Make New Entry, Make Followup Entry
User name D. Parno
Log entry time 02:35:17 on May 19, 2009
Entry number 274661
keyword=Compton Cavity Performance Depends on Laser PolarizationThe Compton cavity has been unstable over the past few weeks. It appears that this instability is worse for one laser polarization state (laser RIGHT circularly polarized) than for the other. We have noticed that asymmetry statistics have been worse for laser RIGHT, and in some cases the asymmetry plot for laser RIGHT does not make any sense at all.
To explore what's going on here, Gregg and I took a look at trigger rates (as recorded by the new, CMU FADC-based Compton crate) over the past few months. In all the following figures, the red histogram represents a right-polarized laser; the blue histogram represents a left-polarized laser; and the violet histogram represents the laser off state. Laser on/off is determined via a combination of a real-time logic signal and an EPICS variable. No scaling has been applied to any figure.
Fig. 1 shows trigger rates as they looked during normal d2n running (four-pass). You can see that the rates look essentially identical for the two laser state polarizations. You can also see that each laser polarization has a small secondary peak at the same location as the laser-off (background) peak, meaning that those events have been mislabeled as laser on. We have traced this to a fault in the logic signal (real-time cavity power variable).
Fig. 2 shows trigger rates from 10 days ago (2-pass). Although the primary peaks for laser left and right are still quite similar, we now see that only the laser right histogram shows an erroneous secondary peak at the laser off state. Suddenly, the real-time cavity power variable is only misbehaving when the laser is polarized RIGHT. Also, the misidentification is a larger effect, with an amplitude about 1/6 that of the primary peak (as opposed to less than 1/10 for the earlier data).
Fig. 3 shows trigger rates from yesterday (2-pass). Again, only laser-right events are misidentified as being cavity on. And this time, there are marked differences between the trigger rates for laser left/right, cavity truly on. The peak of the laser right state has shifted perhaps 150 Hz to the left, indicating that either (1) the cavity power is lower for laser RIGHT or (2) the intracavity photon beam is less well aligned with the electron beam at the Compton interaction point for laser RIGHT.
These results reflect informal observations that the cavity appears to
struggle more with locking to the right-polarized laser beam, and that
the cavity performance has degraded over time. It is curious that the
real-time cavity power measurement appears to behaving differently, as
well. This logic was not changed after d2n.
A copy of this log entry has been emailed to: franklin, nanda, camsonne, rom
Figure 1
Figure 2
Figure 3
