Now let's look at spectra. Back on the main GUI window, use the pull-down menu on the top to select:

and a new window labeled SPECTRA should open up looking something like the following:

Note that both the Observation Time Interval and the Spectrum Time Interval shown in this window are the same as previously selected. We must select the time intervals within this overall time for spectral analysis. The flare of interest is finished by about 11:20 UT so we will change the spectrum time interval to end at that time. Click on the "Change..." button and set the interval duration to 1200 s and return.

We now want to divide the flare up into useful time intervals. We must also be sure to select some intervals, both before and after the flare if possible, that can be used for background determination. The default is to divide the Spectrum Time Interval into equal one minute intervals. The default time bins can be changed by pressing the "Define Bins Manually" button in the third box down from the top of the window where the current Time Bins are listed in the pull-down box. For now we will accept the default of 20 intervals each 60 s long.

Next we must decide what energy bins to use for our spectra. This is done using the second box from the top that initially lists the default energy bins for Binning Code 7. Generally you want to pick energy bins fine enough to distinguish between the different emission models you want to study, but not so fine that the statistical noise is overwhelming. Binning Code 7 gives bins that are a bit coarse for detailed continuum spectral work (such as separating out thermal, superhot thermal, and power-law components, for instance). There are several other good default choices, or you can construct your own binning by pressing the "Define Bins Manually..." button. You can see what default energy bins are available by clicking on the "Show Binning Codes" button. We will use one of the other default sets by selecting Binning Code = 1. To do this, click on the pull-down menu next to the words "Binning Code" and select "1." Binning Code 1 gives us 1-keV bins from 3 keV up to 100 keV.

Keep in mind that the total number of bytes of memory Nb that you will need to store all of these spectra depends on the number of time bins Nt and the number of energy bins Ne that you specify according to the following relation:

In our case, Nt = 20 and Ne = 97, so we will only use up 838 kbytes. For more detailed analysis, you will want to choose finer time bins as short as half a rotation or about 2-s long but you must be careful not to choose too many time intervals and energy bins so that Nb exceeds the available memory space.

The next panel lets us select detector segments: front and rear for any or all of the 9 detectors. The document Behind, Beneath and Before HESSI Spectroscopy or "BBB" discusses the detector segments in detail. Basically, the front (F) segments detect most of the hard x-rays up to 100 keV and the rear (R) segments detect most of the gamma-rays above 100 keV. The flare we are analyzing is only detectable up to 100 keV so we choose just the front segments. This is the default setting.

We should also do something that you'll probably want to do most of the time in real life: remove from consideration the front segment of the detectors under Grids 2 and 7. (We refer to these detectors as "G2" and "G7" for short). Detector G2 after the first couple of weeks in orbit could only be used in the unsegmented mode and has a resolution of about 10 keV and a threshold energy of about 25 keV. Detector G7 has an energy resolution in its front segment of about 3 keV instead of the ~1 keV FWHM at hard x-ray energies that the other detectors have. To deselect these two detectors, click on the "Change" button in the fourth box down in the Spectra window that shows the selected detectors and segments. A new window appears that allows you to select or deselect any segment. Click on the checked boxes next to 2F and 7F to deselect them. Note that all the other front segments are selected and all the rear segments are deselected. Click on the "Accept" button to return to the Spectra window.

You should also be aware that Detector 8 also has problems when the transmitter is on and is using the aft antenna. There is some noise pickup that results in a higher background at these times in this detector channel. You can check for this by plotting the light curve for this detector alone and comparing it with the light curves for the other detectors. We don't have to worry about it for this flare.

Their are two more settings that can be changed before generating a plot. First, to look at the spectrum in units of counts flux, select that option in the "Spectra" window, on the same row as "Semi-Calibrated", in the "Units" section. The second option is to change the flags that are displayed in the Time History. This is done by selecting the "Change" button on "Show Flags on Time History", and this window will pop up:

Select "All", and then press "Accept" to close this window. Add notes on usefulness of flags?

At this stage the SPECTRA window should look as follows:

We can see the time history of the flare we are analyzing by pressing the "Plot Time History" button at the bottom of the SPECTRA window. Note that since this is the first time we are actually extracting the information about individual counts from the flight data files, it takes a little while for this step to execute - about 2 minutes on my 800-MHz laptop. The actual time taken depends on the number of counts recorded in the interval of interest. So depending on the speed of your computer you must be patient. It's a good idea to have the main IDL window visible before you click on this button so that you can see any error messages in case the program crashes. You can generally tell if everything is going OK by noting occasional disk activity while you are waiting. The time history for this flare will appear in the main GUI window and look like the following:

This is the counting rate time history summed over all the energy bins we selected. You can view the light curve for any individual energy bin or any group of bins, either separately or summed by clicking on the "Plot_Control" pull-down menu and select "XY-Plot Display Options." In the resulting window, you can also change the appearance of the plot such as switch to a linear scale or change the maximum and minimum values of the Y-axis.

