ISO SWS Frequently Asked Questions (FAQ)
Last update: 12-Jun-95
1) How do I run the SWS Observation Time Estimator (SOTE) at IPAC?
Click here for a usage
summary and template files which you can cut and paste.
2) The TDT's I am getting from the PH software are very high
compared to those output by the SWS Observation Time Estimator (SOTE).
What's the deal?
Yes. Unfortunately, the overheads have increased significantly since Phase 1.
The major new overheads include High Precision Pointing (HPP), extra telemetry
uplink time ("uplink jitter"), and signal-to-noise ratios measured during the
testing phase which for the grating detectors are ~70% to 90% of those assumed
by SOTE. All SWS AOT's now have HPP built in at the level of the AOT-to-OCT
logic. The HPP time overhead amounts to 98 sec. (Note that the AOT-to-OCT
logic no longer imposes an extra 117 sec overhead for each SWS aperture, as it
did during April and early May.)
A summary of the new overheads is as follows:
Item Item Cumulative Total Cumulative
--------- --------------- ----------------
Target acquisition:
Slew: 180 sec
Pointing request: 19 sec
Telemetry: 6 sec
Total: 205 sec 205 sec
HPP:
HPP: 98 sec 98 sec 303 sec
Uplink jitter: 6*(#ICSs) 11 sec minimum TBD sec
3) What is the recipe for computing the new SWS AOT TDTs?
As of 3-Apr-95, the PH is returning TDTs in line with the new realities
described above. You can use the following recipes to predict what the PH will
return for SWS AOT TDTs using the output of SOTE. However, note that as of PGA
v4.0, the PGA itself computes correct OTT's for all SWS AOT's, so this
procedure should no longer be necessary for most observers:
- AOT SWS01
Speed OTT (sec)* TDT%
------- --------- ----
Speed 1 1230 1410
Speed 2 1986 2166
Speed 3 3601 3781
Speed 4 6623 6803
*OTT includes integration time, instumental overheads,
HPP (now only for 1 aperture, not all 3), and new telemetry overheads.
OTT is fixed and depends only on the scan speed is indicated here.
%TDT = OTT + 180 sec for normal target acquisition (assumed here).
= OTT + 20 sec for a concatenated target acquisition.
(Source: Beware 1.8, Douwe Bientema, and PH [after 3-Apr-95])
NOTE: During Phase 1, the SOTE TDTs were 1158, 1878, 3498, and 6498 sec
for Speeds 1, 2, 3 & 4 respectively. The corresponding TDTs returned
by PGA V4.0 and the PH during the period 3 April -- 15 May 1995
were 1606, 2363, 3977, & 6999 sec. The new times in the table above
reflect simply subtracting off 98+98=196 sec from the earlier
implementation which used 2 extra HPP operations that are no longer
required.
- AOT SWS02
[1.50 to 2.00] * [OTT(SOTE)] incorrect S/N in SOTE; telemetry pads;
HPP (1 aperture).
+ 180 acquisition; 20 sec if concatenated.
-----------------------------
TDT : _______
- AOT SWS06
[1.15 to 1.25] * [OTT(SOTE)] incorrect S/N in SOTE; telemetry pads
HPP (1 aperture).
+ 180 acquisition; 20 sec if concatenated.
-----------------------------
TDT : _______
- AOT SWS07
Experience has shown that a recipe such as those above does not
predict the TDT returned from the PH software. Instead, an SWS07
TDT can be best estimated by simply doubling the TDT (not OTT)
computed by SOTE:
TDT ~ 2 * TDT(SOTE).
4) For AOT SWS01, should the flux I input into the PGA be the source flux,
or the total (source plus background) flux?
This parameter is used to set
the detector gains, to avoid saturation and to assure optimal signal-to-noise
will be achieved. Thus, input the total (line, plus source continuum, plus
Galactic background (FIR cirrus), plus Zodical emission (if significant) flux.
The Galactic and Zodi emission can be easily estimated using ISKY or IBIS.
5) Do I need a background/reference observation?
At SWS wavelengths,
unless your target is sitting along the line of sight to a Galactic cloud with
significant dust emission, the background will be dominated by the zodiacal
light. The typical zodical emission in an SWS aperture is ~0.3 Jy. Of course it
depends in detail on the exact location of your target, and on the direction of
viewing from the satellite at the time of observation (the solar elongation,
etc.). To be safe you should use IRSKY (or IBIS for the zodiacal backgrounds)
to measure the expected background flux density in the SWS aperture(s) you are
using. If the background is a significant fraction of the expected flux density
in your target, and if you want SPECTROPHOTOMETRY (absolute calibration of the
continuum and lines) and/or accurate line EQUIVALENT WIDTHS, you should take an
offset/background/reference integration and subtract it from the target
spectrum. On the other hand, if all you're interested in is LINE FLUXES AND
LINE RATIOs, and the noise in the background is not too high, you can probably
get away without a background measurement. This is because the continuum around
the line(s) of interest will also be sampled by the spectrograph, so regardless
of the exact flux and shape of the infrared background which is added to your
source's underlying continuum, you can fit the continuum surrounding the
line(s) and subtract it, leaving only the emission or absorption lines. A
detailed analysis requires a proper propagation of errors (noise) in each
specific case. BE CAREFUL!
6) Will there be a new version of SOTE?
