ASCA Observation Planning Guide
AO-7 version, 1998 June 1
Important UpdateIt is predicted that, by the end of AO-5, 1-CCD Faint mode and 2-CCD Bright mode of SIS will require the use of level discrimination set at about 0.48 keV, due to the continued increase in the flickering pixel rate. See below for details.
IntroductionThis document suggests guidelines to ASCA PIs in planning an observation, after it has been accepted (although it might also be useful in writing a proposal, NASA Research Announcement contains more extensive documentation for this).
Both ISAS and the ASCA GOF at NASA/Goddard Space Flight Center have a system of rotating duty scientists to help PIs plan the details of their observations (the former is for Japanese and European PIs, while the latter is for US PIs). This document is intended as explanatory; final decisions should be made after consultation between the PIs and the duty scientists. Since ISAS uses a semi-automatic system to generate 10 days of satellite commands, which must be done at least 3 days ahead of the first observation, the deadline for the detail planning should be considered as 2 weeks ahead of the observation date.
To maintain a sufficient level of power supply, the Solar Paddle must
be pointed close to the Sun. There is a hard operational limit of
25 deg, meaning that ASCA can observe targets 65-115 deg away from the Sun.
The preferred limit is 75-112, to increase the margin for safety;
observations outside this range are discouraged, and sometimes impossible
due to other operational constraints. The time of the
year during which a target is observable with ASCA can be estimated using the
provided by the ASCA GOF.
There is an additional constraint that the two star trackers must not point to within 20 deg of the Moon (this limit may be conservative; however, ghost images can again cause the Safe Hold mode). This can usually be avoided by adjusting the roll angle, but such an option may not be always available if there are other constraints (e.g., time-critical observation).
Most non-time-critical observations are in fact scheduled to have a Sun angle in the range 80-110 deg. This is because the quality of SIS data degrades due to scattered sunlight at lower Sun angles. However, consecutive observations at large Sun angles tend to warm the CCDs, which also has serious detrimental effects to the data quality.
Because of the square shape of the CCD chips, particular roll angles may be required to fit an extended source onto the SIS field of view. Note that such roll-critical observations are time-critical (also see Source Placement below.
Many PIs are interested in coordinated observations with ground-based telescopes. This can be achieved, to a degree, without making the ASCA observation time-critical. To do so, wait for the long-term timeline to become available (which may, however, be too late for observations at the beginning of a new AO period). Schedule your ground-based observation for the ``week'' in which your ASCA observation is scheduled. (A ``week'' in this context does not necessarily correspond to 7 days, but this term is used for convenience. Long-term timeline does not specify what time during the week a particular observation will be made.) Notify Dr. Nagase at ISAS (or the duty scientist assigned to you), and hope the long-term timeline does not have to change. If the circumstances allow, we will attempt to arrange the short-term timeline to make your ASCA and ground-based observation as close in time as possible; however, this is on a best-effort basis and we do not guarantee successful coordination unless the proposal was marked time-critical at the time of submission.
Use default position for point source.
In the case of observations of a single point source, the primary concern is to place the target so as to collect as many photons as possible. We need to consider several different factors in this:
- Telescope vignetting (optical axes, the detector positions with maximum effective area, differ slightly from instrument to instrument).
- The size of CCD chips, the gaps between them, and different characteristics of the individual chips --- we strongly recommend to collect all source photons onto a single chip per SIS.
- The grid in GIS window, which lowers the quantum efficiency.
In all the above, the uncertainty in the locking attitude must be taken into account. Currently it appears that the locking of spacecraft on the target coordinates is accurate to 0.01 deg.
The preferred chip is SIS0 Chip 1/SIS1 Chip 3; they are both well-calibrated (relative to the other chips) and suffer comparatively little from hot pixels and light leaks. We have developed a set of default positions, which are shown in an accompanying diagram together with the optical axis positions and other information regarding the size and orientation of the CCDs. We believe the "1-CCD default" to be the best position to observe a single point source in most circumstances.
Rely on ISAS expertise for extended sources.
For observations of an extended source or multiple sources, a particular roll angle may be advantageous. Normally, however, the actual roll angle is determined after the scheduling has been done and very close to the actual observation. It is therefore very important to understand that a specific roll angle may or may not be available.
It is best to let the operations team at ISAS decide on the actual roll angle, after providing them with positions of all the features/sources of interest (with priorities) or an X-ray image of the region (preferably annotated to show important features), on a best effort basis. The duty scientist (either at ISAS or at GSFC) should act as the intermediary to reduce confusion. Note that roll angle for any given date is severely limited by the Sun-angle constraint so if any observation is truly roll-critical, this has to be checked well in advance of the timeline being made.
Use PH mode for GIS
The spectral response of GIS in MPC mode is highly uncertain because it is impossible to perform position-dependent gain correction on individual photons. Except when timing analysis of an extremely bright (> 100 cps) source is sought, PH mode is always advantageous over MPC mode.
Use default bit distribution
Each PH mode event occupies 32 bits in the telemetry; the default configuration uses 1 bit for sensor ID, 10 for PH, 8 each for X and Y position etc. (click here for more details). In principle, the on-board processor can be commanded to use different number of bits for different attributes of the events, and the downstream software should be able to accommodate such non-standard configurations.
