Results 1 to 20 of 20
Thread: Optical Flats
04-16-2008, 01:01 AM #1
I'm trying to get a better grip on the use of Optical Flats in hopes that I can properly calibrate my own Micrometers when needed. I have certified Grade 2 Webber Gage Blocks I use regularly, but they don't tell me if the anvils are parallel. I bellieve I understand using 4 Optical Flats of different thickness with a 0 to 1" micrometer to check for anvil parallelism in four "quadrants" (for lack of a better term), but how do you check larger Micrometers for parallelism with them? I can't imagine how wringing gage blocks to them would indicate parallelism.....
04-16-2008, 01:14 AM #2
That's exactly how you check larger mikes to ensure spindle face parallelism: wring a micrometer check flat on a suitable 1" increment gage block and look for the straightness and parallelism of the fringes in the usual way on both spindle and anvil. Work clean and you will never scratch the apparatus.
Last edited by Forrest Addy; 04-16-2008 at 07:04 AM.
04-16-2008, 01:27 AM #3
Well there we go........The search for flats is on.
Are four flats a requirement, or will one do it?
I kind of figured this one would attract you!
04-16-2008, 01:43 AM #4
What Forrest said and you can read up on a bunch of different techniques using optical flats in "Advanced Machine Work" by Robert H Smith (1925) from Lindsay Books ISBN 0-917914-23-6. Section 15 in the book is titled "Mechanical Measurements with Light Waves". There are probably more recent descriptions of using light and optical flats but this is the one that came off my current bookshelf.
Lindsay Books page on Advanced Machine Work.
Google Boooks has earlier editions for browsing but they don't have the section on using optical flats
Anyhow... as Forrest said: Just stack up some blocks AND the optical flat for the measurement you want to make. Easiest is to use an OF that has parallel flat surfaces (not all of them do). There is no need to get 4 optical flats... just make up the different increments using the blocks. You can also check your blocks as you go. It is handy to have a large-ish (3" dia) steel flat or OF to use as a stage. You can even make up your own steel flat but that can be very time consuming.
04-16-2008, 01:55 AM #5
You will also need a monochromatic light source. I have a nice old Do-All monolight I got off eBay for less than $50. There is on on eBay right now (Item number 330227025547 ) that is different than mine. He wants $35 just for shipping. They tend to be heavy buggers though due to the transformer. Mine is a mercury lamp... puts out some clean yellow light AND lots of UV. Has a filter to remove the UV.
There are some old timer tricks for making a filter for a regular lamp. The book mentions a ground glass filter or parchment paper (or mebbe stretched pig skin ) and a regular electric light or sunlight. Main thing is you need a bright source of light in a single wavelength so that it isn't swamped by the other wavelengths for the other light source(s).
04-16-2008, 02:00 AM #6
Thank you David,
I'll track down a copy for sure as it looks quite interresting. Never been much of a reader, only ever read one book from cover to cover before..........."Holes, Contours, and Surfaces" by Richard Moore.
Greatly appreciate the help!
04-16-2008, 02:16 AM #7
I lucked out and snagged a set of 4 Miti blocks. Pristine. $30. Ebay. Ok, tool gloat over. The blocks work somewhat even under fluorescent light but mono is better.
The flats are the "orthodox" way to check, but if you have a very precise gage balls (especially if you have four that differ by 0.0625) you can do a cursory check. If each ball registers the same value no matter where on the anvil/spindle you measure, this may be good enough of a practical test for your purposes.
04-16-2008, 04:11 AM #8
I assume you clamp an optical flat against the reflective surface of a gage block so that it acts like a mirror that allows you to see fringes on the micrometer tip face that are reflected off the gage block face. Seems you should be able to see fringes between the gage block surface and the optical flat and also see the fringes between the micrometer tip and the other side of the optical flat. That would give some indication of parallelism of the faces of the micrometer. I have never thought of worrying about the parallelism of micrometer faces to the level of interference fringes so I may be confused about this. Optical flats generally have a "good" side that meets the specification which is usually 1/10 wavelength of red light or around 25 microinches (0.6 micron). The surfaces of optical flats are generally not parallel to the same accuracy as the flatness of the best side as the second surface is not as important in the normal use of the optical flat which is checking flatness and not parallelism. A nominal 6" diameter optical flat is around 1" thick so it wouldn't be easy to use on a 1" micrometer. A 4" optical flat is about 3/4" thick usually. I have some 2" flats that are about a 1/2" thick though. I guess parallelism is a bigger issue on large micrometers and there you would have to make a big gage block build up. We have pink and green monochromatic lights for use with optical flats. You can get a mercury vapor or high pressure halogen street light that are used for night lights and that works very well for a monochromatic light.
