A Multiplane Foucault Tester
This page will describe the construction of a Slitless Foucault Tester that can be used to test mirrors along any diagonal without having to move the mirror.  It is controlled by James Lerch's RTAFT program.
This is an overview of the tester.  Unique to this tester is the cradle that allows the entire tester to be rotated through 180 degrees, so testing can be done on any mirror diagonal without the need to rotate the mirror itself. The tester is mounted on an optical rail, which is mounted on an optical bench. This provides a flexible but rigid framework for testing.
The test image is captured with a black and white CCTV camera connected to a PCI frame grabber. The camera has a zoom lens to allow a full screen image to be obtained for mirrors with any focal ratio. On a 266MHz PII computer, images can be captured at 10 frames per second.  Although this is a little slow, testing of a 10" f/4.5 mirror using 8 zones can be done in under 9 minutes.
The light source consists of a red LED with variable intensity.  The knife on the RHS can be moved to create a narrow slit when using a ronchi screen or other tests involving diffraction effects.  The source/knife edge can be positioned using the screw visible on the left. In the future, this screw may be controlled by a stepper to provide a computer controlled knife edge.  However, if the tester is accurately aligned with the optical axis of the mirror, moving the knife edge during a test shouldn't be necessary. Below the red LED is a high intensity mini-lamp used to help position the tester and mirror during initial setup.
The tester is mounted on gimbles to allow multi-axis testing.  As the entire tester is mounted on an X/Y table, initial alignment can be set so that a series of tests can be performed with little additional adjustment.
One of the more difficult tasks in mirror testing is the positioning of the mirror, even when using a CCTV camera to view the image. This stand can be remotely controlled in both X and Y, greatly simplifying the initial effort.  

The moveable part of the stand is mounted on trunnions that can be adjusted so the X and Y pivot points are aligned with the front surface of the mirror. Otherwise, with large thick mirrors, the optical axis of the mirror can move as the mirror is positioned.
The platform X and Y axes are controlled by 200 step/rev stepper motors with 30:1gear boxes. In this way, the axes can be positioned to within 0.0001" although typically steps of .0005" are used.
The motors are driven by the Robofoucault program through electonics designed to be used with Mel Bartel's SCOPE.EXE program.
Test data are captured by RTAFT and typically analysed using Dave Rowe and James Lerch's FigureXP program.  A spread sheet is also used to compare measured knife edge positions with theoretical values.
Because RTAFT assigns a value of zero to the knife edge position for Zone 1 when it is first found for a given test, it is only possible to compare actual readings between different diagonals by selecting a repeat option of 4 (say) and pausing between each pass to adjust the diagonal.  Perhaps future versions of RTAFT will allow multiple tests to be run without resetting Zone 1 to zero for each test.