However, we recommend the use of the default bit distribution, unless there is a compelling reason to do otherwise. This would simplify the operation (ASCA operation is very labor-intensive; introducing additional complications can possibly lead to errors).
The only reason for not using the default bit distribution has been to increase the time resolution, from the default values of 500 msec at medium bit rate and 62.5 msec at high bit rate. These can be improved by a factor of up to 1024 by changing the bit allocation. Even then, it is probably wise not to use arbitrary combination but use an oft-used alternative. Consult the duty scientist for details.
Stay within telemetry limits
After launch, the ASCA team unexpectedly discovered the presence of hot and flickering pixels --- those that give out spurious signals that the on-board software mistakes for photon events, some permanently active (hot), some that come and go (flickering). Analysis software can eliminate the effects of these successfully (except in fast mode); however, they occupy valuable telemetry slots. Moreover, the flickering pixel rate is increasing, which is a manifestation of the cumulative radiation damage to the CCDs.
Limits for source count rate (counts/second per CCD chip) are:
A "-" indicates that telemetry saturation will occur even when source counts are negligible. The above limits reflect the increased event threshold being used since the beginning of AO-4 period.
Estimated count rate limits at the beginning of 1997
Data mode Faint Bright Fast Clocking mode 2 1 2 1 (1) Bit rate High 15 40 105 225 477 Medium - ~1 ~2 17 41
Some CCD chips are affected more by the hot and flickering pixels. For 1-CCD mode, we have assumed the use of the default chips (SIS-0 Chip 1 and SIS-1 Chip 3), while SIS-0 Chip 0 should be avoided in Bright 2-CCD mode observations.
The rate of flickering pixels is slowly increasing. It is predicted that, by the end of the AO-5 period, 1-CCD Faint mode and 2-CCD Bright mode may become telemetry saturated during times of medium bit rate. To avoid this, the use of level discrimination (see below; probably at 0.48 keV) will be necessary.The ISAS duty scientists will keep track of this important issue, and will advise the PIs and Goddard duty scientists accordingly.
Be prepared to use level discrimination
Most flickering pixels have low PHA values (below 0.5 keV), barely above the event threshold. At these energies, the SIS does not have a large effective area, and the responses are calibrated with limited accuracy. Therefore, the use of level discrimination is highly effective in removing flickering pixel events from telemetry without, in many cases, unduly reducing the scientific return of the observation.
As noted above, the use of level discrimination may become routine for 1-CCD Faint mode and 2-CCD Bright mode at medium bit rate, by the end of the AO-5 period.
In addition, if the CCDs become warm (> -60C, > -59C in particular), the flickering pixel rate increases; in such a case, the duty scientist at ISAS or at KSC may use level discrimination to avoid telemetry saturation.
If your target is highly absorbed, you can raise the level discriminator setting to higher energies (2 keV, for example, in extreme cases) to avoid most of the flickering pixels.
Use faint mode as much as possible
We have discovered two effects in the on-board processing that affect the quality of SIS spectral data: echo effect and dark frame error (DFE). Both effects can, in principle, be corrected on a photon-by-photon basis when the faint mode is used, although the increasing number of active pixels is making it difficult to correct for the latter (the RDD effect). For best spectroscopic results on bright sources, the use of faint mode appears very advantageous at this point, if this is possible within the telemetry limit.
Spectroscopic calibration of fast mode is still very uncertain, for two reasons: The XRT calibration for a square region (which fast mode data require) is uncertain; the secular gain change is dependent on the position of the photons on the chip, which is not recorded in fast mode. Therefore, fast mode data are, and likely to remain to be, less accurately calibrated than imaging mode data.
2-CCD mode can increase the accuracy of background estimates
Background subtraction for SIS data can be done locally or by using background fields (which the ASCA GOF has made available). If the Galactic background is particularly important (objects on the Galactic plane, for example), then local background subtraction is preferred. Unfortunately, 1-CCD mode observations leave little room for background regions, even for a single point source (although not impossible); in such a case, the use of 2-CCD mode may be advantageous.
Note, however, this usually involves a trade of better background subtraction for somewhat poorer calibration as well as reduced efficiency (see the extended source section below).
Consult duty scientist for very bright sources
We are currently exploring different options for very bright sources (> 50 cps), for which telemetry saturation and photon pile-up can be significant. Possibilities are the use of fast mode, using only the high bit rate data, and the use of area mask to reduce count rate, or to analyze data with photon pile-ups. None of the above appears to be a perfect solution. Consult your duty scientist for the least of the evils.
Extended sources --- avoid 4-CCD mode
4-CCD mode data suffer strongly from the cumulative effect of radiation damage, while 1-CCD mode data remain relatively unaffected. Therefore, the ASCA GOF advise against any use of 4-CCD mode. For extended sources, 2-CCD mode with complimentary chips may still be viable. Although 2-CCD mode is also affected by the RDD effects, for the AO-5 period, it is predicted to produce usable data when telemetry saturation can be avoided..
Please consult your duty scientist for the latest information.
If you have any questions concerning ASCA, visit our Feedback form.