A customer called last week with a problem on their packaging line. one of the machines was running out of sync with the feed screw, and making a hell of a mess of the labels.

I took a look and found that the (poorly documented, of course) machine had an Allen-Bradley controller that was apparently misbehaving. After an hour of warmup where it would work perfectly, even though the "Product Present" prox sensor was inactive (and showed as such on the controller's LED strip), it was still making "stick-a-label-on" output. This output went through a time-delay relay to produce a pulse that triggered an air solenoid that.... well the details aren't really important.

Having ZERO information on the AB box or its programming, and having no access to the programming cable or software, I took a different route. It was pretty obvious what the thing had to do to work properly - it involved looking at three inputs, going through a state table and generating a timed pulse at the correct time.

I took a couple of quad OPTO-22 modules (one input,one output) and literally electrical taped them together with an arduino board wired to the I's and O's with pigtails hanging off to connect to the existing controller. All told, didn't take much more than an hour to get the stuff together.

I wired it in place, and the programming took about another hour to get the timings correct. And VOILA! the machine is fixed.

So now I'm thinking that maybe there's a product to be made - a board that holds a couple of Quad OPT-22 modules, a microcontroller, terminal blocks and enough power supply to run from 24VDC or 120/240VAC. Programmable in the language of your choice (mine is C, but whatever can run on an AVR).

As if I really need another project... I could bang this one out pretty quick. And I KNOW I could use a few now and then - it would have been good to have one sitting on the shelf ready to go last week.

Any thoughts?