{"id":6804,"date":"2017-03-24T11:41:35","date_gmt":"2017-03-24T21:41:35","guid":{"rendered":"http:\/\/www.eaobservatory.org\/jcmt\/?page_id=6804"},"modified":"2017-03-24T12:58:22","modified_gmt":"2017-03-24T22:58:22","slug":"heterodyne-observing-modes-more-details","status":"publish","type":"page","link":"https:\/\/www.eaobservatory.org\/jcmt\/instrumentation\/heterodyne\/observing-modes\/heterodyne-observing-modes-more-details\/","title":{"rendered":"Heterodyne Observing modes – more details"},"content":{"rendered":"

Main Heterodyne Observing modes<\/a> page<\/p>\n

(Note: the loop control parameters mentioned below cannot be changed in MSBs; the TSS can modify some of them if necessary)<\/p>\n

Jiggle-chop observations<\/h3>\n

Duration:<\/h4>\n

The figure below illustrates the duration of observations based on the UT timestamps in the headers. The duration scales linearly with the requested integration time for single-point beam-switched observations (black dots and line). JCHOP maps with separate offs<\/i> (red dots) follow the same relation which is not surprising since the motion of the secondary is not a dominating factor. JCHOP maps with shared offs<\/i> also scale linearly with the requested integration time (blue dots and blue line), but require significantly less time than their separate counterpart and, as expected, this effect scales non-linearly with the number of points in the jiggle (cyan points — 3×3, 4×4, 5×5 pattern, respectively — and blue line). The durations in the plot include an additional ~60secs before the start UT is written or closing the files after the end UT timestamp. The approximate relations thus become (np: number of points in jiggle; t_on: requested ‘on’-only integration time per jiggle-point):<\/p>\n

\n\n\n\n\n\n
Jiggle-Chop<\/th>\nDuration (secs)<\/th>\n<\/tr>\n
Separate off<\/i> or 1×1 BMSW<\/td>\n2.3 * np * t_on + 100<\/td>\n<\/tr>\n
Shared off<\/i><\/td>\n1.27 * [np +sqrt(np)]* t_on + 100<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n

\"jchp_time\"<\/a><\/p>\n<\/div>\n

<\/a><\/p>\n

Noise:<\/h4>\n

The figure below shows the noise in Jiggle-Chop observations. The top panel compares the observed noise to the expected noise, for chopped observations ranging from single-position to 5×5 jiggles, both using shared<\/i> and separate offs<\/i>. Overall the agreement is excellent, but the data indicate an excess noise of about 4% (the black line).<\/p>\n

