Null tests for oblate spheroids

In most real cases requiring simple null optics, the optical path difference cannot be reduced to zero but can be kept at a fractional wavelength level so that interferometric data reduction can be used to account for the residual error. In other cases, computer-generated holograms may be used to obtain apparently straight fringes when the desired surface is obtained. Two examples, one involving an f/2.5 concave oblate spheroid and the other a Paul-Baker secondary, are examined. It is shown that although the null tests are not generally perfect, the residual error is small and the tests are simple.