Log entry time 22:21:36 on March 30, 2010

Entry number 314764

I noticed this in pseudo-production from Sunday night, but didn't get around to logging it until now. There is about 2k+ intensity noise - much more than we had when we started a week ago, I think - and there are also periodic variations in intensity.

Figure 1 is a plot of bcm3 vs. event number, for Run 2405 (taken last night on owl shift). There certainly appears to be some periodic drop in intensity. How periodic?

I zoomed in to the first beam-on period, (m_ev_num<15000) and plotted the M tree avg_bcm3. I counted 10 periods of the sharp drop outs, and measured the corresponding event numbers, to calculate a period (exactly 1200 events = 5 seconds). Starting from the first drop, I marked a line every 1200 events. This is figure 2, and the result is pretty clear.

Given that this is 5 seconds on the nose, there must be some system causing it. I think there is some mention of this in the accelerator elog as well, but I don't have access to it right now. Also, this is not obvious a problem (doesn't show up badly in charge asymmetry, see figure 3 of charge asymmetry vs time). The large bcm width is an uncomfortable problem (and out of spec, if anyone is keeping track... not important yet). Maybe the air conditioner in the source laser hut is on again?

Figure 1

Figure 2

Figure 3