Space Environment Center
GOES
Solar X-ray Imager
|
||||
| GOES
Solar X-ray Imager |
||||
This web page and its links contain an overview of the GOES-N (13) SXI instrument and its operations. These links include:
Further information on the engineering aspects of the instrument can be found in the GOES-N DataBook section on SXI.
To help forecast geospace events of solar origin, the following SXI requirements have been set:
Figure 1: The GOES-13 SXI Accomodation
The SXI instrument is housed in the yoke assembly of the GOES-N spacecraft, as depicted in Figure 1. The main body of the spacecraft points continuously at Earth, but the yoke assembly on which the solar array panels are mounted counter-rotates to face the Sun at all times. The east-west rotation of the yoke and solar array can be adjusted as described below, and the SXI can be pointed north-south over a limited range independently.
The key features of the spacecraft accomodation include:
The SXI instrument consists of a telescope assembly (Figure 2) and its
associated electronics boxes. The telescope itself can be thought
of in three basic physical segments: The objective/mirror section, the metering
assembly, and the focal plane/camera section.
Figure 2: Telescope Schematic
Broadband soft X-rays in the spectral region SXI nominally observes, 0.6-6.0 nm, cannot be focused by transmissive optics or by "normal-incidence" mirrors. It is necessary to use grazing incidence mirrors, typically shaped like a cylinder, with the X-rays making one or more bounces at very shallow angles. GOES-13 SXI uses a modified Wolter design with a single piece of glass fashioned into two sections. In a break with past designs, both sections have a hyperboloidal figure, which provides a flatter overall focus. Like all such optical systems, the GOES-N SXI design suffers from a geometric vignetting that increases linearly with displacement from image center. The magnitude of the vignetting approaches 13% for features lying 20 arcmin off-axis.
Figure 3: Detector Summary
The GOES-13 SXI detector is a back-illuminated CCD with very high quantum efficiency in the applicable wavelength ranges. Its basic properties are summarized in Figure 3. The CCD has a 528x588 array of 16-micron square pixels, corresponding to a spatial resolution of 5 arcsec on the solar disk. The long dimension is oriented in the apparent direction of E/W solar image drift. For standard patrol imaging, a 512x512 pixel area centered on the Sun is extracted for downlink.
This detector has a number of important features and artifacts.
The CCD has two gain settings, "High" and "Low", which control the amplification of incoming photons as they are converted into electrons. The "High" gain setting results in more electrons being collected for a given photon, and it is used primarily for test purposes. The standard operational setting is "Low", which provides a greater dynamic range. At the nominal operating temperature of -40C, the mean dynamic range is of the order of a few hundred at "Low" gain, and about a factor of 5 less at "High" gain.
The CCD can be actively heated and cooled to maintain a suitable operating environment. The CCD will be periodically baked to restore efficiency lost to routine on-orbit degradation.
The SXI CCD has a mean charge capture (MCC) of 65% (as measured at 13.3A), so it introduces only minimal spreading in point source images. The full-system measured on-axis RMS point source size is about 5 arcsec, which is the pixel size.
Yohkoh SXT and GOES-12 SXI images suffer from prominent "blooming" during bright flares. That is, some of the electrons in an overloaded pixel bleed into the neighboring pixels, creating uni- or bi-directional spike-like artifacts. The SXI-N CCD is equipped with electronic drains that should eliminate this behavior in most instances.
Compared with the GOES-12 SXI detector, the SXI-N CCD features a very low noise level and a relatively uniform, flat background. Periodic monitoring of the dark background, including bad pixels, can be performed using a special blue LED source.
The data rate for the SXI allows for the transfer of one 512x512 pixel standard image in about 40 seconds. This allows about 20 seconds for other telemetry, i.e., SXI actually has a downlink capability of 3 full images in 2 minutes. It also has a data storage buffer capable of holding about 4 full images, which is very useful in managing the downlink of multiple images taken during an observing minute.
The electronics are capable of reading all or select parts of the CCD image, and and there is provision for further processing aimed at reducing image size for faster downlink. There are two amps for redundancy, and their relative gains as a function of temperature are well documented.
The "permanent" memory for SXI sequence and control is contained in EEProm. This memory retains the image command sets through a temporary shutdown, so that a full upload need not be done to get the instrument back into operation. A basic command set is loaded before launch, but the EEProm can be updated as the instrument settles into operational mode.
The mirror assembly and metering tube are equipped with heaters to moderate seasonal variations in the thermal environment.