Once you are satisfied with the appearance of the time history plot, close the XY-Plot Options window.

Now let's look at the spectrum we're getting by pressing the button on the bottom of the SPECTRA window labeled "Plot Spectrum." The result will appear in the main GUI window and should look like this.

This is the count spectrum, meaning it's the spectrum of the detector counting rates vs. the energy that the detectors recorded for each count. It is summed over all of the time intervals that we selected. Again, you can plot any one or a selected group of intervals by clicking on the "Plot_Control" pull-down menu and select "XY-Plot Display Options" as before.

First, notice that the spectrum is rising from 3 to about 12 keV. This is not a characteristic of the Sun; it's due to absorption in passive material in front of the detectors (thermal blankets, beryllium detector windows, and most importantly the aluminum attenuators, the thinner of which was in place for this flare). From 12 keV to 100 keV the spectrum falls roughly as the true solar spectrum does. The background spectrum contributes a greater fraction of the total at the higher energies. Also, the sharp peaks that can be seen in the spectrum at energies above 50 keV are lines from the germanium detectors and are not of solar origin.

The rest of this document will be devoted to converting this count spectrum to a photon spectrum, i.e. what's incident on the spacecraft from the Sun. The simplest thing to do is to divide this spectrum by the efficiency of the system as a function of energy. This will remove the effects of absorption from the materials in front. To do this, simply click on the check box next to "Use Semi-calibrated Data" and then click on the Plot Spectrum button again. You should now see something like this:

The instrumental absorption which caused the spectrum to roll over below 12 keV is gone. Unfortunately, the background is still not subtracted and the flattening at higher energies above ~30 keV is not real. Later versions of this program will allow the background to be computed and removed before the photon spectrum is computed. The slight bump at 6 - 7 keV is probably the real iron-line feature from the flare.

"Semicalibrated" photon spectra of the type shown above can be used with caution for science when all of the following conditions hold:

1. Only the front segments are used.
2. The flare spectrum is at least ten times the background spectrum for all the energies being examined.
3. Only energies below about 75 keV (50 keV to be conservative) are being examined. Above 75 keV, the redistribution of photons to lower energies by Compton scattering becomes more and more significant.
4. Fluxes at energies below 10 keV should only be considered accurate if both attenuators are out. With the attenuators in, the 3 - 10 keV band can be dominated be germanium K-shell escape events from the 13 - 20 keV band.

To determine the incident photon spectrum more accurately, we must use the full Spectral Response Matrix (SRM). (Note that this is called the Detector Response Matrix or DRM in SPEX). Each line of the SRM gives the total response of the instrument to photons of a particular energy. The diagonal terms of this matrix represent the total system efficiency for recognizing a photon energy directly. Off-axis elements represent the probability that photons are redistributed to another (generally lower) energy. For hard x-rays, more photons are recorded with their true energy than with a lower energy, but above 100 keV this is reversed. This is why the "semicalibrated" procedure, which simply takes the diagonal elements of the response and divides the count spectrum by them to get a photon spectrum, is inaccurate above 50-100 keV or so, even without considering the issue of background. Note also, that using only the diagonal elements gives inaccurate results below 10 keV if either shutter is in the FOV because of the germanium K-shell escape events.

To analyze count flux data in SPEX, a count rate spectrum and a response matrix must be read into the active session. The response matrix is stored in the array drm within the SPEX common blocks. When using the HESSI spectrum object or GUI to create the count rate spectrum file and response matrix file ( filewrite on the CLI and write output from the GUI), the name of the response matrix file is written in the RESPFILE parameter in the header of the spectrum FITS file. A different drm is needed to analyze the count rate obtained with each attenuator state.

The way we are using the HESSI spectrum object right now, you only need one srmfile for each attenuator state for each detector combination for every binning, specify that with dfile, and then use read_drm to create the drm array used with SPEX. There is no distinction by flare at this time even though we autoname the srmfile with the flare date and time, 11/15/02. In the future there will be some information about the average position of the source from the imaging axis. You can even write all of the spectrum data for a flare into one file and then change the drm as needed. I recommend that you adopt a standard binning code and a standard set of detectors and create srm files for all three of the used attenuator states. Then you should name the srm files to something like hsi_srm_generic_bincodeX_detZZZZ_attN.fits. X would be the binning code (one of those from the GUI or your own system), ZZZZ would indicate your choice of segments, and N would indicate the attenuator state. You can produce these srmfiles from the spectrum object (sp_obj->filewrite,/fits,/build,... ) or from the SPECTRUM GUI by choosing a start time when the instrument is in the needed attenuator state.

This page last updated: June 27, 2011