Guidelines for designing EIS studies
v2, 2009 March 23, Peter Young
Raster & Studies
A study is a collection of one or more rasters. When it comes
timeline planning (i.e., when a day's observations are planned), it is
studies that are inserted onto the timeline, not rasters. It is thus
the study definition that is exported to MSSL for testing, not the
A raster can either be a raster scan (i.e., scanning a 2D area
Sun by moving the EIS mirror), or a sit-and-stare (i.e., keeping the
slit fixed on an area of the Sun.
Rasters are much more complicated to design than studies,
are simply a combination of pre-defined rasters. All the information
about spatial area, slit numbers, emission lines, etc., are contained
in the raster.
First of all you must have an idea of what type of observation
to make. Questions you may want to ask are:
- Do I want spectroscopic information, or are images OK? This
determines whether you want to use one of the slots (40", 266") or one
of the narrow slits (1", 2").
- Raster or sit-and-stare?
- Do I want just a few lines for basic temperature
discrimination? or many lines for detailed diagnostics?
Compression & data rate
Early on in the mission the average EIS data rate was around
50 kbit/s. However in 2007 December the Hinode X-band antenna failed,
and so science data had to be transmitted using the S-band antenna
which offers a much lower data rate. For example, during 2009 February
the EIS data rate averaged around 7 kbits/s.
Before 2007 December it was typical for the Hinode instruments
observe for 24 hours/day at fairly steady data rates. The situation is
much more flexible now and a detailed description is given on the EIS
wiki. For example EIS can still run studies at the previous high data
rate, but it will only be for a few hours and no other studies can be
run during an OP period (2-3 days).
For this reason study designers should think in terms of the data
volume their studies produce and compare this with typical data
recorder (DR) allocations given on the Hinode Daily Events webpage.
Previous to the loss of the X-band antenna, most users chose
compression for their studies which gives around a factor 2 lossless
compression. Using JPEG compression is now much more common and general
guidelines are that Q-factors of 92 or higher are reasonable for narrow
slit data, while lower Q-factors can be used for slot data. Compression
factors are discussed further on the EIS wiki.
Trade-offs: window width vs. slit length vs. number of lines
The varieties of rasters are almost endless, and developing a
raster is something of an art form. The ultimate limiting factor is the
EIS data rate and the user must trade off four parameters against each
other in order to keep the data rate down to 50 kbits/s. These are:
- The wavelength window width.
- The number of pixels along the slit.
- The number of wavelength windows.
- Exposure time.
CDS study designers will be experienced in performing this
trade-off, for others a certain amount of trial-and-error will be
One recommendation from CDS experience: you may design a study
observe a loop, or bright point, or some other solar feature, but then
find that you actually catch an interesting flare or transition region
brightening, for example. You will then curse yourself for not
including Fe XXIV, O V, or some other ion. Thus, unless you really want
big spatial coverage or high cadence, I would recommend always trying
to have a good selection of emission lines.
Wavelength window width
For EIS, the size of wavelength windows can be set in blocks
pixels (i.e., 8, 16, 24, etc.), and the windows can have different
sizes. (This contrasts with CDS, where each window had to have the same
size, but you could set the size to any value.)
Before the loss of the X-band antenna, choosing window widths
pixels was recommended. However, users are now strongly constrained by
the data volume of their studies and so it may be preferable to go to
24 or 16 pixels. Bear in mind that high velocity events may be missed
if 16 pixel windows are used.
Given the high instrument sensitivity in the Fe XII 195 line,
should consider having a larger width for this line. Similarly, since
flare lines can show large blueshifts (400 km/s, for example) then
broader windows should be considered for Fe XXIII and Fe XXIV. For
example a 400 km/s blueshifted component for Fe XXIV 192.0 will be 11
pixels from the rest wavelength of the line.
Note that there are regions of the spectrum where you may want
choose a much larger window in order to pick up several useful lines in
a single window. An example is the region around 256.3 to 258.3
angstroms where around 10-12 lines are found.
Choosing your exposure time
There are no hard-and-fast rules due to the wide variation of
sensitivity with wavelength, but the following guidelines can be taken
- For quiet Sun, good line profiles for Fe XII 195 can be
obtained in as little as 10 second exposures. To get good profiles in a
number of different lines requires 50-60 seconds.
- For active regions, 1 second exposures give useful profiles
for Fe XII 195. For other strong lines (e.g., the Fe XIII 203/202
density diagnostic) 5-10 second exposures are fine. Some of the strong
lines are likely to saturate if you go to exposure times beyond 30
seconds, but these exposure times are needed to pick up weak lines such
as the transition region lines.
Remember that if you use the 2" slit, you will get twice as
so exposure times should be adjusted accordingly.
Why use the 2" slit?
As well as allowing you to use shorter exposure times (see
above), the 2" slit also allows you to raster a region more quickly
(since you need half as many steps to do it).
The downsides are that you lose spatial resolution in the
and that the spectral resolution is also degraded.
Choosing emission lines
Please refer to the separate document Choosing Emission Lines
for EIS Studies.
page mantained by Dr.P.R.Young,
last revised on 23-Aug-2013