% anymore for example.

My first circuit looks like this:

3 load cell amplifier circuit

After looking for some solutions I found out that a low pass filter at the Data-Pins of the load cell might help so I changed the circuit to this:

enter image description here

That did not actually help though since I'm still getting noise. So I'm still left with the question how to fix both interference and noise.

I don't really know a lot about circuits so I'm sorry if I made some obvious mistake or left out important information. I'd appreciate any help!

This is one of the amplifier circuits which connects via wire to the stm32. The 3 upper traces under the IC have been cut.

Circuit

Load cell wires connect here:

enter image description here

\$\endgroup\$
8
  • \$\begingroup\$ Circuit layout is a big contributor to what you describe, posting a picture of your setup might help us understand where the problem would be. I would suggest the the hx711 might have been a better choice for your application - at least there wouldn't be analog signals floating around and you'd get superior resolution as well. \$\endgroup\$
    – Kartman
    Oct 23, 2021 at 11:49
  • \$\begingroup\$ Sorry, I added a picture of one of the actual boards. I actually copied the design from someone who then mid project decided to only sell boards instead of supplying plans. That's why I'm in a bit of an awkward position trying to figure things out myself. \$\endgroup\$
    – Carl
    Oct 23, 2021 at 12:19
  • \$\begingroup\$ Disconnect all but one load cell electrically. Do you still get an interaction between the pedals? \$\endgroup\$
    – Andy aka
    Oct 23, 2021 at 12:19
  • \$\begingroup\$ If by disconnect you mean testing only one load cell + ina122, then yes. I get current on all three Data pins. Otherwise I'm not sure what you mean by disconnect, using diodes in some way? \$\endgroup\$
    – Carl
    Oct 23, 2021 at 12:23
  • \$\begingroup\$ No, I mean disconnect i.e. remove the electrical wiring between the load cells and the circuitry. Leave only one load cell properly connected. I have no idea what this means: I get current on all three Data pins - if you still get interactions then you have a mechanical problem. \$\endgroup\$
    – Andy aka
    Oct 23, 2021 at 12:56

1 Answer 1

0
\$\begingroup\$

The filter idea you have is ok but you could use much larger caps. You can use up to 47uF with your 1k resistor. This would have a time constant about 1/20 second which should be ok for a brake/gas pedal response.

Also, your schematic show no connection from the ground of your microcontroller to the Vref of your inAmps. Is it there an not shown or did you not complete the circuit?

Finally, add a large cap across each V+ to V- into each load cell. This should minimize your cross-talk across sense lines at other load cells.

Good luck.

\$\endgroup\$
6
  • \$\begingroup\$ I connected all Vrefs to GND I just didn't want to cross even more wires in the schematic. Can I use electrolytic capacitors for the 47uF? I only have those. Are you talking about the Vin+ and -? What would you categorize as a large capacitor for that? Thank you for the great answer! \$\endgroup\$
    – Carl
    Oct 23, 2021 at 14:08
  • \$\begingroup\$ Yes, you can use electrolytic but make sure the black strip (neg side) is connected at Vin(-) for your filter. You can add them in parallel with your current 100nF ceramic caps. No need to remove them. Keeping the For s "large" cap across Vin/Vout, a 47uF to 1000uF can all be used. You'll have to experiment. Close to the load cell is better (short wires). Finally, avoid any long wires of your unamplifed signals. If long (more than 6") cannot be avoided, twist wires together to minimize picking up radio frequency noise. Big area open loops of wire are bad. Keep things tight & compact \$\endgroup\$ Oct 23, 2021 at 14:28
  • \$\begingroup\$ I've cut the wires of the load cell to about 2" and twisted them so I hope after modifying the circuit with the capacitors the problems are gone. This has been a great help, thank you very much! \$\endgroup\$
    – Carl
    Oct 23, 2021 at 14:37
  • \$\begingroup\$ Great. I'm interested how it works out for you. Please report back good or bad results. \$\endgroup\$ Oct 23, 2021 at 15:19
  • \$\begingroup\$ I added 3 47uF capacitors for each Vout and from Vin+ to Vin- and the result has been great! I even tried 100uF and 220uF but 47uF felt like the best response/noise ratio. Thank you again for your help everything works perfectly now. \$\endgroup\$
    – Carl
    Oct 25, 2021 at 20:00

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.