User name R. Michaels
Log entry time 17:04:12 on March 7,2007
Entry number 195088
Followups:
keyword=raster current pattern explainedFirst of all, I adjusted the offsets so that the raster data are
far away from pedestal. I may have accidentally bumped the screw
driver for the BPM signals too. They look ok, but watch for possible
pedestal shift after this point.
The raster current has spikes. I looked on a scope. The signal
is sent upstairs and has a peak-to-peak size of 45 Volts and the
correct time structure (triangle wave, 25 kHz frequency). It
goes into amplifiers which adjust the signal to the dynamic range
of the ADC. Resulting spectrum is fig1. The output of the amplifiers
is like in fig 2 (a slight exaggeration). It has oscillations which
effectively "freeze" the value for a period of time of order 1 usec.
Beginning at the top of the cycle, the time zones labelled 1,2,3,4
etc in the figure are progressivly smaller. This explains precisely
the ADC spectrum (big peaks on the end, smaller in the middle).
Note, other amplifiers in the crate had the same symptom.
The question is what to do ? First issue is : does it matter ?
Experiment can decide. Secondly, Brandon said that GeN didn't
see these spikes, so we should verify that. The difference
in GeN is that we used VME-based amplifiers (no reason, just for
convenience). Maybe they are better. If so, we can deploy them
but it will require scrounging some VME slots and about 4 hours
of work in hall A.
FIGURE 1
FIGURE 2
