Floats seem to be rounded to the nearest integer

Using Arduino Programming Questions
#1

I am having some issues when reporting data to the serial monitor. Below, I have included the relevant parts of the code and sample output. I first initialize variables and then run a series of functions within my void loop. When it comes to printing the data, it seems to be rounding or doing something that I do not quite understand.

For example, looking at the output, acceleration in gs seems to be rounded to the nearest integer even though I set it as a float. Accel (g) is calculated by taking the Raw Accel obtained from the appropriate registers and dividing by 8192, following the sensor specification document. This should result in a decimal answer and not zero or one. What am I doing wrong?

Code (Note that this is not the code in full):

long accelX, accelY, accelZ;
float gForceX, gForceY, gForceZ;
float AccelForceX, AccelForceY, AccelForceZ;

void loop() {
  recordAccelRegisters();
  processAccelData();
  printData();
  delay(3000);
}

void processAccelData(){
  gForceX = accelX/8192;
  gForceY = accelY/8192; 
  gForceZ = accelZ/8192;
  AccelForceX = gForceX*9.80665;
  AccelForceY = gForceY*9.80665; 
  AccelForceZ = gForceZ*9.80665;
}


void printData() {
  Serial.print("Raw Accel");
  Serial.print(" X=");
  Serial.print(accelX);
  Serial.print(" Y=");
  Serial.print(accelY);
  Serial.print(" Z=");
  Serial.println(accelZ);
  Serial.print("Accel (g)");
  Serial.print(" X=");
  Serial.print(gForceX,4);
  Serial.print(" Y=");
  Serial.print(gForceY,4);
  Serial.print(" Z=");
  Serial.println(gForceZ,4);
  Serial.print("Accel (m/s)");
  Serial.print(" X=");
  Serial.print(AccelForceX,4);
  Serial.print(" Y=");
  Serial.print(AccelForceY,4);
  Serial.print(" Z=");
  Serial.println(AccelForceZ,4);

}

Output:

Raw Accel X=110 Y=368 Z=8392
Accel (g) X=0.0000 Y=0.0000 Z=1.0000
Accel (m/s) X=0.0000 Y=0.0000 Z=9.8066
#2 
gForceX = accelX/8192;

Since accelX is a long (which is an integer type), this expression will give you an integer, rounded toward zero. Cast accelX to float before dividing it by 8192 (or divide by 8192.0 instead).

#3

Thank you for your help.
Both your suggestions worked. Would you mind explaining why an integer divided by a float outputs an integer rounded toward zero? And why the second method of dividing by 8192.0 works?

#4

8192 is an integer. Integer/integer results in an integer.

8192.0 is a float. Integer/float produces an output which is a float.

Steve

#5