How accurate can a laser measure? I have a metrology salesman telling me .015" is MAX

How accurate can a laser measure? I have a metrology salesman telling me .015" is MAX

Does anyone know of any laser scanning technology that can accuratley and repeatedly scan surfaces to .0001"? Assuming one can control reflectivity and all that?

Also, does anyone have any experience implementing a CMM with Pro|VERIFY from PTC?

This is the route I wish to take my operation for automatic scanning and verification of parts.

I personally have not been involved in much laser or optical applications. However, here are a couple of links to some FARO products that you may want to take a look at. What is the size of the parts you are trying to scan. FARO has a few different products that may suit what you need. Nothing I am afraid is down to the .0001" tolerance. The FARO Laser ScanArm is hitting +/- .0021" within a 6' volume.

http://www.faro.com/adobe_pdf_mgr.aspx?item=978&pdfTitle=04REF201_057_Laser_ScanArm_V3_pdf_2

There is also the Photon Laser Scanner that is +/- 2mm at 25'.

http://www.faro.com/adobe_pdf_mgr.aspx?item=969&pdfTitle=FARO_Photon_en_pdf

Have you looked at any optical systems such as ones provided by OGP (http://www.ogpnet.com/ogpMain.jsp). I am not sure what the size limitations are but I know of some people that have had a lot of success with their products.

Hope this helps! Good luck and happy measuring.

It's easy to confuse laser scanning with laser measuring.
Two very different things.

I've done this on parts the size of the head of a pin to the size of a entire car body

The resolution of any scanning system is limited by the sensor resolution and the size of the area being scanned. If you limit the size you can get down to a little under one micron but your scanning field is going to be very small. (something like .100 ) If you want to scan a car fender then you are going to see something like .015 at best. The area being scanned is divided by the sensor resolution.

Actually you can use plain white light projected as a stripe and achieve results just as good as a laser. People hear the word laser and expect ultra high accuracies. It just isn't so. If you want to get down below one micron you are into UV light technologies because you need the ultra short wavelengths to avoid diffraction problems.

Machine vision systems such as those by OGP, View Engineering, Ram Optical, and others can do this at about a tenth but they can't scan a large part all at once. Instead they move the sensor (a camera) around to take many pictures so they don't have the speed of a laser scanner. These machines are basically video cmms but since they gather a large number of points at once they are much faster than a touch probe.

Two big questions.
How big is the part to be scanned? What is the size of the smallest feature to be measured?

At these kind of resolutions dust or dirt can be a very big problem. A piece of lint from a rag will throw your measurements off by a mile.

Bob

This is a very complex question which defies a simple answer. Changes in the measuring environment (temperature, etc.) can easily degrade system accuracy by one or two orders of magnitude.

For reference, the system resolution for the Agilent (nee Hewlett-Packard) 5527A/B laser system is given as 10 nanometers for the standard version down to 2.5 nanometers for the high-precision product.

You might want to read their product note 5952-7973 entitled "Achieving Maximum Accuracy and Repeatability". It discusses all the parameters and variables in detail.

You can download it from the Agilent website www.agilent.com or I can email you a copy.

When considering any such equipment, you should get a copy of the manufacturer's datasheet for the product. Statements from sales people are worth less than what you pay for them.

- Leigh

Good info !!

All very important things to control. Temp control and recording the temperature at which a reading was taken would both be requirements.

Right now, I have +-.001 across a 5"x1.25"1.125" part with a sloping profile.. but I wouldn't invest in any system that can get to at least .001 across a 40"x20"x20" volume.

I would have thought that laser, sonar, lidar.. what-ever would have some sort of redundancy check built in to scan the same point from multiple angles and use any internal heuristics to figure out the point to a known accuracy and repeatabilty. As for the focal length I guess that's why the head has to be moveable like a traditional CMM setup. I'd think this would be breakthru technology that somebody would be working on.

These are all great concepts. Hopefully, someday, the tech will catch up.

Does anyone use Pro/VERIFY to integrate with thier exhisting metrology equipment?

Take a look at this web site for a very nice laser-based measurement system. It will do more than you need and on large parts. You can get the system with most of the associated accessories for only $200K. I got offered a slightly used system for a deal price of $149K. I think Rolm or whatever also part of Hexagon is a little cheaper. Just what you need if you are building jet fighters in your garage.

http://www.leica-geosystems.com/metrology/en/lgs_406.htm

I work at Agilent, and my buddy that works on the laser measuring systems showed me how accurate they are. He set up the laser on one end of a workbench that can support 1000lbs, and set up the receiver on the other end. When he sets a ballpoint pen on the middle of the bench, you can see the reading of the laser change by three significant digits! Blows me away.

As a number of replies indicate, there is a significant amount of "stuff" out there, but the usefulness of any of it is completely dependent on what exactly you are trying to accomplish. It isn't clear if you want any certain level of speed of measurement, what may be the maximum envelope, if this is for a production enviroment or total blue-sky R&D, or any other factors which would greatly influence the use (and usability) of the hardware. (Especially what budget there may be for this equipment...)

There are laser triangulation and tracker technologies which can get within a few microns on surface contour scans, but that can take a while to complete, and the resolution may be less for larger surfaces, and increased distance. There are distance-measuring interferometer setups, such as HP/Agilent hardware, that can measure to sub-nanometer resolution under the right conditions, but can't measure areas, only a distance to a point. You might be interested in some of the auto industry stuff that acquires/analyzes surface scan "point cloud" data for 3D contour checking. Another thing to consider is whether you will utilize any of your existing equipment as part of the metrology setup, and if so, what errors you bring to the table with that equipment. Have you looked into machine-tool based probing?

As CarbideBob notes, optical "fringe" projection (coupled with a sufficiently accurate CCD camera and optical system, with the appropriate software) can do a great job analyzing surface contours on areas.