Were the SAD to be turned off, the orbital motion of the GOES spacecraft would draw the solar disk across the CCD face at a rate of about 3 pixels (15 arcsec) per second. When the SAD is put into motion to track the Sun, whatever its mode (continuous single step, continuous stutter step, or variable stepping with quiet period) the solar disk still moves about the CCD face. These motions include any uneveness in the SAD drive rate as well as the jitter it induces in the spacecraft bus by its very movement. These residual motions contribute to image smear, and there are two options for minimizing their effects. One is active on-chip image compensation using the high-cadence HASS determinations of instrument pointing relative to solar disk center. The charge on the CCD is shifted very rapidly in the E/W direction to track the motion measured by the HASS. There is no N/S correction. The other option is to impose a constant drift correction, in which the charge on the CCD is shifted to pace the mean solar drift motion. Images can be taken in any case with no compensation, but these would be for test purposes.
The overall performance of the GOES-13 SXI is similar in spatial resolution to the Yohkoh SXT telescope in full-image mode. Its spectral sensitivity bridges the gap between the Yohkoh SXT and the GOES-12 SXI. Optical and spectral performance is covered in detail in the Instrument Performance Summary. Finally, the GOES-N SXI offers even higher sustained cadence than the GOES-12 SXI, as well as much greater flexibility in the mix of images.
Individual images are characterized by six primary parameters. The first three (integration time, filter, and intent category) are rather straightforward and are very relevant to image interpretation. The last three (extraction, binning, and compression) can impact the image information content, but are often not as critical as the others. Details on the relevant keywords in FITS products and annotations in PNG products can be found in Data Products and Software .
The integration time is the duration of the integration period for the image taken. The detector is mechanically shuttered. The exposure duration can be set from 1 ms to approximately 60 seconds. Images can be synchronized with the SAD quiet period (if enabled).
In addition to a fixed polyimide entrance filter (intended to strongly suppress the longer-wave UV and visible solar spectrum), the GOES-13 SXI has analysis filters mounted in two wheels just in front of the CCD detector. These help discriminate between solar sources of emission with different spectral signatures. Each filter has a bandpass designed to be sensitive to a certain temperature range. There are four basic types of filters: the polyimide filters, the tin filters (being flown for the first time), an aluminum filter, and the beryllium filters. There are three polyimide filters and three beryllium filters: in both cases there is a primary and backup thin filter (one in each wheel, for redundancy) and a single thick one. There is a primary and backup tin filter, and just one aluminum filter. Finally, there are also "Open" (no filter at all), and "Glass" filters, which are used for instrument test purposes.
The beryllium filters are sensitive to the hottest temperatures. The polyimide filters include all the flux in the beryllium filters, but add cooler, longer wavelengths. The tin filter is similar to the thin polyimide, but offers enhanced response to cool coronal material and a flatter overall profile of transmission as a function of wavelength; it comes closest to the GOES-12 SXI "OPEN" position. The aluminum filter is sensitive to hot active region and flare material and is included to complement the spectral response of the beryllium filters. The "Open" filter position provides images dominated by long-wave UV and visible emissions, and the "Glass" filter blocks X-rays while passing longer wave radiation.
For additional flexibility of exposure, the basic filters can be ganged. That is, running the two beryllium thins in tandem produces a beryllium "medium", and similarly for the polyimides and tin filters. While it is possible to mix filter types (e.g., combine a polyimide with a beryllium) this capability is unlikely to be used in practice, unless pinhole defects or multiple filter failures occur.
Note that PNG images produced by the GOES-N ground system are color-coded by filter type: green is for polyimide, yellow is for tin, blue is for aluminum or beryllium, and grey is for calibration images, such as darks.
To simplify the selection of images, 'intent categories' have been assigned to different image types. These intent categories include:
The basic concept is that a range of exposure times and filters provide images that best represent certain types of solar features. Thus, a 1 sec image using the thin Polyimide filter is categorized as 'Coronal Structure'. It provides high sensitivity and good dynamic range for seeing fainter features. Alternatively, a 100 ms thin beryllium image is attuned to typical active region emissions, and a 10 ms thick beryllium image is more suited to the hot, bright plasma of intense flares. The label "Test" is applied to all other imagery, such as backgrounds and other calibrations.
Portions of an SXI image can be selectively downlinked to save transmission time and enable more images to be taken in any observing minute. ("Extracted" images are also called "partial" or "windowed" images.)
SXI images can be binned with a choice of 2x2, 4x4, or 8x8 on-chip pixel averaging to reduce downlinked image size.
SXI images can also be compressed. The image can be downlinked in 8-bit format, and the conversion from the normal 12-bit format is effected by a choice of two lookup tables. By default, one imposes a square-root compression, while the other simply retains the high 8 bits of the full 12-bit word. These tables can be modified by ground command.
