deletions | additions
diff --git a/subsection_Calibration_Scans_and_Overhead__.tex b/subsection_Calibration_Scans_and_Overhead__.tex
index 374b385..95658cc 100644
--- a/subsection_Calibration_Scans_and_Overhead__.tex
+++ b/subsection_Calibration_Scans_and_Overhead__.tex
...
\begin{enumerate}
\item START ($T_{START}$): a fixed startup overhead (to slew to first source, do SETUP scan) of $T_{START}=10$m,
\item FLUX ($T_{FLUX}$): a fixed flux density and polarization calibration (angle if high polarization, leakage if low polarization source), of duration of $T_{FLUX}=$2m. This (or the POL source) serves as bandpass/delay calibrator,
\item POL ($T_{POL}$): a fixed polarization angle or leakage calibration (whichever is not covered by the flux density calibrator), of duration (including slew from previous source) $T_{POL}=$3--8m. This is an alternate bandpass calibrator if needed (if using ALTPOL, this is one of the different PA scans
instead), instead). Not needed if the GCAL and FLUX sources suffice for leakage and/or angle calibration.
\item CAL1 ($T_{CAL1}$): a fixed initial scan on the gain/phase calibrator, including slew to calibrator and to the first OTF position, of duration
$t_{CAL1}=$4--8m, $t_{CAL1}=$4--8m. Note: if bright enough and observed over a good range of parallactic angles, can serve as the leakage calibrator also. In some cases might serve as a polarization angle calibrator.
\item OBS (N X $\Delta t_{GCAL}$): each $\Delta t_{GCAL}=$15--30m block of target observing contains $\Delta t_{GCAL}-t_{GCAL}$ on-source and one gain calibrator scan of duration $t_{GCAL}=$60--120s,
\item ALTPOL ($T_{ALTPOL}$): {optional ---} a polarization leakage calibration of 1 or 2 additional scans of one of the other calibrators, with $T_{ALTPOL}=$2--10m depending on what source is used.
\end{enumerate}
...
X & 30 & 10 & 2 & 9 & 9 & 0 \\
Y & 25 & 10 & 2 & 6 & 7 & 0 \\
Z & 20 & 8 & 2 & 5 & 5 & 0 \\
P & 15 & 8 & 3 & 0 & 4 & 0 \\
Q & 36 & 10 & 2 & 8 & 8 & 8 \\
\end{tabular}
\end{center}
\end{table}
Note that scenario Z to get 20m fixed calibration duration requires locating the first source in mid-wrap South to reduce the starting slew needed, and having nearby POL and target areas, and thus is probably only achievable over limited number of tiles.
Scenario P assumes you can use the GCAL source for polarization leakage (or angle, whichever is not covered by FLUX) and minimal slews, and probably represents the best case.
The OTF scanning itself is subject to overheads, given by the time needed to move from the end of one row and move to and get up to speed scanning the next row. In the current OPT (April 2016) an extra duration $t_{OTF}$ of around 16--18 seconds is required. This greatly increases the final overhead versus the on-target time in OTF for even longer block durations and OTF scan lengths. We think the overheads should be 6 seconds or less. The total overhead factor is given by
\begin{equation}
...
\caption{\label{tab:overheadGCAL} Gain Calibration Overheads (GCAL-OH) recurring at a given interval $\Delta t_{GCAL}$ between calibration, for various calibration scan durations $t_{GCAL}$ (1--2m).}
\begin{tabular}{cccc}
{\bf GCAL Interval $\Delta t_{GCAL}$ (min)} & {\bf GCAL-OH ($t_{GCAL}=$60s)} & {\bf GCAL-OH ($t_{GCAL}=$90s)} & {\bf GCAL-OH ($t_{GCAL}=$120s)}\\
15.0 &
7.4\% 7.1\% & 11.1\% & 15.4\% \\
18.0 & 5.9\% & 9.1\% & 12.5\% \\
20.0 & 5.3\% & 8.1\% & 11.1\% \\
30.0 & 3.4\% & 5.3\% & 7.1\% \\
\end{tabular}
...
