OT - Machinists/Mechanical Designers

I have a client that needs to measure/verify flatness of a surface to within less than 1/100". Size of surface is about 8 ft.X8 ft. Not my normal area of expertise so thought I would take advantage of all the experience here. Looking for anything ready made for purchase or a "home-built" solution. Might be able to give somebody a "referral" if it works out.

Thanks for looking,

Kirk

dial indicator on base block

What kind of surface?

how thick is the material? Lay on a mill table an use indicator in chuck.

.1" isn't really accurate. You could easily do that with a water level

Excuse me .01" could still do it with a good water level

Some kind of lazer might tell how flat something is and I'm sure there are other ways but I think what might work would be with a long straight edge, like a really good level. You could put the level on 2 precision blocks at each end and then use a dial indicator or internal micrometer to measure to see how much variance there is along the length of the level. You could move the level to different locations to see if readings change. Not real scientific but might be close enough without having to spend a lot of money. Dave

Eight foot by eight foot is one large area to check with a dial indicator.

I would guess that the best way would be a ground straight edge longer that eight foot. If there is such a thing. It would take one honkin grinder to grind a straight edge that long.

The big mold shops would have machines that would get that close if you have something portable.

When I was still at the factory we would check something like that by placing it on a granite block that had been honed to perfect flatness and then use a dial indicator that was setting on that same granite block, but we were checking for far greater precision than what you are asking about.

An 8 foot by 8 foot piece would be nearly impossible to check with the method I described, and if tolerances within 1/100" is what you need, maybe you could find an 8' straight edge, lay it down on the surface, and then use a feeler guage to check for variances.

Spoke to my contact today and got a little more info. He may not have all the requirements or they might be playing things close to the vest for now.

This is a vertical assembly that they want to check as part of a QA/QC process at completion of assembly. Tolerance is now .002 if economical. Limits of their manufacturing process/space prevent them from moving assembly to a fixed point so device needs to move to assembly.

Find yourself a big CMM to stick it on or make your own gantry with an indicator... 0.1 is a mile in the machining world. Heck could probably get away with a tight string grid over the part and a scale... Find the lowest and highest spots... Delta shouldn't be more than 0.100 inches...

What you need for your 0.010 inch requirement is a cast iron base plate. They can be made in smooth top or a T-slot design to hold fixtures and machine parts. The base plate is cast with an underside web in a grid pattern to support the load and maintain flatness. They are made at Bay Cast Tech in Bay City, Michigan. baycasttech.com/companies/bct/plates.htm

If this is steel then the piece could/should be Blanchard ground.

There is also a device called an Auto-collimator. This is a large movable lapping device that laps in surface plates for Metrology inspection labs. (I have never seen this done, only a picture in an old text book)

Most surface plates are accurate and flat to within .0005 inches. The largest surface plate I have ever seen is a 6 foot by 12 foot granite plate.

.002 English or .002 mm? Very big difference. .002 mm is only about 80 millionths of an inch. Autocollimators are measurement devices used to measure flatness on a surface plate. It is called "shooting the plate". Autocolliminators are electronic now, and very accurate. They aren't hard to use, but pretty fussy - and the surface is going to have to be pretty smooth. I believe that the machinist handbook will give instructions, and a description of a grid pattern to use. A plate that large is going to flex, gravity wil affect it. So will heat. And after moving, it will probably change some more. I would shoot it after it has been installed, if it is a critical use item. A couple of items come to mind:

1) the plate should locate on 3 points. More is not better.

2) the plate clamping should be limited to the area within a triangle formed by the locating points.

3) A Michigan outfit, Mitchell surface plate, can check it for you, and they have certs. I guess that $500 or so would cover it? They used to do our plates when I worked for GM. I don't know where you are at, so I don't know if that is practical. There are other folks out there, but they are the guys we used to use.

4) Buy a Machinist's handbook.

Have a nice day!!!

Kirk,

I have a 6 ft straight edge (arched top) ,it
is what you see in the machine rebuilding industry
when someone is hand scraping ways.
It's heavy ,two people to handle it safely ,also
have several shorter ones 2,3,4 ft and several
parallels (matched and 3 or 4 ft long) ,real
gravity seekers .

george

Find a Measurement or Calibration firm with a "Romer" or "Faro" arm (portable CMM). Both companies make arms that will measure well within your .002" over 8 feet.