First Year Performance Paper - Section 3.3
3.3 Timing calibration
The time calibration of the DOM is based on a procedure called Reciprocal Active Pulsing , in which the phase and frequency of each DOM's free running local oscillator (20 MHz) are determined relative to a master GPS-controlled oscillator at the surface by transmitting a fast bipolar pulse at known intervals from the DOR card to the DOM. After receiving this pulse the DOM "reciprocates" and sends an identical fast bipolar pulse after a known delay interval to the DOR card. The time calibration waveforms are produced and digitized by the same DACs and ADCs (operating at 20 MSPS) used for digital communication. The 2.5 km long cable slows the rise time of the original fast calibration pulse from a few nanoseconds at the source to a microsecond or more at the receiving end. Fig. 8 shows the received time-calibration waveforms.
The accuracy of this calibration method arises from the reciprocity or symmetry of the pulsing system, in which the pulses generated at each end have the same shape and the dispersed, attenuated pulses received at each end have the same shape. In this limit, the one-way transit time is equal to one-half the round-trip time minus the known delay, regardless of which feature of the waveform (e.g., leading edge or crossover) is taken as the fiducial mark. Reciprocity is achieved because the calibration signals follow the same electronic path through the same components for both transmission and reception. Reciprocity is also verifiable, because the calibration waveforms are digitized in both the DOM and DOR; these waveforms can be compared to determine any differences in shape that might introduce systematic errors. The round-trip travel times for different DOMs depend on the electrical length of the cable connecting the DOR and DOM. This is illustrated in Fig. 9. The transit times increase for deeper sensors on the string (DOMs with numbers 1-60), and are similar to each other for the surface DOMs (shown as numbers 61-76). At present, calibrations are done automatically every three seconds, although longer intervals up to ten seconds may be used in the future. The variation in the round-trip travel time from one calibration to the next provides the basic measurement of the precision of the time calibration procedure. The best resolution is obtained using crossover timing, for which the typical roundtrip resolution is less than 2 ns RMS (Fig. 10).
An LED on the main board is used to measure the photo-electron transit time in the PMT and other delays in the photon signal path. The statistical and systematic precision with which photon arrival times can be determined depends on additional factors, such as PMT performance and waveform analysis. This is measured in experiments using flasher signals and down-going muons to be described in subsequent sections.