Anyone Familiar w/ OMRON F-10 Pattern Sensor?

I'm trying to set up a machine with an OMRON F-10 Pattern matching sensor and having a bit of a problem interpreting the interpreted, so called, instructions which came with it.

I'm trying to detect the presence / absence of a groove in a part. The sensor has the capability of doing a plain (contrast averaging) measurement or a pattern measurement. I'm pretty sure I want pattern matching.

Also, using some downloadable, free, (you get what you pay for) software you have the ability to futz with the detection area and pattern area and something called a registered area but the whole thing is a little fuzzy and the area's seem to change at will. I really need to get a good grasp on this to mask out some bright spots in the camera's field of view.

Also a note, we have pretty much determined that it will be better to teach the system on the absence of the groove as that image is more stable than the one with the groove (i.e. the groove width varies slightly). When the groove is absent, there is still a chamfer in the view which displays as a bright horizontal line in the field of view. This is the pattern I want to remember.

I know there are probably other ways to do this but this is the system I have now and really can't change it.

I have it close but I still get false readings maybe 1 in 250 parts which is way unacceptable on a 2 shift a day machine w/ a 19 sec cycle time. So far I have only rejected good parts but I'm sure if I don't have a good handle on what this thing is doing it's only a matter of time before I pass a bad part. This is the last inspection machine before these parts get shipped to a big-4 auto mfg. and you know what happens if they find 1 bad part

Any insight into setting this thing up would be greatly appreciated.

Norm,

I have used the F10 unit, it is a vision pattern match sensor. From your description of the problems you are experiencing let's start with the sensor it's self.

The F10 when set up in pattern match will only give you a "good signal" when the pattern matches exactly. Which explains why when you are looking for the groove it does not all ways work, (i.e. the groove width varies slightly).

As for the bright spots, again the F10 is a vision sensor, which means it is effected by ambent light sources. The F10 has it's own light source and if possible only use this source, it is a high frequency light as opposed to 60HZ ambent lighting.

When the ambent lighting is effecting the immage which the sensor see's it could be either in the dark or light half of the cycle, which may explain your area seeming to change at will.

My experience with the F10 has been excellent, I too have used it on automotive application with out any problems. However you must make sure that the part is exactly the same and that you have the ambent light sheilded from the sensor.

If after you try this call your local Omron guy, I am sure he would be more than happy to assist you with this issue.

Mike

Mike:

Thanks for the tips. No where in the manual was the exactness of the pattern match made clear.

I called Omron and everyone was very friendly but it seems all the vision people were flying back from some seminar in CA friday so I'll have to call back tomorrow (Monday).

Our rep said that the ambient wouldn't be a problem but I'll run it by the Omron guy tomorrow.

Thanks Again

Norm,

I also left out that the part has to be positioned the exact same way every time, if not then you will not get repeatable readings.

Mike

I've only worked on Cognix vision systems but,
The number one cause of problems was lighting, number 2, product positioning.
I have heard that these are the two major problem areas for most vision systems. If you can get, or make, these two variables constant, then your problems should be greatly reduced.

Doug