Last Update: JeffMangum - 02 March 2011
See the "Calibration Device Technical Specifications and Reports" section of the Amplitude Calibration
wiki page for historical background, specs, etc. for the amplitude calibration device (ACD).
CASA Development Issues
(Placeholder for any questions/comments from CASA developers regarding ACD system development needs).
Tsys Measurement System Development
By design, we are supposed to be able to make Tsys measurements as a function of frequency "whenever and wherever needed". In the following we attempt to track the development to this goal.
State of the System as of 2011-02-23
The design of the ACD was meant to allow measurement of Tsys at any time. Following some question-and-answer with the folks actually doing this development (RobertLucas
and Stuartt Corder), got the following facts regarding the current "state of the Tsys measurement":
- ACD Latency: ACD use is currently limited by software latency associated with each subscan. For example, from an actual Tsys measurement, it took 57 seconds to complete a set of three 6 minute scans, or about 13 seconds of overhead associated with each 6 second scan. This software latency is expected to be significantly reduced when subscan sequencing is introduced in ALMA 8.1 (due July 2011). Here is the example cal measurement summary:
We can estimate from the data from uid___A002_X19f3fe_X12:
Scan 1 2011-02-13 19:17:52-19:18:49 3c454.3 3 subs. (CALIBRATE_ATMOSPHERE, CALIBRATE_WVR)
-Ss 1 2011-02-13 19:18:07-19:18:13 ON_SOURCE
-Ss 2 2011-02-13 19:18:24-19:18:30 ON_SOURCE
-Ss 3 2011-02-13 19:18:42-19:18:48 OFF_SOURCE
Scan 2 2011-02-13 19:18:49-19:19:09 3c454.3 1 subs. (CALIBRATE_DELAY, CALIBRATE_PHASE, CALIBRATE_WVR)
-Ss 1 2011-02-13 19:18:58-19:19:09 ON_SOURCE
So we get 57s with 6s subscans.
- Tsys Versus Frequency: Even though it is not currently being used, TelCal can produce Tsys measurements as a function of frequency (i.e. not spectral averages). It requires using TDM spectral mode.
Detailed Description of the SMA Tsys Measurements System
(Thanks to Bob Wilson and Mark Gurwell for this information)
Our continuum detectors cover the nominal 2 GHz band of each receiver plus some additional because of the finite
cutoff of the band defining filter. They are interfaced to the antenna computer through a V/F converter which drives a counter. The program syncdet2 reads and resets the counter more frequently (because if the chopper is running it acts as a synchronous detector), but outputs a proper one second integration each second. These data are corrected for detector non-linearity and expressed in milliwatts in the two position array RM_CONT_DET_MUWATT_V2_F. This data stream is the basis of the Tsys calculation.
The observing script has tsys commands from time to time, usually at the time of switching between source and calibrator. The tsys program commands the ambient load (cabin temperature ~18C) into the beam near a waist, averages four 1 sec. measurements after a suitable delay, moves the ambient load out and then stores the result in variables called RM_HOTLOAD_LOWFREQ_VOLTS_F and RM_HOTLOAD_HIGHFREQ_VOLTS_F along with RM_UNHEATEDLOAD_TEMPERATURE_F.
Each second that the receiver has been seeing the sky for the whole second, a separate thread of syncdet2 retrieves the hotload and sky values and calculates Tsys for each receiver:
Tsys (K) = [Vsky/(Vload-Vsky)] * Tamb (K)
The terminology of Volts here is left over from the earlier interface in which instantaneous volts were read, but is actually milliwats. These values are written out each second into RM_TRX_D and RM_TRX2_D.
We have heated loads, 60C if I remember correctly, but due to receiver non-linearities we have not come to terms with using them yet.