flatness measuring help

MCritchley · Dec 6, 2022

jccaclimber said:
Your described method produces parallelism of the highest 3 points, not flatness of the surface.

True, are you going to have a .0004 dip between those three points? Probably not, this is the practical machinist after all. Being practical makes us money and gets the job done.

jccaclimber · Dec 6, 2022

MCritchley said:
True, are you going to have a .0004 dip between those three points? Probably not, this is the practical machinist after all. Being practical makes us money and gets the job done.

I’d say no, but apparently the customer thinks it’s a problem.
I still think an air based functional test is correct for this application.

CarbideBob · Dec 6, 2022

I do not get the whole set up and slide your indicator stand on a plate thing.
Is your plate this good? (Rule of ten)

ttrager · Dec 8, 2022

Color the surface to be checked, then quickly lap on a surface plate or a piece of glass with very fine sandpaper . . . lightly, then look to see if any high spots have been cleared off? Or, lap one side true/flat using the same technique and inspecting for a consistent scratch pattern under magnification, then using a .00005 indicator on a stand for the opposite side?

I'll be interested to see what your final solution was.

jim rozen · Dec 8, 2022

Late to the party, but to measure a matte surface with an interferometer, one can apply something like floor wax to the surface, buff it up and that will allow the measurment. In the past we had to check flatness on silicon slabs (2X3 inch) to verifiy flatness down to +/- one micron, using a Zygo interferometer. An old hand here pointed this trick out and it worked great. Measurements to one fringe of red light worked out.

brucecu · Dec 9, 2022

Is there a waviness specification? How are the parts mounted for grinding? Are they wrung into the fixture or floated on wax for side 1? Is there spring back from the grinding or inherent stress in the ceramic? Did you try dye penetrant for porosity, scratches or cracks? Any change in raw material supply or forming method? In application how are they sure the flat face is parallel to the surface of the piece being picked up? What is the vacuum requirement?

M. Roberts · Dec 9, 2022

Jim, Good solution, but not sure if it is applicable to these parts; I think the issue is that the surface is so small, that would someone even to be able to see the light bands if in fact that an optical flat would work. Maybe set up a camera, and enlarge the image? IDK. I'm pretty sure that an interferometer isn't cheap.
Mark

M. Roberts · Dec 9, 2022

Bruce, good questions! No, there is not a waviness called out on the print. I do not know how the vendor is finishing the back, flat side of the part. You bring up a good point regarding the possibility of porosity; I do know that there is a density check performed on the parent block in which "tubes" are hole popped and wire EDM'd out to provide an oversized blank for grinding. I brought up the same issue as you mentioned; there is not a callout on the print regarding parallelism between the opposing faces, and while an individual part may pass a vacuum leak decay (the part must hold 24in/hg for 2min), in an assembly one has to be concerned about the parts being co-planar to each other....we will solve it if it takes all the money they got!!

brucecu · Dec 9, 2022

Years ago I was involved on a project with similar requirements. An ultrasonic time of flight instrument was used to measure flatness and coincidentally could measure interfaces also. Resolution was on the order of one micron -- right in the ballpark for your application. Sonoscan or similar was the instrument.

Mechanola · Dec 9, 2022

The rings are typical lapping parts. No worries, if you have a lapping machine or know someone who is into lapping. 200 can be worked on together. Parallel and to measure within 50 millionths of an inch

rhb · Dec 10, 2022

If you have an optical flat and monochromatic light try polishing the surface for inspection.

In the telescope making community, they polish the surface with a felt buff to increase the reflectivity for testing against an optical flat during grinding.

I am currently trying to figure out how to use an optical flat to test a surface plate, but don't yet have a good monochromatic light assembled.

The potential inspection cost savings make an optical solution very attractive if it can be done as one would expect. Don't rely on doing it, but it's really worth it if you can.

I've got some 1200 grit SiC. I'll charge a small felt pad and try polishing a corner of a "black granite" (i.e. gabbro :-) surface plate and post a photo of the plate reflection. I've also got rouge, but 1200 is pretty fine. Probably do both just for fun.

rhb · Dec 10, 2022

After I posted I went back and looked at some stuff I didn't read entirely. Talk about a hard metrology problem! Though I still think a flat is the right solution because of time to inspect.

As for a process change solution, an iron surface plate charged with fine diamond grit would probably correct any errors as quickly as an inspection would find them. Quick series of figure eights on both faces before cleaning and packing.

I'd certainly inspect the failed samples very closely.

