Technical Notes262 views

From Arc-Seconds to Microns: Reading Conversion and Bridge Span in Autocollimator Measurement

1 arc-second is 4.85 μrad, i.e. about 0.97 μm height difference on a 200 mm bridge; wrong conversion invalidates straightness evaluation.

华旋光电技术部
From Arc-Seconds to Microns: Reading Conversion and Bridge Span in Autocollimator Measurement

Converting angle to length

An autocollimator reads angle directly, in arc-seconds, while straightness and flatness need linear units in microns. The conversion is 1 arc-second = 4.848 μrad: a beam deflected by one arc-second moves the spot about 4.85 μm at a distance of one metre. In bridge-method straightness measurement the mirror sits on a bridge of span L, and the height difference between the front and rear feet is Δh=L·tanθ≈L·θ with θ in radians. Practical conversion factors: for L=100 mm, one arc-second is about 0.485 μm; for L=200 mm about 0.97 μm; for L=500 mm about 2.42 μm.

Choosing the span

A shorter span gives finer height resolution per unit angle but increases the number of stations, the accumulated error and the labour. A longer span is faster but misses short-wavelength undulations. The rule of thumb is a span no greater than one tenth of the measured length and no greater than half the shortest wavelength of interest. A 2 m guideway is usually measured with a 200 mm bridge, where a 0.1 arc-second autocollimator resolves about 0.1 μm in height, adequate for precision machine-tool guideways.

Instrument constants and ranges

A dual-axis photoelectric autocollimator with 0.1 arc-second resolution typically outputs linearized angle values; the operator only enters the bridge span L in the measurement wizard, and the software accumulates Δh=L·θ station by station. Beware that some instruments display a slope value such as 0.005 mm/m, which is itself an angle: 0.005 mm/m equals 5 μrad, about 1.03 arc-seconds. Mixing these two conventions is the most common conversion mistake in the field.

Controlling accumulated error

In step-by-step accumulation, a per-step reading error ε accumulates as √n for random behavior or as n for systematic behavior, with n the number of steps. Ten stations measured with a 0.1 arc-second instrument accumulate roughly 0.32 arc-seconds randomly, about 0.31 μm on a 200 mm span. Long runs therefore call for a higher accuracy class, with indication error within 0.2 arc-seconds, verified periodically per JJG 202 so that indication error and repeatability remain traceable.

⇄0CompareProduct comparison