.RC. 26th June 2014, 03:16 PM I was looking on youtube last night and the recent discussion on levels, I came across some laser interferometer videos..
They did a good job of explaining the principle... If I have it right, essentially the device works like this...
You have the laser unit that emits a pure known wavelength of light... This goes into the interferometer unit... The interferometer unit splits the beam, with it sending one beam to the receiver unit and sending the other beam on it's way to the reflector.. The reflector some distance away reflects the beam of light straight back to the interferometer which also then shoots that beam back to the receiver unit... So there are two beams of light going into the receiver unit.
What the receiver unit does is look at where the two beams converge, and it will either produce a light spot of a dark spot depending on the distance the reflector is from the interferometer.. This is because light travels sort of as a wave... So if when the two beams hit the detector unit in the receiver both hit at the top of the wave or the bottom, they produce a light spot on the detector... If they are out by 180 degrees, it is dark... (or is it the other way around)...
So since the wavelength distance is known, the distance the reflector has to move to turn a dark spot into a light spot is also known and just has to be counted.. So 100 000 dark/light spot counts equals x amount of distance... Hence why the units are so extremely sensitive..
All well and good for measuring distance...
So how do they measure angles?
RayG 26th June 2014, 03:44 PM It's all done with mirrors.... :) I'll see if I can find a good diagram.
Ray
RayG 26th June 2014, 04:08 PM See if this explains the method.. lots of mirrors...
317937317938317939
The setup is the same as flatness measurement, when we did the surface plate calibration, it is the same as using differential levels.
In 10x mode the angular resolution is 0.1 arc seconds or 0.5 millionths per inch.. the range is limited to +-10 degrees.
When you are measuring for squareness, that is, the angle you want to measure is 90 degrees you use the optical square.
the actual meaurement is the difference in path length caused by the corner reflectors being at an angle, so you are measuring the angle.
Here's a better diagram showing that.
317940
Ray
eskimo 26th June 2014, 04:28 PM all this stuff about dark and light begs this question... which is faster...dark or light...:D
RayG 26th June 2014, 04:35 PM It's worth putting that into context of mm/meter, when you are talking about levels.
0.1 arc seconds is 0.00048 mm/meter 5 millionths in 10 inches
1.0 arc seconds is 0.00485 mm/meter
5.0 arc seconds is 0.02424 mm/meter ... that's about where the most sensitive machinist vial levels start
20.0 arc seconds is 0.09696 mm/meter ... ( about 1 thou in 10 inches )
Ray
Another way of looking at these ridiculously small angles..
10 microns/meter is 2 arc seconds.. average dust particle is 2-10 microns,
So if you had a single dust particle under one end of a meter long straight edge that would be an angle of around 2 arc seconds..
RayG 26th June 2014, 04:36 PM all this stuff about dark and light begs this question... which is faster...dark or light...:D
Douglass Adams answered that one, nothing travels faster than bad news...
eskimo 26th June 2014, 04:40 PM It's worth putting that into context of mm/meter, when you are talking about levels.
0.1 arc seconds is 0.00048 mm/meter 5 millionths in 10 inches
1.0 arc seconds is 0.00485 mm/meter
5.0 arc seconds is 0.02424 mm/meter ... that's about where the most sensitive machinist vial levels start
20.0 arc seconds is 0.09696 mm/meter ... ( about 1 thou in 10 inches )
Ray
can we just stick to thou..or microns....i have enough trouble with these without bringing another measurement into it
how do you know all this stuff Ray
.RC. 26th June 2014, 04:58 PM Can the interferometer determine the direction of angle, or does it only know an angular displacement has taken place, but not the direction..
I also see my above explanation of the workings was incorrect... The interferometer device does not send two light beams back to the detector, but it is the interferometer itself that does the recombining of the split light beam and sends that signal to the detector, and the detector reads it...
This new explanation explains to me how the angular measurement works as changing the distance between the interferometer and the reflector does nothing as far as changing the light/dark bits..... Only angular changes will change the light/dark bits...
If we wish to also discuss level sensitivities... A Talyvel 1, I think measures to 0.2 arc seconds, the Talyvel 2 and above does 0.1 arc seconds and better... I believe for surface plate checking you need 0.1 arc seconds or better....
RayG 26th June 2014, 04:59 PM how do you know all this stuff Ray
I saw Hitchhikers Guide to the Galaxy on TV :D
RayG 26th June 2014, 05:04 PM Can the interferometer determine the direction of angle, or does it only know an angular displacement has taken place, but not the direction...
You get the direction as well, depending on which beam is longer or shorter the +-sign on the display changes. ( there is a reverse switch on the front panel if you want it the other way round )
Stustoys 26th June 2014, 05:21 PM So, is the distance between the retroreflectors printed on it? or is there some other way to work that out?(or I guess they could all come from the factory set to a standard number). With out knowing that to some stupidly small number you have a problem right?
Stuart
GSRocket 26th June 2014, 05:21 PM It's all done with mirrors....
Ray
Smoke and mirrors...apparently. Light and dark, as clear as day and night.:no:
RayG 26th June 2014, 05:24 PM If we wish to also discuss level sensitivities... A Talyvel 1, I think measures to 0.1 arc seconds, the Talyvel 2 and above does 0.1 arc seconds and better... I believe for surface plate checking you need 0.1 arc seconds or better....
GQ had a Taylor Hobson, don't remember the model... might be still kicking around somewhere?
.RC. 26th June 2014, 05:25 PM You get the direction as well, depending on which beam is longer or shorter the +-sign on the display changes. ( there is a reverse switch on the front panel if you want it the other way round )
Ahh OK... I just found Agilent/HP must have all the manuals online and they mention a doppler effect... This must be how they measure beam length to determine direction of angle...
I wonder how many millions of dollars they poured into research way back when to create the laser interferometer..
GQ had a Taylor Hobson, don't remember the model... might be still kicking around somewhere?
It is a Talyvel 1... I had to edit my above post, the Talyvel 1 was 0.2 arc seconds, not 0.1 as I originally mentioned... AFAIK Greg still has his..
RayG 26th June 2014, 05:34 PM So, is the distance between the retroreflectors printed on it? or is there some other way to work that out?(or I guess they could all come from the factory set to a standard number). With out knowing that to some stupidly small number you have a problem right?
Stuart
The distance between cube corners is standardized at 2.0625 inches.. known to within 1 part in 10,000 ( that's their quoted tolerance ) so +- 2 tenths, and yes, that affects the ultimate accuracy.
So if you couldn't know the absolute angle to better than 1 in 10,000.. worst case would be when you are measuring close to the 10 degree limit of it's range, 1 in 10000 would be 0.001 degrees ( ~3 arc seconds ) but the error scales with the measurement, so smaller angles would have greater absolute accuracy? ( I think that's right.. )
Ray
| |