• State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 81 subscribers
  • Views 587 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

MSP430I2041: Load Cell Interface Noise Issue - Seeking Assistance

Sadeen Alaa
Prodigy 30 points
Part Number: MSP430I2041


I am experiencing a challenging issue while trying to interface a load cell with the MSP430 microcontroller, specifically by sending the load cell signal to the A0.0+ and A0.0- channels. Here's a detailed breakdown of my hardware and software configurations and the problem I'm encountering:

Hardware Configuration:

  • I've set up an RC low-pass filter on the signal path before it reaches A0.0+ and A0.0-, consisting of a 100nF capacitor and a 1k resistor.
  • I also tried to connect the signal directly to the A0.0+ and A0.0- pins without hardware low pass filter.
  • The load cell is powered by a 3.3V supply from the MSP430.
  • Additionally, I've tried both with and without a 100nF capacitor between pins 9 and 10, as recommended in the datasheet.

Software Configuration:

  • I've experimented with different gain settings (1, 2, 8, and 16) and found that gain 1 yields the best results.
  • Oversampling is set to 256.
  • SD24LSBTOG is configured as 1.
  • I've tested both single and continuous conversion modes.
  • The grouping is disabled.
  • Data format is set to binary.
  • I'm using the internal reference voltage.
  • I've verified that all register values align with the expected configuration through debugging.

Problem: The primary issue I'm facing is that the raw data obtained from the SD24 module is excessively noisy, which directly affects the accuracy of the weight data calculated from the SD24. To troubleshoot, I've confirmed that the load cell itself is not the source of the problem by successfully interfacing it with another microcontroller (STM8s with HX711). However, when I use the MSP430i2041 microcontroller to replace both the HX711 and STM8, I encounter this noisy signal issue.

The noise in the signal is causing the weight measurements to fluctuate significantly, resulting in inaccurate readings. I'm seeking assistance in determining whether this particular MSP430 model has the capability to replace the HX711 effectively or if there might be a software-related issue causing the noise problem.

Any guidance or suggestions on how to address this noise issue and improve the accuracy of the load cell interface with the MSP430i2041 would be greatly appreciated.

  • 0
    Guru 100745 points

    1) Powering the load cell from the MSP430 sounds like you're doing it through a GPIO. The commercial load cells I'm familiar with have fairly low impedances (350 Ohm typical), and as the bridge goes unbalanced it could (try to) draw considerable current through the pin (rule of thumb: max 4mA) which could cause the bridge source voltage to droop. If you can, try feeding Vbridge from a "stiff" lab power supply.

    2) As I recall OSR=256 gives about 4ksps, compared to the HX711's max 80sps. Have you tried decimating (by averaging) your readings down to around 80sps?

    3) What does the noise look like? There may be a clue there. 

  • 0
    Guru 25175 points
    Sadeen Alaa said:
    I've set up an RC low-pass filter on the signal path before it reaches A0.0+ and A0.0-, consisting of a 100nF capacitor and a 1k resistor

    For a differential input you usually need five components to filter both common mode and differential noise. A series resistor on each signal line and three capacitors. One each from signal to ground plus one more between the signal lines.

    Also, the recommendation to keep the sampler happy is a corner less than 10KHz but the step response tells me that the signal needs to be cutoff way below that.

    You don't mention it but I have to assume that you are using shielded twisted pair wire.

**Attention** This is a public forum