Hi,
I am trying to implement a parabolic function in Arduino.
The specs I want to create are:
starting point: [0,0]
middle point: [60,100]
end point: [120,0]
The function would be: (1/36)(-(X-60)^2)+100
When I add this function as an Arduino code:
((1/36)*((-pow(X-60,2))))+100;
It only gives the outcome: 100, when I print it..
What am I doing wrong?
You are using integer arithmetic. Use floats.
Thanks for the quick reply,
Unfortunately it still doesn't work..
float RTCVar;
float seconds = 0;
int unsigned long millisVar = 0;
void setup() {
Serial.begin(9600);
}
void loop() {
seconds = floor((millis() - millisVar) / 1000);
RTCVar = ((1/36)*((-pow(seconds-60,2))))+100;
Serial.print(seconds);
Serial.print(" - ");
Serial.println(RTCVar);
delay(200);
}
Your operands are still integers.
Use floats. (hint: 1 is an integer, 1.0 is a float)
i.e. 1/36 = 0
1.0/36.0 = .027777777
A little bit of algebra changes that to
RTCvar = (120.*seconds - seconds*seconds)/36.;
Pete
The pow() function will bite you in the ass, too. It expects float arguments and produces a float output. Raising a value to a power of two using pow() is incredibly wasteful. A bit shift will be orders of magnitude faster, if the value is a integer. If not, simply multiplying by the value is faster.
Of course, you are storing an integer value in a float, so you will have problems deciding whether to treat seconds as an int or a float.
bit shifting multiplies by 2, it does not square!
If you're squaring an integer, an integer multiply is faster than pow.
That is true, but it is more than a simple bit-shift.
(and pow() might have a trap for integer powers and do a multiply itself.)
el_supremo:
A little bit of algebra changes that toRTCvar = (120.*seconds - seconds*seconds)/36.;Pete
Finishing the job and eliminating one multiply:
RTCvar = (120.0 - seconds) * seconds / 36.0 ;
Well indeed, this is a more easy way to create that parabola. Thank you all!!