Simulations
Unlike most particle physics experiments, which can be brought to a test beam, atmospheric Cherenkov telescopes (ACTs) cannot be directly calibrated. The combination of the nature of electromagnetic airshowers, cherenkov radiation, and the complex geometry of our optics make an analytical solution of the total efficiency of Solar Two impossible. This forces us to rely heavily on simulations to understand detector efficiencies and energy response. Our simulations consist of four Monte Carlo programs piped together so that the output of each can be studied individually and the interaction between them can be studied as a whole.
The Programs
The first program in the simulation chain is kascade. Kascade is a Monte Carlo simulating the physics of the particle showers that occur in the atmosphere. It traces out segments of the paths of all the particles created as the gamma rays interact with nuclei and produce electron positron pairs.
The output of kascade is piped in as input to cherenk, our second program in the chain. Cherenk is a Monte Carlo simulating the physics of cherenkov radiation. The atmosphere is modeled, the index of refraction and emission angles calculated, and atmospheric absorption and the effects of the earth's magnetic field are taken care of. Cherenkov light is propogated from particle segments down to the ground.
The output of cherenk is piped in as input to s2optics, the third program in the chain. S2optics is a Monte Carlo simulating all of the optical components of Solar Two. It reflects photons from cherenk off of the heliostats and ray traces them to the secondary mirror, reflects them, ray traces them to the camera and then through the winston cones attached to the faces of our photomultiplier tubes. S2optics has played a crucial role in the calibration of our experiment. Simulated photons from the sun are used to test the size and shape of the sun reflected off of a heliostat onto the side of our tower. In reality, we do this regularly to make sure our heliostats are pointing and tracking as they should.
S2optics can also simulate photons from stars as they drift across the field of view of a heliostat. A simulated signal as a star drifts across our field of view is compared to actual star drift data as a further test that the total throughput of simulations matches reality.
Simulated star drift scan compared to actual star drift signal (notice the excellent agreement in RMS and general shape of the signal as the star crosses the field of view)
y-axis - number of counts; x-axis - displaced angle from heliostat pointing coordinate in degrees this indicates the each heliostat has ~1 degree field of view
Information about the simulated photons that make it all the way through the optics is piped to s2elec, the last program in our simulation chain. S2elec gives signal photons a pulse shape characteristic of those produced in Solar Two's photomultiplier (PMT), preamp, amplifier chain. It simulates background (night sky/electronic) noise, discriminator, and trigger. Statistics about the original gamma ray, shower statistics and electronics response are maintained for all showers that trigger.
Comparisons between simulated noise rates and actual noise data for a range of voltages applied to the PMTs and for a range of discriminator threshold settings show excellent agreement.
Noise rates vs. threshold settings reveal that our background is 0.6 photo-electrons(PE)/channel/nanosecond and our threshold is ~4 PE (the black line is simulation, colored lines actual data)
Noise rates vs. high voltage setting (notice the excellent agreement between simulations and data)