04-16-2008, 06:53 AM #9
Here's an eBay auction showing just what we're referring to:
As was said you sight through the flat using a monochromatic light to see the fringes. Given the rough rule where measuring equipment has to be 4 to 10 times more accurate then the smallest increment of measurement, 1/10 of 0.0001 = 10 millionths. That's right in the ball park for a single light band of error in parallelism of the mike faces.
As for a monochromatic light, a colored light bulb in a hooded desk lamp works good enough but it has to be the right color (pink-orange) to comply with the de-facto standard. A bug light is close. Clear red is good enough for me and it makes more vivid fringes.
04-16-2008, 06:58 AM #10
You will need an optical parallel to check parallelism of micrometer faces properly, and the monochromatic light as mentioned.
04-16-2008, 08:20 AM #11
Just bought a new set of four Mitutoyo Optical Parallels just like the ones in your link above. New is good on high precision stuff!
Just need to find "the book" tonight
I'm the only one that touches my measuring instruments, were all bought new, and I've never dropped anything but dial calipers, so it will be interresting for sure.
Thanks to everyone for all the help!!
04-16-2008, 09:29 AM #12
How are you doing your mic calibration?
I use ISO 3611(1978), which is the same as whats on the calibration tag that comes with new mic's. increments for measurement are:
x.000 max range
This takes care of full range, and also of 1/4 turns. Good mic's will wring to gage blocks if the faces are flat, I know mine do.
04-16-2008, 02:24 PM #13
Thanks for the info. I downloaded a copy of a standard procedure for micrometer certification a while back. Was going to look at it this morning and forgot. Can't remember if it was a MIL Spec. or an ISO standard. If I remember correctly, I downloaded it from a metrology site where they did calibrations.
I'll read again it when I get home tonight.
04-16-2008, 05:15 PM #14
A good source for monochromatic light is one of the COLORED led flashlights. Not the white ones!
04-16-2008, 06:01 PM #15
Very good point (I shoulda thought of that!). The are pretty common.
04-20-2008, 10:58 AM #16
I have a nice Van Keurken light wave micrometer 1-3 inches that is stated to measure to 2-1/2 millionths.
But not a monochromatic light source. A couple of you guys say the colored led's can be used, not white which are comonly available cheap. Where do you get colored led's and how much are they?
04-20-2008, 01:03 PM #17
LED color is based on wavelength; blue light is about 60% the wavelength of red. If you are shooting for standardization look for LED emmission spectrum close to 632nm which is the helium neon gas discharge light used with optical flats for generations.
04-20-2008, 01:41 PM #18
The most common illumination for optical flats has always been green, at a wavelength of 5461Å. This is the source which is sold by optical flat manufacturers, such as Edmund Optics http://www.edmundoptics.com/onlineca...=1659&search=1
The shorter the wavelength, the greater the accuracy. By definition, green will yield more accurate measurements than red/orange, which is also used.
It appears that modern LED technology has expanded the color range available. LEDs are an excellent source for optical work since they are inherently monochromatic, thus eliminating filter requirements with the attendant loss of brightness.
04-20-2008, 01:58 PM #19
Not to argue but for years Van Keuren was he de facto standard for optical flats and machine shop gaging. Every shop posessiong a lapping machine had a VK monochromatic light and a scratched up flat neatby. Its light was neon pink corresponding to 630 nm (6300 Angstroms then). Given 630 nm another defacto standard sprang up whereby a single interferance fringe represented 1/2 wavelength of light or 11.4 microinches (I think. I gotta look it up.)
If green is the light freq to use then that's a new one on this old dinosaur. Next I suppose you'll say horses are out.
04-20-2008, 02:45 PM #20
That's correct. I was just looking through the 1948 Van Keuren catalog/handbook, in which they introduce that light source. It mentions that the company pioneered optical precision measurement in 1920.
The red/orange monochromatic source uses a helium-filled discharge tube operating at high voltage (3000 volts). This is one of the few natural sources of monochromatic visible light. I believe they still offer a source of this type. That is implied by the photo on their website, but they give no specs on the source.
Another commonly-available source was the high-pressure mercury vapor lamp. These produce good intensity, but require heavy filtration to achieve monochromaticity, and require substantial power, thus generating heat. And they're expensive. I don't know what color was used for measurement, since I've never seen a living example of this product, nor a spec sheet.
I don't know where this fits chronologically WRT the helium discharge tube. I suspect mercury was the first available source. Mercury lamps were in common use for industrial illumination in the 1920's.
The third version used small fluorescent bulbs of the desk lamp variety, coupled with a precision green monochromatic filter, since these lamps produce almost no energy in the red portion of the spectrum. These became quite popular due to low cost, simplicity, low power consumption, and very low heat. The Edmund product which I linked above is typical. Green has the added advantage of producing higher accuracy, as I mentioned previously.
The current state-of-the-art uses LEDs. I'm sure this will become predominant in the near future just because LEDs are overtaking all other technologies as the light source of choice.
But rest assured that horses are still very popular.