Sharp Dust Sensor querie for the timing code

Hi All
I have a question regarding the code for the reading data from the sharp dust sensor. Previously Ive used code similar to what others have used but Im currently revisiting this sensor and from what Ive been reading nobody has taken in to account the time it takes to do an anlogue read when calculating their delays, its almost as though they/I was assuming the anlogue read is instananeous but it is significant compared to 280 microsecond delay.
What most people are doing is :

Set the LED
delay 280 microseconds
analog read the output (~111microseconds)
delay 40 microseconds
Turn off LED.

But it has occured to me that the analog read takes about 111mircoseconds on average. Which would make the actual analogue read at about 280+111 microseconds. Which means the 391microsecond pulse width of the LED on is actually going to be 280+111+40 which is greater than the pulse width in the data sheet. Will this be a problem.

My second thought was that analog read is going to take place further along the output curve as long as this is consistant this may not matter and that the LED is still on during this measurement longer than indicated in the data sheet.

Set the LED
delay 280 minus 111 (time to carry out an analogue read) in microseconds
analog read the output
delay 40 microseconds
Turn off LED.

Is it better to subtract the analog read time from the 280microseconds so that the first delay is 280-111(time for an anlogue read), then do the anlogue read and delay 40microseconds and switch off the LED.

I dont know if Im thinking in circles and making an issue where there isnt one. Could this be a reason for people and I previously experiencing noisey data when ignoring the analogue conversion time.

Does anybody have any thoughts on this?
Andy

You are right... have a look at this post ConnectraNet – Random tinkering with Things