CAMAC Progress Report

                    Week Ending: November 30, 2001
                               Week: 44


Tom Folkers:
     1. Spent week on site testing new hardware and software.
     2. Found the computer room to be a very noisy environment. Much 
        work will need to be done to squelch the noise to a reasonable 
        level.

Tom Sargent:
     1. I got the last bug out of the servo code that was revealed by 
        the test setup and we took it to the telescope. 
     2. The system did not perform any better than the previous 
        approach. It overshoots about 100 arcsec.  Close analysis of 
        the commanded velocity curve shows a very regular 200 
        arcsec/sec oscillation with a period of about 5Hz.  I think 
        this is at least in part because the algorithm expects very 
        low noise in the velocities and we discovered that the noise 
        in the encoders signals is significant.
     3. We also looked at the noise in the encoder signals. It is 
        rather large - about 0.5%. When the telescope is stationary, 
        the encoders show about 200 counts of noise out of 32768.  If 
        one axis is moved while the other is held stationary, the 
        noise in the stationary one gets worse. Frequency analysis 
        shows prominent 180Hz and 60Hz noise.  This is improved by a 
        factor of 2 by switching to use the old encoder signal cables 
        that go to the Krupp rack.  Another factor of 2 can be 
        obtained by increasing the voltage amplitude at the source 
        from 4.5 v to about 10 and then dividing it down to 4.5 at the 
        A/D input.  We also discovered that the Traquair A/D's do not 
        have 1KHz filters on their inputs as we expected. 
     4. I suspect the servo algorithm can be made to work by not 
        feeding back the error signal on every position measurement 
        cycle. Driving it at 100Hz apparently stimulates the 5Hz 
        velocity oscillation.  Some more damping is needed.  Also the 
        algorithm has to be changed to not be so sensitive to noise in 
        the velocity signal.

Pete Hubbard:
     1. Travel to the SMT for servo testing. With the telescope not 
        moving we collected sin and cos  voltages to look at how much 
        noise is present. The highest amplitude frequency component 
        was 180 HZ with encoder ELV1. The P-P average noise is about 
        200mv. We also tried using the cable that feed's the antenna 
        CAMAC chassis and the noise was 100mv. Out temporary cables 
        were picking up more noise.
     2. Set up new collection system in lab for more testing to see 
        how much noise may be coming from the encoders themselves.

Dave Ashby:
     1. No Report.