\begin{table}[h]
\begin{center}
\caption{\label{tab:overheadFCAL}Fixed overheads for various assumed fixed calibration (FCAL) durations $T_{FCAL}$. ``Block Duration'' $T_{BLK}$ refers to the length of a single self-contained SB or the sum of SB lengths that can share a single fixed calibration.}
\begin{tabular}{ccccc} \begin{tabular}{cccccc}
{\bf Block Duration $T_{BLK}$ (hrs)} & {\bf FCAL-OH
($T_{FCAL}=$15m)} & {\bf FCAL-OH ($T_{FCAL}=$20m)} & {\bf FCAL-OH ($T_{FCAL}=$25m)} & {\bf FCAL-OH ($T_{FCAL}=$30m)} & {\bf FCAL-OH ($T_{FCAL}=$36m)} \\
4.0 &
6.7\% & 9.1\% & 11.6\% & 14.3\% & 17.6\% \\
6.0 &
4.3\% & 5.9\% & 7.5\% & 9.1\% & 11.1\% \\
8.0 &
3.2\% & 4.3\% & 5.5\% & 6.7\% & 8.1\% \\
10.0 &
2.6\% & 3.4\% & 4.3\% & 5.3\% & 6.3\% \\
12.0 &
2.1\% & 2.9\% & 3.6\% & 4.3\% & 5.3\% \\
18.0 &
1.4\% & 1.9\% & 2.4\% & 2.9\% & 3.4\% \\
24.0 &
1.1\% & 1.4\% & 1.8\% & 2.1\% & 2.6\% \\
\end{tabular}
\end{center}
\end{table}
...
\begin{table}[h]
\begin{center}
\caption{\label{tab:overheadALL}Final total overheads for different scenarios.}
\begin{tabular}{ccccc} \begin{tabular}{cccccc}
{\bf Block Duration $T_{BLK}$ (hrs)} & {\bf TOT-OH (CASE A)} & {\bf TOT-OH (CASE B)} & {\bf TOT-OH (CASE C)} & {\bf TOT-OH (CASE D)}
& {\bf TOT-OH (CASE E)} \\
4.0 &
21.0\% 20.3\% & 20.7\% & 28.1\% & 31.2\% & 35.0\% \\
6.0 & 17.5\% &
17.2\% & 23.4\% & 25.3\% & 27.5\% \\
8.0 &
15.7\% 16.3\% & 15.4\% & 21.1\% & 22.5\% & 24.1\% \\
10.0 &
14.7\% 15.6\% & 14.4\% & 19.7\& & 20.8\% & 22.0\% \\
12.0 &
14.2\% 15.1\% & 13.8\% & 18.9\% & 19.7\% & 20.8\% \\
18.0 &
13.1\% 14.3\% & 12.7\% & 17.5\% & 18.1\% & 18.7\% \\
24.0 &
12.5\% 13.9\% & 12.2\% & 16.8\% & 17.2\% & 17.8\% \\
\end{tabular}
\end{center}
\end{table}
...
\caption{\label{tab:overheadCASES}Scenarios for calculating overheads. All these cases assume a $t_{OTF}=$6s turnaround and a OTF stripe length of $10^\circ$ at nominal scan speed $\dot{\theta}=3.3^\prime$/s.}
\begin{tabular}{cccc}
{\bf Case} & {\bf FCAL Duration $T_{FCAL}$ (min)} & {\bf GCAL Interval $\Delta t_{GCAL}$ (min)} & {\bf GCAL Duration $t_{GCAL}$ (sec)} \\
A &
15 & 18 & 90 \\
B & 20 & 15 & 60 \\
B C & 25 & 15 & 90 \\
C D & 30 & 15 & 90 \\
D E & 36 & 20 & 120 \\
\end{tabular}
\end{center}
\end{table}
In the VLASS proposal and TIP sent for review, the ALL-SKY component for the full survey was assumed to incur an overhead of 17\%, while for the pilot design we assumed a 20\% overhead. In order for the pilot to have a total overhead near 20\%, the numbers for CASE
C D (inserting a 90s gain calibrator scan every 15m, with 30m fixed calibration)
indicate that we want to have a total block length of 10 hours or more. If we observe in self-contained 4 hour blocks and insert a 90s gain calibration every 15m with 25m fixed calibration (CASE
B) C) then the overhead will be 28.1\%, or 21.0\% if we reduce the gain calibration scans to 60s and the fixed calibration to 20m instead of 30m (CASE
B), reaching 20.3\% for 90s gain calibration scans every 18m with only 15m of fixed calibration at start (CASE A).
Thus, sharing Note that CASE A is possible, as shown by test schedules on COSMOS. Sharing calibration between multiple 4-hour SB is highly desirable. If we were to use longer 8-hour blocks, then using a 90s gain calibration every 15m with 25m fixed calibration (CASE
B) C) will give a 21.1\% overhead, close to the target value for the pilot.
