


Arrived Monday 8-July-02 and left on Thursday 11-Jul-02


1) Balanced all of the inner loops. 
        Pilot valve #6 was replaced because it had significantly more
phase lag than all the other valves.
        Pilot valve #2, #5, #12,  required 1.0 volts of dither, and #16
required 1.4 volts of dither to get good following. 
        All of the other valves are at 0.5 volts of dither.
        Pilot valve #2 seemed to drift over time, we mechanically
balanced it twice during the week, it probably should be replaced. 


2) Verified inner loop tuning, tried to match responses.


3) Balanced all of the outer loops.


4) Running simulated crossings, showed significant ringing. The system
had a single notch filter set to a center frequency of 10.0 Hz and a
bandwidth of 100.0 Hz . Last week, this filter was not providing enough
protection from bridge resonance and oil column resonance's. This filter
had about 0.023 seconds of delay and 0.85 amplitude at about 2.0 Hz. 


5) The result of one week of iterations was the following:


 4 Notch filters
        #1 -   6 Hz center, 3 Hz band, -1 depth
        #2 - 13 Hz center, 5 Hz band, -1 depth
        #3 - 25 Hz center, 5 Hz band, -1 depth
        #4 - 33 Hz center, 5 Hz band, -1 depth


 Low pass filter on controller output at 15.0 Hz


6) The 6 Hz notch may not be required for control, the strain control
crossings have 25 Hz data and they are being played 5 times slower than
actual, hence they are programming 5 Hz which they then say they don't
like.


7) The PID controllers were all previously set to a P gain of 1.5. It
turns out the center actuators (9 and 10) like a value around 1.15 and
the ends(1, 2, 17, 18) can handle a gain of 2.0. This could be further
optimized, but I ran out of time. Pat thought it was ok. It is the
opposite of what I expected, the lower spring rate in the center should
allow for more gain, right?


8) After returning to MTS, I differenced the code and found that I
forgot to put the 4th notch filter on the write list. So it is not being
written to the settings file. It is defaulted to the 33 Hz setting. I
can build a new version and send e-mail it to Pat, but I am waiting to
see if there are any other changes.


Cheers,
Marlin

PS. Brad and Rod, the updates source is on my computer. 