The expected noise can thus be calculated using:<\/p>\n

    rms = 1.04 * Tsys*1.23\/sqrt(df*t_on) * sqrt(1+1\/sqrt(np))\r\n\r\n    1.04: 4% excess noise, presently unexplained\r\n    1.23: Correlator digitization factor\r\n    df:   channel-spacing in Hz\r\n    t_on: 'on'-only integration time in sec\r\n    np:   1 for separate offs else the number of point \r\n          sharing an off. For patterns not exceeding 9x9\r\n          this is the number of points in the jiggle pattern.\r\n\r\n<\/pre>\n
The bottom panel shows the normalized noise: rms*sqrt(int. time) as a function of the duration of the observations with the noise calculated using the equation above indicated by the black crosses. As expected the normalized noise is independent of the duration and type of the observations, indicating no unexplained sysystematic effects, with one significant exception: all observations with shared<\/i> offs have about 20% less noise than their separate<\/i>
\ncounterparts. Together with the previous figure above, this panel illustrates the point made when discussing separate offs<\/i>: not only do observations with shared<\/i> offs take less time for the same amount of integration time, the resulting noise is lower as well.<\/p>\n
\"jchp_rms\"<\/a><\/p>\n
Jiggle-Position Switch observations<\/h3>\n
Loop control:<\/h4>\n
There are two parameters that control the execution of the JPSSWobservation: the step_time<\/i>, the max_time_between_refs<\/i>. As for JCHOP, the step_time is the dwell time per jiggle position and has a minimum of 0.1s. The max_time_between_refs is the maximum time that can be spend observing jiggle positions before a position switch to the sky reference is require; its current default is 30s. Note that in order to minimize the overhead of the position switch, the rate of going to the sky position is less than for JCHOP which can result in worse baselines.<\/p>\n
The dwell time per jiggle position is calculated such that it is as long as possible to allow the pattern to be completed within the max_time_between_refs, but not to exceed 2s per point. As for JCHOP there is a continuum mode<\/i> which changes the max_time_between_refs to 5s at the cost of more overhead. Note that position-switch observations are not well suited for measuring the continuum level, both because of their typically larger switch, as well as the relatively long time between sky references even in continuum mode.<\/p>\n
Duration:<\/h4>\n
The figure below illustrates the duration of observations based on the UT timestamps in the headers. The duration scales linearly with the requested integration time for single-point position-switched observations (black squares and line). JPSSW maps with separate offs<\/i> (red circles) follow the same relation which is not surprising since the motion of the secondary is not a dominating factor. JPSSW maps with shared offs<\/i> also scale linearly with the requested integration time (blue circles and blue line), but require significantly less time than their separate <\/i>counterparts. Contrary to JCHOP observations, and for reasons not obvious, the same linear relation fits patterns of different size (cyan circles — 4×4, 5×5 pattern, respectively — and blue line). The durations in the plot include an additional ~60secs before the start UT is written or closing the files after the end UT timestamp. The approximate relations thus become (np: number of points in jiggle; t_on: requested ‘on’-only integration time per jiggle-point):<\/p>\n
\n\n\n\n\n\n
Jiggle-PSSW<\/th>\nDuration (secs)<\/th>\n<\/tr>\n
Separate off<\/i> or 1×1 PSSW<\/td>\n2.45 * np * t_on + 80<\/td>\n<\/tr>\n
Shared off<\/i><\/td>\n1.75 * np * t_on + 80<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
 <\/p>\n
\"jpsw_time\"<\/a><\/p>\n<\/div>\n
 <\/a><\/p>\n
Noise:<\/h4>\n
The figure below shows the noise in Jiggle-PSSW observations. The top panel compares the observed noise to the expected noise, for PSSW observations ranging from single-position to 5×5 jiggles, both using shared<\/i> and separate offs<\/i>. As for JCHOP the data indicate an excess noise of about 4% (the black line). The bottom panel shows the normalized noise: rms*sqrt(int. time) as a function of the duration of the observations with the noise calculated using the equation above indicated by the black crosses.<\/p>\n
See Jiggle-Chop for a more detailed discussion<\/em><\/a><\/p>\n
\"jpsw_rms\"<\/a><\/p>\n
Grid Position-Switch<\/h3>\n
Loop control:<\/h4>\n
There are two parameters that control the execution of the GPSSW observation: the step_time_grid_pssw<\/i>, the max_time_between_refs<\/i>. As for the other modes the step_time is the minimum<\/i><\/b> dwell time per point, but for GPSSW it’s default value is 1.0s rather than the 0.1s for jiggles. The max_time_between_refs, default 30s, is the maximum time that can be spend observing grid positions before a position switch to the sky reference is require. The rate of going to the sky position is less than for the equivalent JCHOP observation, which can result in worse baselines.<\/p>\n
Similar to JPSSW, there is continuum mode<\/i> for GPSSW, which changes the max_time_between_refs to 5s at the cost of more overhead. Note that position-switch observations are not well suited for measuring the continuum level both, because of their typically larger switch, as well as the relatively long time between sky references even in continuum mode.<\/p>\n
Duration:<\/h4>\n
The above conclusion is illustrated in the plot below, which shows the duration of observations based on the UT timestamps in the headers. The duration scales linearly with the requested integration time for single-point position-switched observations (black squares and line). GPSSW maps with shared offs<\/i> also scale linearly with the requested integration time (blue squares and blue line), but require a bit more time than their 1×1 counterpart due to the telescope motions within the grid. Although the overheads increase for larger maps with more points (cyan points — 4×4, 5×5 pattern, respectively –) the increase is not very large. On the other hand, short maps with separate offs<\/i> (red squares and line) incur an excessive overhead due to the many switches to the sky reference up to the point
\nthe integration time per grid-point exceeds 15s when the distinction between separate<\/i> and shared offs<\/i> vanishes. The durations in the plot include an additional ~60secs before the start UT is written or closing the files after the end UT timestamp. The approximate relations thus become (np: number of points in jiggle; t_on: requested ‘on’-only integration time per jiggle-point):<\/p>\n
\n\n\n\n\n\n\n
Grid-PSSW<\/th>\nDuration (secs)<\/th>\n<\/tr>\n
1×1 PSSW<\/td>\n2.45 * t_on + 80<\/td>\n<\/tr>\n
Shared off<\/i> and separate \u2265 15s<\/i><\/td>\n2.65 * np * t_on + 80<\/td>\n<\/tr>\n
Separate off < 15s<\/i><\/td>\n2.0 * np * t_on + 190<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
 <\/p>\n
\"gpsw_time\"<\/a><\/p>\n
 <\/a><\/p>\n
Noise:<\/h4>\n
The figure below shows the noise in Jiggle-PSSW observations. The top panel compares the observed noise to the expected noise, for PSSW observations ranging from single-position to 5×5 grids, both using shared<\/i> and separate offs<\/i>. As for JCHOP the data indicate an excess noise of about 4% (the black line). The bottom panel shows the normalized noise: rms*sqrt(int. time) as a function of the duration of the observations with the expected noise indicated by the black crosses. The equation used is the same as given in the discussion of JCHOP observations, but ‘np’ is a bit less straightforward:<\/p>\n
    rms = 1.04 * Tsys*1.23\/sqrt(df*t_on) * sqrt(1+1\/sqrt(np))\r\n\r\n     np:  number of points with shared offs<\/i> or 1 if none:\r\n          max(1,int(30\/t_on)) i.e. the number of t_ons that fit \r\n          in 30sec, but no less than 1. (30secs = time_between_refs.)\r\n\r\n<\/pre>\n
Observations with shared offs<\/i> but on<\/i> integration times in excess of 15s are excuted the same as observations with separate offs<\/i>.<\/p>\n
See Jiggle-Chop for more details.<\/em><\/p>\n
\"gpsw_rms\"<\/a><\/p>\n
 <\/a><\/p>\n
Comparing Observing modes<\/h3>\n
The figure below shows the duration all observing modes discussed above and separates into three groups:<\/p>\n