NO! The SWS Instrument
Dedicated Team has abandoned SOTE for computation of FINAL TDTs. Version 6.0
(20 June 1994) was the last version of SOTE. It should be regarded as a tool
to approximate the On Target Time (OTT), and to do "what if" experiments to
figure out what S/N you will need to ask for to trigger the next step up (or
down) in integration time, since this is quantized by the detector reset
intervals of 1, 2, 4, or 8 sec. The recipes explained in item (3) above can be
used to estimate the new overheads and TDTs, based on SOTE's computed OTTs (and
TDT for SWS07). However, as of mid April 1995, the Proposal Handler (PH)
software returns real TDTs by implimenting the new AOT-to-OCT logic for SWS.
Likewise, PGA Version 4.0, released 24-April-95, now contains the current
AOT-to-OCT logic and computes OTT correctly for all SWS AOTs.
7) What is the minimum (maximum) line-to-continuum (continuum-to-line)
ratio that I can expect the SWS to be able to measure?
Tests have shown that line-to-continuum ratios as small as ~1% (3-sigma?)
can be measured with the SWS grating. See also question (8).
8) My source will have weak unresolved lines on a strong continuum.
To detect the line at a given signal-to-noise ratio, what signal-to-noise
ratio should I input into the SWS PGA (or SOTE) software for the total
(line + continuum) flux density?
S/N(total)=S/N(line)*{1+[F_nu(continuum)*d_lambda/F(line)]},
where F_lam(continuum) is the expected flux density [W m^-2 um^-1] in
the continuum, d_lambda [um] is the line width (use one resolution
element), and F(line) is the expected total line flux [W m^-2] in the
resolution element containing the line. For example, if you want
S/N=10 in a line with total flux 6e-17 [W/m^2], and the continuum is
150 Jy (1.8e-14 W/m^2/um], the S/N required on the total flus is 1810!
See Sections 6.4 and 6.5 of the LWS Observers Manual for a
good desciption of line-to-continuum and S/N issues.
9) Where can I find accurate wavelengths for common
astrophyical emission lines observable with the SWS?
Here is a good start---
10) What is the read-out-noise of the integrating amplifiers for the SWS
detectors, and how does it limit the sensitivity of my observations?
The minimum system noise (read-noise) for the grating detectors, measured at a
signal power at which the photon noise is negligible, ranges 30--45 e-, or
0.32--1.00 uV/s at gain=16 and reset time = 2 sec (Table 5.1, p. 27, SWS
Observer's Manual). The expected S/N ratio per resolution element for the SWS
is expressed as:
S * t_r * sqrt(n_int)
S/N = ----------------------------------------,
sqrt{N_r.o.n.^2 + (N_d^2 + N_s^2*S)*t_r}
where S is the signal flux [uV/s] (use the curves in Figs. 5-3 - 5.8), t_r is the
detector reset time in [s] and n_int is the number of integrations per
resolution element (see Table 6.3, p. 50), and N_r.o.n., N_d, and N_s are the
read-out noise [uV/s], dark noise [uV/s], and signal shot noise [sqrt(uV/s)]
(see Table 5.2, p. 27).
An observation is said to be "read-noise limited" when N_r.o.n. (sigma_r.o.n.)
is greater than the other noise sources. (An observation is said to be
"background limited" when sigma_b.g. >> sigma_r.o.n.) For example, the Si:As
detector (12.0-29.5 um) has N_r.o.n.=1 [uV/s], N_d=0.41[uV/s] and N_s=0.14
[(uV/s)^0.5]. A read-noise limited observation for this detector would require
N_r.o.n. >> (N_d^2 + N_s^2*S)*t_r, or
1 >> (0.17 + 0.02*S)*t_r.
As an example, for an observation of a source with S_nu=0.5 Jy at 18.0 um,
S=3.0[uV/s/Jy]*0.5[Jy]=1.5[uV/s], so we inquire whether 1 >> 0.2*t_r.
Using t_r=1 or t_r=2 (e.g., the only possibilities for SWS01), we find
that indeed the observation is read-noise limited. However, using SWS02 or
SWS06 with t_r=8 (the maximum), the observation is not read-noise limited.
The above recipe can of course be used to get an idea of the limitations of
your SWS observations at other wavelengths and flux levels. The key data are in
Table 5.2 (p. 27) [or Table 5.3, p. 39 for the FP] and Table 6.3 (p. 50).
11) What is the expected precision of the SWS calibrations?
- Flux Calibration:
The final flux calibration will depend on how well the memory (hystersis)
effects of the detectors can be controlled and modeled under in-orbit
conditions. The goal for the ABSOLUTE calibration of grating and FP spectra is
30%. The goal for the RELATIVE calibration of grating and FP spectra is 10%.
- Wavelength Calibration:
The wavelength calibration will be to 1/5 of a resolution element, or ~75 km/s,
for the grating, and to 1/3 of a resolution element, or ~3 km/s for the FP.
- Instrumental Profile:
Characterization of the FWHM of the grating and FP instrumental profile is
expected to be accurate to ~10%. It will be limited by various effects,
including memory effects, mechanical stability, pointing sability, etc. The
instrumental profile is not extacted to be known well enough to perform
deconvolution.
References
ISO SWS Observer's Manual
ISO LWS Observer's Manual
ISO PGA Reference Manual
Various memos and communications from ESTEC and the SWS Instrument Dedicated Team.
Disclaimer
These are personal notes of J. Mazzarella which
are in no way complete or intended to be a substitute for the original
SWS document set. They are made available in case they may be
of use to others.
ISO SWS Information (Back)