Caveat

So in “The right tool for the job” I suggested you should avoid buying more vision system than you need. By that I meant that if a sensor will give you the required present/absent response, then resist the temptation to buy the high res, PC-based solution with all the “Pat” tools (sorry Cognex, not picking on you.) But there is a caveat: never buy less than you need, because if you do, at some point the system will fail to perform as you require.

So if you’re truly gifted you’ll buy a vision system that does just what you want, now and for the next ten years. Unfortunately most of us do not have crystal balls, (if we did we’d be in the circus,) and so are unable to anticipate what might be asked of our system. And you can bet your inheritance from Aunt Gertie that one day some V.P. is going to walk through your plant and ask why that vision system can’t read the date code on the box passing in front of it. And when he does, do you want to be the one to tell him that you’ll have to throw that system away and buy a different one?

I didn’t think so.

So when you ‘size’ your next vision system, anticipate the need to add some functionality in the future. Personally, I’m reluctant to use the basic, ‘low end’ sensors when for just a few hundred dollars more I can get something that gives me some ‘headroom’ for future growth. No, I don’t use CameraLink cameras with a quad core processor when a vision sensor will do, but neither to I buy the most basic presence-absence detector I can find, even though it might be good enough.

Frankly, no one ever got fired for buying Cognex.