Somewhat OT - Anyone know optical sensors, etc?

Joe, How about a video camera with a crosshairs or bullseye. I'm not a guru at all on this, but I've done mechanical stuff for video inspection of different parts.

Joe, you have the right idea a cheap photo sensor that will trigger from the dark paint is perfect.

This is a simple application , look in the AB catalog.

We use em on presses all the time to detect material problems , marked from the supplier.

Best

Best Daveb

I don't need it to trigger the drilling per say. The customer would like it to turn on a light so that they know when to clamp and press the start buttons on my drills, etc. The light controls will be built right into their ePanel.

It could be automated, but then the danger of a mistake is higher on these high-cost items if there was a dark knot, etc.

I'll check out the AB Catalog.

Regards, Joe Agro, Jr. (800) 871-5022

If your worried about a dark knot etc. why not use a different marker colour or even a paint marker. "Florescent green" would work. Or a spray can and a small hole template made from cardboard etc.

Are the photo sensors sensitive enough to pick up specific colors only? i.e. florescent green?

"Joe AutoDrill" wrote in news:ro6Ag.4817$Th2.4196@trndny02:

Cognex vision system.

"Joe AutoDrill" wrote in news:877Ag.2281$nU1.213@trndny03:

The cognex can look for a specific pattern, eliminating the possibility of mistaking a knot for a mark.

Joe, These are the best in the biz, commonly used for web (film )registration) they can sense about tpye of color etc.

Take a look:

Joe, I'm from NJ too, Hightstown Exit 8. I've used quite a lot of photoeyes over the years, and this problem is quite tough. Detecting a black or colored mark on a background of a CONSISTENT color is easy. Doing it on a piece of wood is quite difficult. In the greenhouse industry I did a barcode reader using labels made from retro reflective material. This material reflects 100 time more light than a white surface, yet you'd be amazed at how many times a reflection off a glass wall or shiny piece of metal caused us trouble.

I'm not understanding the whole problem. You say the TigerStop is slow. In what way, it looks like it could feed as fast as 35 inches per second? Time to program? How are you going to make the black marks, with a tape measure? Is the manual marking PLUS the time to align the mark to drill going to be quicker? I've done a lot of automation projects for really cheap industries. If you give me more information I can probably help you out.

Gary H. Lucas

"Roger_N" wrote in news:FPbAg.12$ snipped-for-privacy@news.sisna.com:

The cheap banner vision sensors are not good for much. We tested one, for a simple job. The app guys at banner build up a test and tested it, gave us the assurances it would work, naturally, it failed miserably. We had thier applications guys come in and try to get it to work, they couldn't either. All of the other banner products we use (bunches of them) are excellent. The fiber optical beam-thru sensors are infallable and reliable as dirt. The camera's though, are crap. They just haven't been in the business of vision sensing long enough. Go with a Cognex or a DVT. (We have well in excess of 80 cameras in operation in automation in our plant.)

Now that's funny, even to a NJ guy like me. We seem to be the only state where your location is determined by exit numbers.

I'm too far off the turnpike to be marked by an exit because I can get home equally fast from the GS Parkway or 280 West exits.

Well beyond what I would want to tackle.

Actually, my customer says it is too slow. I'm not sure what they mean by that, but I'm simply researching some ideas for them. Received an order for a single unit yesterday so I know they are going to mvoe forward. The more I help them automate with info, the more units I can potentially sell them.

Probably time to program. I think every single piece has a slightly different pattern on it. TigerStop works very well if the pattern is repeatable from what I understand.

Probably crayon or something similar. Truth be told, I don't know.

Imagine telephone poles. Now imagine that each one has to have a 2" hole drilled through it at 5 differnt locations. Now imagine the holes are all on the same plane but in different locations for each pole. The poles are manually loaded onto a bed to be drilled and the locations somehow marked. The drill comes up from below and drills through 5-18" of material each time depending on pole size.

Because his locations differ per pole and because he doesn't have an auto-feeder for the poles, he imagines that he can somehow mark the poles and manually locate them for drilling. His tolerances are obviously not too close (Ever see a woodworker use a micrometer? Nah...)

YES, most of the work I've done with photo sensors is in the automotive manufacturing but you would be amazed what can be had. Most of what we used was for dimensioning parts and doing part rejections. Colour recognition can either be setup in software or like others have mentioned could be done using special lense filters and lights so that only the colour used would even be seen. THINK florescent markers and black light or something like that. You want to make sure that the user can also see the mark though.

We used DVT sensors. Lots of software to work with and they provide ActiveX control of the devices so you can custom program them and monitor them over TCP/IP network. One thing you REALLY REALLY need to factor in here is whether or not you can isolate the vision sensor from the movement of the machine. Vision sensors don't do well when they are shaking and moving because of vibration etc.

Another idea that may / may not work. What about building some sort of "Feeler" a set of arms with small rollers that would drop into a small hole that was pre-drilled. The operator could drill a small 1/2" by 1/2" hole that the feeler arm would drop into. THis would eliminate vision problems, but not natural holes in the wood? Just thinking outside the box.

"HotRod" wrote in news: snipped-for-privacy@corp.supernews.com:

Yup, in addition, lighting is an important factor.

Figure out a way to interface a digital tape measure to it.

Joe, Okay I have a better understanding now. You could mark the poles using the retroreflective tape that I used to use to make my labels in the greenhouse. Viewed against wood these would be absolutely readable, even in direct sunlight, better than any camera or color sensor could possibly be. I know this because my barcode was read at 100 feet per minute, in total darkness to full sun. Unless you have actually tried reading marks in that range of lighting you can't believe how hard it is to do. The photoeye to use is a Banner SM502A. This eye has visible red beam about 1/4" in diameter and can read retroreflective tape from 12" away with no problem. I used to use two side by side. When the tape comes by the first eye you slow down the motion, and creep to the second eye and stop. My watering machines do this with DC drives from a speed of 100 feet per minute and stop within 1/8"

Gary H. Lucas

This may be expensive... maybe not. I saw one of them crazy robotic lumber saws on "Modern Marvels" a while back. Basically it grabbed the log then squeezing it between wheels which drive down it's length to a calculated dimension where it was cut.

Perhaps a similar idea of having the lumber mechanically pushed through a set of rubber wheels. They'd have encoders that tell the programmed system how far it's traveled at which time the device pushing the lumber stops for drilling. If the lumber is not super critical, one could use spring loaded wheels with small teeth on them for better accuracy. Or, the wheels could actually drive the system like a cnc?

-- Bill

Thats the best idea so far......good thinking.

Use a cheap Idec plc with 8-12 I/O and bingo you got it.

Regards Daveb