How to calibrate a load cell?

Question

I'm working with a load cell and HX711. There was no weight on the sensor but I kept getting strange values, I thought it was broken or something so I used my second load cell but the same thing happened again. I'm wondering what is wrong with it ? If anyone have worked with it or have an idea how to work with it, please don't hesitate to give your opinions.

 #include "HX711.h"
#include <Wire.h> 

#define DOUT  A0
#define CLK   A1

#define DEC_POINT  2
#define STABLE  1

float offset=0;
float calibration_factor = 1;
//float real_weight = 1.560;//kg
float real_weight = 0.5;//kg

HX711 scale(DOUT, CLK);

unsigned char state=0;
long  FindZeroFactor();
float get_units_kg();
void  ReadWeight();
void  FindCalibrationFactor();

void setup()
{
  Serial.begin(115200);
  Serial.println();
  Serial.println("Auto Calibrate Program");
  Serial.println("Send 'a' to Find Zero Factor (Please Remove all weight from scale)");
  Serial.println("Send 'b' to Find Calibration Factor (Please insert know the weight on the scales)");
  Serial.println("Send 'c' Show weight on the scales");
}
void loop() 
{
  if(Serial.available())
  {
    char temp = Serial.read();
    if(temp=='a')
      state=1;
    if(temp=='b')
      state=2;   
    if(temp=='c')
      state=3;
  }

  switch(state)
  {
    case 0:
    break;
    case 1:
      FindZeroFactor();
   //   ReadWeight();
      state=0;
    break;
    case 2:
     FindCalibrationFactor();
     state=0;
    break;
    case 3:
      ReadWeight();
      delay(200);
    break;
    case 4:

    break;

  }
}

long FindZeroFactor()
{
     Serial.println("Find Zero Factor");
     Serial.println("Please wait .....");
     scale.set_scale();
     scale.tare(); 
     long zero_factor = scale.read_average(20); 
     Serial.print("Zero factor: ");
     Serial.println(zero_factor);
     return(zero_factor);
}

void FindCalibrationFactor()
{
  unsigned char flag_stable=0;
  unsigned int decpoint=1;
  for(unsigned char i=0;i<DEC_POINT+1;i++ )
    decpoint = decpoint*10;
  while(1)
  {
      scale.set_scale(calibration_factor); //Adjust to this calibration factor
      Serial.print("Reading: ");
      float read_weight = get_units_kg();
      String data = String(read_weight, DEC_POINT);
      Serial.print(data);
      Serial.print(" kg"); 
      Serial.print(" calibration_factor: ");
      Serial.print(calibration_factor);
      Serial.println();
      long r_weight      = (real_weight*decpoint);
      long int_read_weight = read_weight*decpoint;
      Serial.print(r_weight);
      Serial.print(" , ");
      Serial.println(int_read_weight);
      long x;
      if(r_weight == int_read_weight)
      {
        flag_stable++;
        if(flag_stable>=STABLE)
        {
          Serial.print("Calibration Factor is = ");
          Serial.println(calibration_factor);
          break;
         }
        }
       if(r_weight > int_read_weight)
          {
            x = r_weight - int_read_weight;
            if(x > 100)
              calibration_factor -= 1000; 
            else if(x > 100)
              calibration_factor -= 10;
            else
              calibration_factor -= 1;
            flag_stable=0;
          }
          if(r_weight < int_read_weight)
          {
            x =  int_read_weight-r_weight;
            if(x > 100)
              calibration_factor += 1000; 
            else if(x > 10)
              calibration_factor += 10;
            else
              calibration_factor += 1; 
            flag_stable=0; 
           }  
  }
}

float get_units_kg()
{
  return(scale.get_units()*0.453592);
}
void ReadWeight()
{
  scale.set_scale(calibration_factor); //Adjust to this calibration factor
  Serial.print("Reading: ");
  String data = String(get_units_kg()+offset, DEC_POINT);
  Serial.print(data);
  Serial.println(" kg");
}

Answer 1

The first thing that about Load Cells that always comes to my mind is the circuit which amplifies the load cell's signal, Usually the load cell signal changes in a low mv range, so you have to put a high gain circuit in the way of the signal, which amplifies signal and NOISE. You have to use an instrumentation amplifier for this purpose, otherwise, your signal would become so noisy.

And try to implement a low pass digital filter too, it should compensate the fluctuations and would give you smoother signal.

Answer 2

scale.set_scale(calibration_factor); 

needs to be changed to :

scale.set_scale(calibration_factor/[the reference weight you have used]);