What's new
What's new

High-precision spheres

The real Leigh

Diamond
Joined
Nov 23, 2005
Location
Maryland
How would you like to get an RFQ for spheres with an accuracy spec of two tenths of a millionth of an inch? :nutter:

I wonder how you'd measure such an error limit. :willy_nilly:

Story on the BBC website http://www.bbc.co.uk/news/science-environment-13286241

- Leigh
 
Possibly the tolerance is due to the "zero of ignorance"?

My first question to the client would be: "How are you gonna measure this?"

Arminius
 
Last edited:
Possibly the tolerance is due to the "zero of ignorance"?

My first question to the client would be: "How are you gonna measure this?"

Arminius

You set the ball on a 3 point contact ...sorta like a V block, but 3 contact points

Bowling balls are...well they were ground in a Cincinnati Centerless Grinder line of 3 machines... and then final finishing was a Centerless Lapping....but that accuracy requested?... ummm, lotza luck
 
Their balls may be round, but do they match?
If you asking about the bowling balls... yes, well within a couple tenths... .0002 TIR round and diameter for a days run. This was in the mid 1960's ... no telling what can be done hese days
 
Possibly the tolerance is due to the "zero of ignorance"?

My first question to the client would be: "How are you gonna measure this?"

Arminius


To quote from the Stanford university website:-

Challenge: How to measure the roundness of a sphere at the precision level of 1/10th of one millionth of an inch? The British instrument company, Rank, Taylor, and Hobson, created the Talyrond instrument for measuring the sphericity or roundness of GP-B gyroscope rotors using a stylus mounted on a round spindle to encircle a gyroscope rotor. However, they could not produce a spindle that was itself perfectly round, and thus the spindle introduced error into the measurement.




Solution: Combine the errors in the spindle's roundness with the errors in the sphere being measured. Then, rotate the sphere to a new position and repeat the measurement.
The measurement errors in the roundness of the spindle remain constant, while the measurement errors in the sphere change with each new position. After repeating this process several times, it is possible to separate out the constant spindle error. (The spindle roundness must be checked from time to time, to ensure that it has not changed.)
Result: The spindle roundness errors were calculated and stored in a computer, so they could be reused with different spheres. For each rotor, 16 great-circle measurements were made in the perpendicular plane and one final measurement was made around its equator, tying all the vertical measurements together. The spindle errors were subtracted out of the sphericity measurements, and then the sphericity measurements were translated into contour maps.
 
Tolerances like that are typically in the province of interferometric technology, where resolutions can be achieved in small fractions of a nanometer. The process indicated by Mark Rand is conceptually the same across most any type of extreme-precision measurement process, wherein the metrology system error contribution is characterized, and then applied to the series of measurements as a subtracted value. Like any measurement process at that level, environmental stability is absolutely critical.
 
sounds like

To quote from the Stanford university website:-

Challenge: How to measure the roundness of a sphere at the precision level of 1/10th of one millionth of an inch? The British instrument company, Rank, Taylor, and Hobson, created the Talyrond instrument for measuring the sphericity or roundness of GP-B gyroscope rotors using a stylus mounted on a round spindle to encircle a gyroscope rotor. However, they could not produce a spindle that was itself perfectly round, and thus the spindle introduced error into the measurement.




Solution: Combine the errors in the spindle's roundness with the errors in the sphere being measured. Then, rotate the sphere to a new position and repeat the measurement.
The measurement errors in the roundness of the spindle remain constant, while the measurement errors in the sphere change with each new position. After repeating this process several times, it is possible to separate out the constant spindle error. (The spindle roundness must be checked from time to time, to ensure that it has not changed.)
Result: The spindle roundness errors were calculated and stored in a computer, so they could be reused with different spheres. For each rotor, 16 great-circle measurements were made in the perpendicular plane and one final measurement was made around its equator, tying all the vertical measurements together. The spindle errors were subtracted out of the sphericity measurements, and then the sphericity measurements were translated into contour maps.

Sounds like Way scraping on large machines...."Symetrical distribution of errors" ? Or, scraping of a primary true suface. Fun Fun.
 
ivory tower

I made (5) 3 1/2 OR 4" dia monel ball valve spheres with 2" dia hole thru and a hasteloy stem, tolerance was +.0002, -.0000 with some mirror rms finish. I made them on ww2 equipment. Was not super difficult but was very labor intensive. They did not have time to wait 8 mo. for Hitachi to make them, and they had no way to check them after I finished. Being well paid and rushed I produced expected results with a Norton precision ball cutter grinder and hand work but where is the limit to precision? Nothing is really flat or round on a nano scale because the closer you go to particle size, the larger the voids between them. An indicator with a ball end of one molecue dia would fall in large craters on a surface we might deem perfect. A ball resting on a surface would sink in and the ball surface would flatten on a nano scale. I would think so even with diamond depending on the forces. Just imagine what quark soup would look like way down below the particle scale. I know my understanding of this is not exactly clear but these scales are so far beyond the 1/billionth on an inch scale that my ivory tower has been destroyed by muons. Read the story of the gravitational gyro measuring satellites and the precision achieved for that experiment. I thought I did good by hand using leather gloves and 3000H clover compound for those balls. It is all a matter of perspective.
 
ivory tower

I made (5) 3 1/2 OR 4" dia monel ball valve spheres with 2" dia hole thru and a hasteloy stem, tolerance was +.0002, -.0000 with some mirror rms finish. I made them on ww2 equipment. Was not super difficult but was very labor intensive. They did not have time to wait 8 mo. for Hitachi to make them, and they had no way to check them after I finished. Being well paid and rushed I produced expected results with a Norton precision ball cutter grinder and hand work but where is the limit to precision? Nothing is really flat or round on a nano scale because the closer you go to particle size, the larger the voids between them. An indicator with a ball end of one molecue dia would fall in large craters on a surface we might deem perfect. A ball resting on a surface would sink in and the ball surface would flatten on a nano scale. I would think so even with diamond depending on the forces. Just imagine what quark soup would look like way down below the particle scale. I know my understanding of this is not exactly clear but these scales are so far beyond the 1/billionth on an inch scale that my ivory tower has been destroyed by muons. Read the story of the gravitational gyro measuring satellites and the precision achieved for that experiment. I thought I did good by hand using leather gloves and 3000H clover compound for those balls. It is all a matter of perspective.
 








 
Back
Top