These were the steps taken (as well as I can remember them now) to replace the broken piezoelectric actuator (PZT) on the SIS active mirror.

1) The system was not stabilizing the artificial star image well, so we put a sinusoidal signal into the channels of the active mirror high voltage amplifier one at a time to see if the mirror was responding. An offset and oscillating signal is seen on the oscilloscope showing the quadrant detector error signal when the PZT was working. PZT #2 was found by this method to be unresponsive.

2) Swapped amplifier used to feed PZT #2 with one that had been demonstrated to work in step 1. PZT #2 was still not responding.

3) Swapped the cable to PZT #2 at the input to the central mirror assembly with one that had been shown to work in steps 1 and 2. The PZT was still unresponsive.

4) At this point, we knew that the problem had to be inside the mirror assembly, so the mirror slide was removed from the octagon.

5) Checked the feed-through in the mirror slide assembly for continuity and for shorts. Unfortunately, this required cutting the cable to the PZT. The feed-through was found to be operating properly.

6) Removed the SIS mirror from its holder on the mirror slide.

7) Tried to unscrew the PZT from the mirror mount. The PZT screw had been lock-tighted to prevent loosening because of the vibrations of the PZT and the screw was a Phillips head screw, so the head became rapidly stripped. The only alternative was to drill the screw out of the PZT (the PZT are tapped so this does no harm to the mounting). This was done with the mirror in place on the mount. In the future, it would probably be best to remove the mirror before the screw is drilled (hopefully, the screws will not always have to be drilled out to replace the PZTs) to avoid damage to the mirror and since we had to remove th mirror to get the PZT out anyway. It may not be necessary to remove the mirror if the screw comes out properly. After removal of the old PZT, a new one was mounted. Again lock-tight was used, but this time a hex head screw was used to try and avoid the problem of stripping the head. The mirror was then re-mounted.

8) The mirror in its mount was put back onto the mirror slide and re-cabled. The PZT come with a BNC connector attached, but originally, these had been cut off and the co-axial cable (an extra length was put in to account for the motion of the mirror slide) was spliced together. On advice from Grant Matsushige, we left the connector on the PZT cable and installed a connector onto the remaining cable. This will allow for easier PZT changes in the future. (Only the broken PZT cable was modified.) The connectors were wrapped in insulating tape after they were connected together.

9) The mirror slide was then put back into the mirror assembly on the octagon.

10) The PZTs were again tested as described in step 1. All PZTs were shown to be working.

11) Tests were then done to see if the system would stabilize the artificial star image (with a PZT on the artificial star running). The system worked very well at low frequencies (0 - 5 Hz), but performance decreased rapidly at higher frequencies. This is normal according to Olivier LeFevre.

We used the only replacement PZT that we had in replacing the broken one, so I have ordered 2 more for spares. The part is a Micro-Kinetics #6095-25, 0-1000 Vdc, 0-25 m piezo-electric actuator. The spares were ordered from Micro-Kinetics Corp.,1881 McGraw Ave, Irving, CA 93714-5733, (714)476-9911 and were $720 each as of 11 August.