Got it working. The problems were the delayMicroseconds input being too large (see comments) and the frequency mismatches.

I set up a script to time the code segments of interest and that allowed me to select the best wait duration. It looks like this:

while (i < 256) {
    /* your code to time here */
    i++;
}
i = 0;
unsigned long time = 10000000 + micros();
Serial.println(time);

Make sure to time the empty loop and subtract that from your readings w/ code. Take the difference between time read outs, subtract the printing statement overhead (which is dependent on how many digits it is printing, hence the 10000000 + micros() to keep # of digits constant) and divide by 256.

I measured 4 different combinations to find the closest frequency match:

The final code ended up looking a lot like the one from this link; counting off cycles of the on-off 38 kHz pulse.

Got it working. The problems were the delayMicroseconds input being too large (see comments) and the frequency mismatches.

I set up a script to time the code segments of interest and that allowed me to select the best wait duration. It looks like this:

while (i < 256) {
    /* your code to time here */
    i++;
}
i = 0;
unsigned long time = 10000000 + micros();
Serial.println(time);

Make sure to time the empty loop and subtract that from your readings w/ code. Take the difference between time read outs, subtract the printing statement overhead (which is dependent on how many digits it is printing, hence the 10000000 + micros() to keep # of digits constant) and divide by 256.

I measured 4 different combinations to find the closest frequency match:

The final code ended up looking a lot like the one from this link; counting off cycles of the on-off 38 kHz pulse.

Got it working. The problems were the delayMicroseconds input being too large (see comments) and the frequency mismatches.

I set up a script to time the code segments of interest and that allowed me to select the best wait duration. It looks like this:

while (i < 256) {
    /* your code to time here */
    i++;
}
i = 0;
unsigned long time = 10000000 + micros();
Serial.println(time);

Make sure to time the empty loop and subtract that from your readings w/ code. Take the difference between time read outs, subtract the printing statement overhead (which is dependent on how many digits it is printing, hence the 10000000 + micros() to keep # of digits constant) and divide by 256.

I measured 4 different combinations to find the closest frequency match:

The final code ended up looking a lot like the one from this link; counting off cycles of the on-off 38 kHz pulse.

Got it working. The problems were the delayMicroseconds input being too large (see comments) and the frequency mismatches.

I set up a script to time the code segments of interest and that allowed me to select the best wait duration. It looks like this:

while (i < 256) {
    /* your code to time here */
    i++;
}
i = 0;
unsigned long time = 10000000 + micros();
Serial.println(time);

Make sure to time the empty loop and subtract that from your readings w/ code. Take the difference between time read outs, subtract the printing statement overhead (which is dependent on how many digits it is printing, hence the 10000000 + micros() to keep # of digits constant) and divide by 256.

I measured 4 different combinations to find the closest frequency match:

The final code ended up looking a lot like the one from this link; counting off cycles of the on-off 38 kHz pulse.