Test Method: In-Vacuum Wavefront Error of a Solid Retroreflector

1Scope

This method covers the determination of RMS wavefront error of a solid corner-cube retroreflector, measured in double pass at 632.8 nm, with the article held at a specified temperature under vacuum.

It does not cover environmental screening. Screening establishes that an article survives an environment and returns working; this method establishes what the article does while the environment is applied. The two answer different questions and neither substitutes for the other. A retroreflector that comes out of a thermal cycle intact has been asked nothing at all about where its beam is pointing while it is cold, and pointing is the entire function of a retroreflector.

2Apparatus

Source

Stabilized helium-neon at 632.8 nm. The wavelength is a stated condition of every result this method produces, not a footnote to it.

Interferometer

Phase-shifting Fizeau, viewing the article through the chamber window in double pass.

Reference element

Transmission element with a current calibration certificate reporting expanded uncertainty at a stated coverage factor.

Chamber

Thermal-vacuum chamber with an optical-quality window, capable of the profile in section 4.

Fixture

Kinematic mount, treated as part of the article throughout, because it contracts too.

Instrumentation

Article-mounted temperature sensors at not fewer than three points, on the same calibration schedule as the optical instruments.

3Article preparation and the question, written down first

Nothing is measured until the question exists on paper: which parameter, at which condition, to which tolerance, and what disposition follows from the answer. A test run without a written acceptance criterion produces a number, not a result.

The article is cleaned, inspected, and mounted to the fixture at ambient, and an ambient reference measurement is taken through the window before pump-down. That reference is what everything in vacuum is read against, and it is the reason the window term in the uncertainty budget can be bounded at all.

4Environmental profile

Pressure

Pumped to 1×10⁻⁶ Torr at the chamber and held for the duration of the profile.

Temperature range

−180 °C to +150 °C, measured at the test article rather than at the shroud.

Soak

Each set point held until the spread across the article's instrumented points is stable within the acquisition window.

Acquisition points

At each set point, and at ambient before and after the profile.

Rate

Set per article. A ramp rate that induces a gradient the article would never see in service produces a number that is true of the test and of nothing else.

5Measurement sequence

1. 5.1

   Acquire the reference measurement at ambient, through the window, with the article mounted.

2. 5.2

   Pump down. Re-acquire at ambient pressure-reduced condition to separate the pump-down contribution from the thermal contribution.

3. 5.3

   Drive to the first set point. Hold to stability. Acquire not fewer than ten consecutive maps without disturbing the article.

4. 5.4

   Repeat at each set point across the range, in both directions, because hysteresis between the ascending and descending legs is a result and not a nuisance.

5. 5.5

   Return to ambient, repressurize, and re-acquire. A measurement that does not come back to its own reference has told you something, and what it has told you is that one of the other measurements is wrong.

6Data reduction

Piston, tilt, and defocus are removed. The analysis mask is defined once, in the written procedure, and applied identically to every map in the campaign: where the mask is drawn changes the RMS that is reported, which is why mask placement appears as a term in the uncertainty budget rather than as an operator preference.

The reported value is the RMS of the residual wavefront over the defined aperture, in nanometres and as a fraction of the measurement wavelength. Both are given, because one of them is what the customer's model wants and the other is what the customer's specification is written in.

7Acceptance criterion, and the guard band

The acceptance criterion is stated per article on the customer's drawing. Where this method is exercised against the published build-out target, that criterion is λ/20 RMS at 632.8 nm.

A criterion alone does not decide acceptance. The decision rule does, and the decision rule has to account for the uncertainty of the measurement making it. This method accepts an article when the measured value plus the expanded uncertainty falls inside the criterion, which is a conservative rule that never accepts an article the measurement cannot defend.

The consequence is stated rather than buried: an uncertainty of a given size consumes that much of the tolerance, and an article measuring close to the limit cannot be accepted under this rule. That is the correct outcome. The alternative is accepting hardware on a number the measurement cannot support, which is how a part passes on the ground and fails in the environment.

8Uncertainty

Every result reported under this method carries the uncertainty budget that produced it. The worked example published alongside this method names the terms, states each distribution and divisor, combines them, and expands the result at a stated coverage factor.

Stating the budget is how a customer decides whether our number is good enough for their margin. A value without a budget is an opinion.

9Traceability

Every instrument in the setup sits on a calibration schedule with an unbroken path to national standards, and the certificate for each is retained with the campaign record. Traceability is not paperwork about the measurement. It is the reason the measurement means anything outside the building that produced it.

No laboratory accreditation currently covers this method. When an accreditation exists, its scope, and the calibration and measurement capability uncertainty it states for this parameter and range, publish next to this method rather than behind a logo on a separate page.

10Records

1. 10.1

   The written question and acceptance criterion, dated, from before the campaign ran.

2. 10.2

   Every acquired map, unreduced, with its environmental record and its instrument state.

3. 10.3

   The reduction parameters, including the mask definition, as applied.

4. 10.4

   The uncertainty budget as evaluated for this campaign, not as copied from this document.

5. 10.5

   Calibration certificates current at the time of measurement for every instrument in the chain.

11What this method does not establish

1. 11.1

   It does not establish that an article will survive the environment. That is screening, and it is a different test.

2. 11.2

   It does not establish performance at any condition not on the profile. A value at −180 °C says nothing about −190 °C, and interpolating between set points is a modeling decision the customer makes, not a measurement this method delivers.

3. 11.3

   It does not establish lifetime, cycling behaviour beyond the profile run, or performance after storage.

4. 11.4

   It does not establish anything about a different article of the same part number. Population statistics require a population, and this method measures one article at a time.

5. 11.5

   It has not been exercised. Nothing has been qualified under it, and it carries no accreditation.