Fig. 1. The NuTeV neutrino detector showing the target calorimeter followed by the downstream muon spectrometer.

Fig. 2. Geometry of one unit of the calorimeter. This unit is repeated 42 times to make up the entire calorimeter. One unit of the calorimeter consists of a scintillation counter sandwiched between two steel plates.

Fig. 3. A schematic drawing of a NuTeV scintillation counter.

Fig. 4. A schematic diagram of the NuTeV calorimeter readout electronics of a counter.

Fig. 5. Typical energy deposition of muons traversing one scintillator counter in units of ADC counts.

Fig. 6. Average counter response to muons traversing as a function of position in the counter. The coordinates are normalized to the half-width of the counter, 1.5 m, on both axes.

Fig. 7. Fractional energy deposition as a function of position for two phototubes in a given counter.

Fig. 8. Sum of phototube gain coefficients, normalized to the average gain over the entire neutrino run period, as a function of temperature of the most upstream region of the calorimeter.

Fig. 9. Average longitudinal hadronic shower profiles of neutrino events in two different shower energy ranges. It can be seen from the plots that large fraction of shower energy is deposited in two to three consecutive counters.

Fig. 10. Relative hadron/muon counter gains which arise from detector non-uniformities unrelated to scintillator thickness.

Fig. 11. Accelerator time structure. Note that the interval between the last neutrino ping and the slow spill calibration beam is only 1.4 s, allowing essentially an in situ calibration.

Fig. 12. NuTeV calibration beamline schematics. The “Ferris wheel” (NTACON) with four different thickness converter material is used to select pure hadrons or electrons. The 6 m long Be filter(NTBBE) is used to select pure muons. The numbers on the left-hand side of each component indicate the relative distance of the component to the primary target (NT8TGT) in meters. Some beam position and intensity monitoring devices are not drawn in this figure because they are irrelevant for this paper.

Fig. 13. A schematic view of the NuTeV calibration beam spectrometer system. The large distances between the chamber stations allow accurate absolute momentum determination.

Fig. 14. PMT signals of the herenkov counter with 160 Torr nitrogen gas. A clean particle separation between π(C₁), K(C₂), and (ped) at 50 GeV is apparent.

Fig. 15. T₁ vs. T₂ for x-view of an SWDC plane. The cluster of points along the line with slope −1 to the left corresponds to the region between the x and x′ wires; these events are used for track fitting. Non-linear time-to-distance effects are visible for very long drift times and for events very close to the sense wires.

Fig. 16. Difference in coordinates, X₁−X₂ for an SWDC chamber plane from a sample of calibration beam tracks. The solid line represents a Gaussian fit on the distribution. The σ of this distribution is 420 μm, implying a spatial resolution of 300 μm.

Fig. 17. (A and B) Track slopes in x and y at the upstream end of calibration spectrometer magnets as determined by chamber tracking. (C and D) Beam profile in x and y at the upstream end of calibration spectrometer magnets as determined by chamber tracking. (E and F) χ² distribution of the upstream chamber fits in x and y views. Solid lines represent the expected normal χ² distributions. Note that both distributions follow the expected normal χ² distributions.

Fig. 18. χ² distribution for X (A) and Y (B) view of calibration beam spectrometer track fit using chambers upstream and downstream of the spectrometer magnet string. The solid lines represent expected normal χ² distributions. Ten points are used in the three-parameter fit.

Fig. 19. A schematic diagram of the Hall probe and the holder location inside the magnet.

Fig. 20. Comparison of linear coefficients in ∫B dℓ vs. Hall probe readout for different EPB dipole/ Hall probe combinations for the NuTeV calibration beam spectrometer magnets.

Fig. 21. Results of quadratic fit to ∫B dℓ vs. Hall probe readout for a typical EPB dipole in the NuTeV calibration spectrometer. ∫B dℓ/H vs. H, where H is the Hall probe reading in Tesla and ∫B dℓ in Tesla-meters is plotted to show the non-linear region at low fields.

Fig. 22. Ratio of ∫B dℓ calculated from magnet shunt current to the value calculated from the Hall probe for a typical NuTeV calibration run.

Fig. 23. A schematic drawing of the NuTeV calibration beam trigger scintillation paddle.

Fig. 24. Typical HV readout values for four randomly selected PMTs. As one can observe, the HV readout does not vary more than 2–5 V.

Fig. 25. Counter response to muons in calibration beam (histogram) and Monte Carlo (crosses) for muon of energies of 15 (A), 50 (B), 100 (C), and 166 GeV (D). DVTEN (HIGH/LOW channel depending on saturation) are muon map corrected and measured in MIPs. All 84 counters contribute to the distributions. Solid lines represent 5-parameter asymmetric Gaussian fits (Eq. (10)) to the distributions.

Fig. 26. Fit parameters (A, B: mean and RMS of the Gaussian part, C and D: asymmetric tail, Eq. (10)) to DVTEN distributions for Monte Carlo (open circles) and calibration beam muons plotted versus incident muon energy (E_μ).

Fig. 27. Counter response to muons summed over 84 counters in calibration beam (histogram) and Monte Carlo (crosses) for muon of energies of 15 (A), 50 (B), 100 (C), and 166 GeV (D).

Fig. 28. Fit parameters (A, B: mean and RMS of the Gaussian part, C and D: asymmetric tail, Eq. (10)) to DVTEN SUM distributions for Monte Carlo (open circles) and calibration beam muons plotted versus incident muon energy (E_μ). DVTEN SUM is the sum of the DVTEN over 84 counters.

Fig. 29. Mean values of DVTEN (A) and DVTEN SUM over 84 counters (C) distributions for Monte Carlo (open circles) and calibration beam muons (solid circles) as well as RMS values of these distributions (B and D).

Fig. 30. Method α: The width (A) and mean (B) from Gaussian fit (solid circle) and histogram statistics (stars) of the difference between “true” ΔE and reconstructed ΔE distributions in muon energy loss bins. (C and D) The same in muon momentum bins.

Fig. 31. Method γ. The fitted widths (A) and means (B) of the difference between true ΔE and reconstructed ΔE distributions in muon energy loss bins. (C, D) The same in muon momentum bins.

Fig. 32. Energy dependence of calibration beam muon depositions of pulse heights below 5 MIPs, over 84 counters, shown in units of GeV/MIP and the parameterization C_μ(E_μ).

Fig. 33. Most probable (A) and mean value (B) of total muon energy loss in NuTeV calorimeter. Comparison of GEANT prediction (open circle) and calibration beam measurement (solid circles) versus muon energy.

Fig. 34. Ratio of the most probable values of ΔE for 50 GeV calibration beam muons to GEANT prediction (15.2 GeV) as a function of muon azimuthal angle.

Fig. 35. Cumulative fractional shower energy as a function of the added number of counters. Note that for all energies, adding 20 counters provides full longitudinal shower containment.

Fig. 36. Remaining time dependence of the ratio of reconstructed calorimeter energy divided by measured beam momentum. The left graph is for 50 GeV runs, and the right is for 100 GeV runs. The horizontal axis is time in units of blocks, where each run block corresponds to a period of about 2 weeks. The responses are obtained after all the time-dependent corrections are applied to the data, but before the final energy scale is set.