A total of four FGS units were built, an “engineering test unit” (ETU) and three for installation on HST. Prior to the second servicing mission the mechanical health of both FGS1 and FGS2 had degraded to the point that the reliability of each unit became questionable. Moreover, it was recognized that the deleterious affect of HST's spherical aberration, to which the FGSs are still subject, could be partially mitigated by replacing the fixed mounted “fold-flat mirror #3” with an articulating mirror assembly (AMA) that can be commanded from the ground to provide the means to align the FGS interferometric elements with HST's optical axis. Therefore the ETU was refurbished with the AMA and made flight-ready for insertion into HST in Servicing Mission 2 (SM2).

FGS2 had been showing chronic trends of mechanical wear, more so than FGS1, and thus had been slated for replacement in SM2. However, a few months before SM2, FGS1 displayed acute mechanical failure symptoms and appeared to be at greater risk than FGS2. Therefore, the refurbished ETU replaced FGS1, and has since been referred to as FGS1r. With the advantage of the AMA, FGS1r has proved itself superior to FGS3 as an astrometric instrument, and has been used as such since 1999.

Meanwhile, the original FGS1, which was returned to Earth at the completion of SM2, was refurbished with new mechanical components and an AMA. Designated FGS2r, it was installed in HST during SM3A. The original FGS2 was returned to Earth and refurbished with the expectation that it would replace the venerable FGS3 in a future servicing mission. However, since mid 2006 FGS2r has been showing problems with the LED on one of it star selector servos. With a failing LED, which is sensed by the FGS firmware to provide closed-loop control of the instrument, FGS2r experienced an increased guide star acquisition failure rate. Therefore, it was replaced in SM4 with the refurbished unit (the original FGS2).

The refurbished FGS installed in SM4 is referred to as FGS2r2. Unlike FGS1r and FGS2r, its optics had been realigned using the specially developed full field of view test set. This enabled the Goodrich optical engineers to eliminate “beam walk”, an effect that causes the interferometer's relative alignment with the HST optical axis to change as the star selectors assembly is rotated to observe stars across the FGS FOV. This beam walk degrades the benefit of the AMA for mitigating HST's spherical aberration. (The AMA can be used to optimize the FGS performance at any one place, but only one place, in the FOV.) The commissioning of FGS2r2 in June 2009, after the AMA adjustment, resulted in near optimal interferometric performance across its entire FOV. Nonetheless, because FGS1r is well calibrated for scientific observations, and has demonstrated superb performance as a science instrument, it continues to be designated as the HST science FGS.