Collecting from arduino pin for specific time interval - loops

Question

I want to run my Arduino for a specific amount of time (say 60 seconds) and collect
data from the analog pin with a specific sampling rate (say four samples a second).
I got the code working in matlab... but I want to use the arduino environment.
Please help me convert it.
a_pin = 0;
fs = 4; % sampling frequency (samplings per second)
mt = 20; % time for measurements
ind = 1;
nind = 1;
last_beep = 0;
tic;
while toc < mt
time(ind) = toc;
v(ind) = a.analogRead(a_pin);
% wait for appropriate time for next measurement
while( nind == ind )
nind = floor(toc*fs) + 1;
end
ind = nind;
end
Ok this is what i have so far in my sketch. Would this measure for 10 seconds taking
readings every 5?
int sensePin = 0;
unsigned long starttime = 0;
unsigned long endtime = 0;
int i = 0;
int n;
const int sizeofv = 50;
int v[sizeofv];
void setup(){
pinMode(sensePin, INPUT);
Serial.begin(9600);
}
void loop() {
starttime = millis();
endtime = starttime;
while ((endtime - starttime) <= 10000) // do this loop for up to 1000mS
{
i = i + 1;
v[i] = analogRead(sensePin);
endtime = millis();
delay(5000);
}
for(n=0; n < sizeofv; n++)
{
Serial.print(v[n]);
Serial.print('\n');
}
while(1) { }
}

Answer 1

I recommend using SimpleTimer library, it does all the work for you and is road proven.

Answer 2

There's one bug in your code, you're incrementing i prior to assigning the first value. So when you'll loop over the v[] array, you'll end up printing v[0] which will have an undefined value, and you'll never print the last recorded value of v (if you did not end up overflow the buffer and write outside its boundaries).
There's one glitch with your code, it's that you make a condition that you need to end it at 10000, and you make two delays of 5000. Theoretically it looks ok, but in practice, you may get a surprise having only one value. That's because code takes time to execute, and an analogRead() is particularly slow.
I think using loops within the loop() function make things more complicated, and add close to no flexibility. So here you have another take you could try, to use the loop() only by tracking states of the loop, based on some simple calculations:
#define TIME_WINDOW 10000
#define NB_STEPS 2
void setup() {
// ...
endtime = millis();
}
bool over = false;
void loop() {
if (!over) {
curtime = endtime - millis();
// get in this block for up to 1000mS
if (curtime <= TIME_WINDOW) {
v[i] = analogRead(sensePin);
// increments only once when 5000 has been reached.
if ((curtime / (TIME_WINDOW/NB_STEPS)) != i)
i = (curtime / (TIME_WINDOW/NB_STEPS));
} else {
for(n=0; n < NB_STEPS; n++)
Serial.println(v[n]);
over = true;
}
}
}
I did not test my code, so it may fail or have typos, but the idea is the following:
we define the time window and the number of measure steps
while the time window is not over:
store the measurement in the current array at current index
if integer division of spent time over time_window is different from current index
store the new value into the current index
e.g.: i = 4000/(10000/2) <=> i = 4000/5000 <=> i = 0 and i = 6000/(10000/2) <=> i = 1
when the time window is over:
print out each result
set over to true thus we're not entering the loop again
But a simpler (non flexible) solution for you would be:
void setup() {
int value1, value2, value3, value4;
value1 = analogRead(sensePin);
delay(2500);
value2 = analogRead(sensePin);
delay(2500);
value3 = analogRead(sensePin);
delay(2500);
value4 = analogRead(sensePin);
Serial.println("first value: %d", value1);
Serial.println("second value: %d", value2);
Serial.println("third value: %d", value3);
Serial.println("fourth value: %d", value4);
}
and there you got your solution!
HTH