There is significant potential that the seal failure is caused by the installation of the part. A hard seal must be flat *and* parallel. Not easy. And insanely difficult if helium is involved.

rhb · Dec 10, 2022

FWIW

2" x 2" Optical Flat - Master Grade /Flatness Gage 1/8 Wave Band, Flawless NEW! | eBay

Made from UV grade fused silica, this optical window has the hardness, & low expansion coefficient for your project needing a precision reference flat or window. Side 1 and side 2 arevery parallel to 10 arc sec.

www.ebay.com

All you need additional is a light source, either low pressure sodium or mercury vapor with a green filter. Laser diodes will work, but they need optics to expand the beam.

jccaclimber · Dec 10, 2022

IME a cell phone with the screen set to a single color is, while a bit dim, sufficient for an optical flat. A CFL bulb and green filter works as well because CFL bulbs make white from a few narrow peaks, one of which is around green.
As to the comments above, add coplanar to the flat and parallel requirement. I wonder how flat the glass being lifted is across that distance?

brucecu · Dec 12, 2022

Can you get away with an empirical opacity test? Take a transparent flat surface and an opaque surface and place a .0005" shim under a corner. Fill the space between them with ink,, food coloring or the like. This becomes your reference. Apply the colored solution to the ceramic ring and place a microscope cover slip on it. Compare the co;or intensity to your "master". I suggest a cover slip to minimize mechanical loading of the ring. At its size,it is easily deflected with small loads and will be subject to springback when the load is removed. Sort of a real world Heisenberg uncertainty principle.

M. Roberts · Dec 14, 2022

Bruce, thank you for the reply....I am not even going to begin to understand your described method....it sounds very simplistic, but messy...I don't think that is a path that we want to go down. Of all of the responses and solutions as to how to measure these parts, my question as to if an optical flat would work has not been answered. To me, that seems the simplest, most economical, and quickest means of inspection, provided that it works......

rhb · Dec 14, 2022

Bruce's suggestion is very amenable to a computerized solution and much easier to automate than checking fringes.

Use blue ink and a red background, kill all the red pixels in the image. sort the saturation values of the blue pixels and plot against the values for a perfect part. It's simple enough that I may code it up just to do it. The one caveat is that JPEGs are lossy and the compression might remove the variations you want to measure.

Have a shallow tray with the ink. Place the face to be tested in the tray to pick up ink and place it on a flat set up so you can see the bottom of the part you are testing, take a photo and process it.

Either Bruce's approach or traditional optical testing will require a proper reference surface.

Flats can be had very cheaply.

2" x 2" Optical Flat - Master Grade /Flatness Gage 1/8 Wave Band, Flawless NEW! | eBay

Made from UV grade fused silica, this optical window has the hardness, & low expansion coefficient for your project needing a precision reference flat or window. Side 1 and side 2 arevery parallel to 10 arc sec.

www.ebay.com

I've got 3 of these arriving tomorrow if I'm lucky. Still trying to source a low cost low pressure sodium light.

rhb · Dec 14, 2022

Try this experiment.

Place an optical flat such that there is thin layer (~1/16") of ink on the top. Put a light and mirror so you can see the part from the underside. Dilute the ink until you can clearly see the difference between the layer between the flat and part and the rest of the ink pool.

XanderB · Dec 16, 2022

M. Roberts said:
Hello. I have small ceramic rings that I need to measure the flatness on; they measure 0.420" ID x 0.460" OD x 0.035" thick; that right, an area of only 0.020" wide. The spec calls out a flatness of 0.00039" (0.01mm)...how would be the best/easiest way to do this? I thought about using an optical flat, but not being that versed, I thought that someone out there might have a better idea....Thanks, Mark

Ehh...pay another shop with a CMM to check a sampling and rock on? Maybe go old school and use jacks. Custom jacks, granite table, & a test indicator

eKretz · Dec 16, 2022

Many of these methods would be very time consuming to check a lot of parts. Personally, I would check a representative sample quantity and if there turns out to be a problem, send them back to the vendor for rework or replacement. It's their responsibility, not yours. You might try a few rework methods (a quick lap on emery or perhaps a lapping plate as a few have suggested might be a good one if the problem turns out to be a flatness issue) and pass the methods along to the vendor with the parts.

flatness measuring help

Hot Rolled

Hot Rolled

Diamond

Aluminum

Diamond

Aluminum

Aluminum

Aluminum

Aluminum

Stainless

Aluminum

Aluminum

Aluminum

Hot Rolled

Aluminum

Aluminum

Aluminum

Aluminum

Plastic

Diamond; Mod Squad

Similar threads