Thats what I do....I have one timer interrupt that runs say at 1KHz (1 millisecond). Then I have a series of "timer variables" that are used by the main line code. Depending upon length of delay I will have some that are 2 bytes and some that are 1 byte. The one byte timers can support delays of up to 255 mSec (about 1/4 second). The 2 byte timers can support delays of 65535 mSec which is just over a minute.

The timer interrupt code looks at each timer variable and if it is non-zero it then decrements the variable.

In the mainline code you simply load a count of milliseconds delay into the timer variable and then in whatever else is going on you can poll this timer variable for a zero/nonzero value. As long as the thing is non zero you know the time delay has not expired.

When using the 16-bit type counters it is necessary to disable interrupts during the setting of the 2 byte variable or during the polling of the value for the timer expire. I sometimes use a technique with the 16-bit variety of allocating a bit flag per timer variable in the bit data area. Then upon setting the timer variable I set the corresponding bit flag. When the timer interrupt decrements a timer variable to zero it then clears the bit flag. Thus when the mainline code needs to poll a timer state it simply has to look at the flag bit and there is no need to disable interrupts during the polling process.

Michael Karas