Run 9 Data: Trigger Simulator Investigation

A look at the behavior of the trigger simulator (in data) as part of the investigation into raising the jet patch ADC thresholds in the simulation discussed here ...


Below, I look at the maximum jet patch ADC spectra as reported by the trigger simulator in the data jet trees. I look at the ADC spectra for the level 0 L2JetHigh and JP1 spectra for three different categories: didFire=1 && shouldFire=1 (so called 'good' events), didFire=0 && shouldFire=1 (so called 'bad' events), and didFire=1 && shouldFire=0.

For the L2JetHigh events, I look at them both with and without the Level-2 filter condition. When the Level-2 condition is applied, the event must pass either the monojet, dijet, or random requirement.



Figure 1: ADC spectra for the different L2JetHigh shouldFire / didFire conditions. The left panel is from a run with JP1 prescale = 2 (10144091) and the right panel is from a run with JP1 prescale = 38 (10164025). The green curve is shouldFire=1 && didFire=1, the blue curve is shouldFire=0 && didFire=1, the red curve is shouldFire=1 && didFire=0, and the black curve is the sum.




Figure 2: ADC spectra for the different JP1 shouldFire / didFire conditions. The left panel is again for PS=2 and the right panel is for PS=38. The color scheme is the same as figure 1.


The large, prescale dependent variation in the didFire=0 && shouldFire=1 numbers in figure 2 are expected as much of those events will come from events which would have fired JP1 but did not due to the trigger being dropped by the prescale. The behavior of the 'bad' didFire=0 && shouldFire=1 events over the 'good' didFire=1 && shouldFire=1  as a function of prescale can be seen in figure 4.


Figure 1 above shows the behavior of the L0 part of the L2JetHigh trigger. The figure below shows the same spectra except that now the monojet, dijet, or random flags of the Level 2 trigger have to be satisfied.



Figure 3: The top panel shows the L2JetHigh spectra from the prescale=2 run (10144091) without the Level 2 condition (left) and with the Level 2 condition (right). The bottom panel has the same format for the prescale=38 (10164025) run.



Figure 3 shows that the 'bad' didFire=0 && shouldFire=1 L2JetHigh events disappear when the Level 2 condition is included.



Below, I show the ratios of number of counts in the good and bad spectra for the two trigger categories. In all cases, I integrate the spectra above the threshold for the trigger I am looking at. All ratios are shown as a function of the JP1 prescale for that particular run.


Figure 4: Plots show the ratios of the number of counts above the appropriate thresholds. Plot titles say what quantities are being compared.