Two timed inputs?

[BoydNielsen]

I have a project where product is fed to two different conveyor belts via an alternating mechanical diverter. I want to add a photocell on each of the two conveyor belts so that when product backs up, the diverter will only shift to the non-backed up conveyor. If both photocells are covered for a certain length of time, the machine will quit feeding product altogether. I figure an operator-controlled adjustable timing (0.0 to 9.9 seconds, in 0.1 second increments) will be able to allow the various product to pass through, then when the photocell(s) is covered for the set time, the software will control the diverter or stop feeding. Then, continue normal diverter operation or restart feeding when the backed up condition clears.

Here's my question: If I use TIMER0 and TIMER1 to measure the time that the photocells are on, then interrupts at the set time, won't there be a time when the TIMER1 interrupt tries to come in while I am trying to handle the TIMER0 interrupt, and vice versa?

Or because we're talking about 0.1 sec (100 msec) increments, instead of interrupts, should I just poll a timer every program scan cycle and just do the math?

[newguy]

I wouldn't worry about the time difference as it's honestly 'pee' in the ocean. Yes, a timer0 interrupt can occur while timer1 is being serviced, and vice-versa, but the time it takes to finish one and service the other is absolutely tiny compared to the resolution of 0.1 sec.

The other option is to set up a single timer to 'tick' at 100 msec or 50 msec. When a photocell becomes blocked, make note of the time (which could be a 16 or 32 bit integer). In the timer interrupt routine, increment a 'blocked' counter for each photocell as long as each is blocked. If that counter exceeds the operator preset value, take action. When the photocell becomes unblocked, reset its counter. I'd use either the external 0 and external 1 interrupts to detect photocell block/unblocking or the port B high nibble interrupt-on-change (pins B4-B7).