Claudio Rivetta and Sten Hansen, Physics Division
Fermi National Accelerator Laboratory Batavia, IL USA 60510
A third measurement trip was made to APO, initially to characterize the rotator axis and to assist Charlie Briegel in setting up to run this axis under the control of an MEI board. The goals were expanded after being informed by French Leger that the azimuth axis preload settings were nowhere near their nominal settings. We were able to get a good data set for all three axes.
The MEI controller can be used to move the rotator axis. A tachometer feedback path for the rotator must be implemented. The friction of the azimuth is double what it was in January and the rotational asymmetry of the azimuth friction has disappeared.
On Wednesday Angela Prosapio and I unpacked the instruments, brought them to the 2.5m telescope enclosure, hooked up the Glentek amplifier and Inland Motor tachometer on the elevation axis, and took a lot of data. Looking at the data that night, it was evident that it differed greatly from the data we took in January. On Thursday morning we found squashed moth guts on the drive disk. After cleaning the disks, a new data run produced results that closely matched our January numbers.
On Thursday morning I took more elevation data and that afternoon Angela and I moved the power amplifier and tachometer to the rotator axis. We collected all the necessary data and we left the system set up for Charlie Briegel's closed-loop tests with the MEI.
On Friday morning Charlie Briegel and I moved the rotator and he read the encoder with the MEI in an open loop configuration. All the pieces worked OK and then during the afternoon we set up and tuned the PID loop. In general it works OK; however, we found the program for tuning the parameters of PID to be inadequate on its own. Something more comprehensive will have to be developed. The other problem is that, due to the harmonic drive reduction mechanism, the motor shaft has an oscillatory response. You cannot see the oscillations on the encoder on the instrument rotator, but by monitoring the motor shaft tachometer voltage the oscillatory response is clearly seen. I'd like to feed back that tach signal into the control loop - either into the analog input to the MEI or into the summing node at the input of the Glentek Box. It is only necessary to run one additional shielded cable from the motor drive to the MEI or Glentek rack. I talked a little to Angela and Paul Czarapata about it. I'll talk later with Paul in more detail.
I made an attempt Friday night at modelling the rotator axis based on the measurements taken during the day, but I had some problems with my simulation. This now becomes a new homework problem, but we know that for phase 1a or 1b the rotator can be moved and parked at any arbitrary position using the MEI controller.
On Saturday with Sten now at APO, we started working on collecting azimuth data while Angela and Charlie worked on the instrument lift. (I think this is the first time the schedule on the telescope has two groups working on it simultaneously!) We spent almost the whole morning trying to understand and get rid of noise induced onto the tach signal. By early afternoon, we gave up and decided to take data with the noise and then filter it out "offline". The azimuth axis has almost twice as much friction as we recorded in January and now the friction is symmetric between directions of rotation. But still we found that if someone was walking on the wind screen or on the building floor, the vibration was coupled to the telescope (the velocity signal from the tach was showing low frequency noise). Sten double checked the coupling between the wind-screen and the telescope and finally we found that Carl Lindenmeyer had put four tabs made of stainless steel on the telescope fork that cover the gap between the fork and the wind-screen. These tabs allow the instrument cart to move easily between the wind-screen and the fork. One of them was touching the wind screen floor and coupling the vibrations between both parts. We bent up this tab with a screw driver and after that, you could jump up and down on the wind-screen and not see anything on the tach signal. We lost the better part of a morning but we now know how sensitive the telescope is.
For the remainder of the afternoon and evening we took the rest of the data, and re-took some of the data that had been corrupted by the noise and declared this to be sufficient for this trip. By 11pm we had those instruments we intended to send back to FNAL packed in their boxes.
Last modified 02/23/99