Extraction, binning, and compression may be applied in any combination to SXI images.
Sequences of images are controlled by a set of 15 tables stored in the SXI computer (click for schematic). The highest level table is the Sequence Control Table (SCT). This table contains 16 distinct sequences of images, only one of which is active during any given observing minute. Each sequence is constructed as a set of pointers, with the SXI system keeping track internally of all necessary counters and logical loops. Each entry in a table contains a pointer to an entry in a lower level table. A key table is the Frame Definition Block (FDB), which contains coded settings for the image parameters discussed above in section 3.
Although the SCT contains 16 sequences, many more can be stored in the SXI memory and can be swapped into the SCT by ground command as the need dictates. Of the resident 16 sequences, about half are reserved for use in special spacecraft events (such as eclipses, station keeping, and thruster flushes), while the rest are available for routine observing. Normally, one such sequence is designed for patrol during times of low (non-flaring) solar activity, one is tailored for use during significant flaring, and another for great (e.g. 'X'-class) flare events. The remainder of the "free" slots normally contain alternate versions of these base sequences.
The basic sequence flow is illustrated with a simple example to be run shortly after launch. This particular sequence (click for chart) is based upon patrol sequence logic developed for the GOES-12 SXI. For simplicity, only one picture is taken per minute, as for GOES-12. The guiding philosophy for patrol sequence design is:
In the example, the "Patrol" block loops through the same 8 image set for 112 minutes. This set includes a thin polyimide pair at 4 minute cadence, and a thick polyimide and a thin beryllium pair at 8 minute cadence. These cover the CS, AR, and FL intent categories, respectively. Four times a day this set is interrupted for 8 minutes to execute a special differential emission measure (DEM) set, and eight times a day it switches to a "Test" set, used here to evaluate the relative merits of alternate exposures.
Later GOES-13 sequences will take advantage of the abililty to routinely downlink multiple images per minute. The most obvious application is to augment the regular patrol sequence by using the 20 sec per minute of downlink left free in the previous example to take additional flare-related images, say, using the aluminum and thick beryllium filters. This can be effected by either taking one additional full-size image every other minute, or by adding binned, compressed, and/or partial images even more frequently. The SXI imaging system enables synchronizing image execution to within a second during the observing minute, so that such additional images can be scheduled accurately.
These more complicated sequences are planned and developed using a spreadsheet layout (click for example). The top 16 rows of this spreadsheet are equivalent to the graphical display for the basic patrol example described above. The bottom four rows (labeled "Flare1" and "Flare2") contain the added flare images, which are a reduced-exposure thick polyimide pair and medium beryllium pair, both at 8 minute cadence. A top to bottom priority in each observing minute (i.e., along the columns in the spreadsheet) is assigned to each potential image, and there is a set of rules (too complicated to explain here) governing which are to be taken and which are to be skipped ("blocked"). The spreadsheet also contains information on which second within the minute an image is to be taken, what size it is, and so forth. From this layout, the tables used to encode the sequence are created for upload to SXI.
It is important to realize that while SXI can be switched to a "flare patrol" sequence by ground command, the usual route will be by autonomous action taken by the on-board software. SXI can be programmed to identify and locate two distinct levels of flare from the most recent, specified image, and to enable either of two "flare sequences" to be put into play. By the same mechanism, SXI reverts back to the "standard" patrol autonomously, once the flaring has subsided. These transitions -- into and out of flare mode -- are initiated at the start of the observing minute, and SXI maintains internal counters such that the cadence and phasing of the standard patrol sequence are preserved throughout. Thus an uninterrupted sequence of the broadest-band imagery continues smoothly while SXI repeatedly jumps in and out of flare mode during busy cycles of solar activity.
Finally, it should be mentioned that SXI can be operated in "Take a Picture" (TAP) mode, in which ground commands are used to take individual exposures. This method is reserved for test purposes.
This Section Under Construction
Operation of the SXI is a joint effort between NOAA/SEC and NOAA/SOCC (Spacecraft Operations and Control Center). Together, with feedback from operational users, they plan changes to the SXI observing program. Typically, new sequences are developed and tested at SEC using an extensive, specialized set of software created by LMSAL. The new table loads containing the encoded sequences are conveyed to SOCC via a shared, secure computer link. They are re-verified by SOCC before being uploaded to SXI for execution.
During post-launch test (PLT), special sets of sequences are run for the first few weeks to verify instrument performance and commanding capabilities. These require frequent changes to and interruptions of imaging sequences. Once these are completed, however, SXI observations are conducted in a more regularized mode anticipating the final operational state.
The PLT period will last for about six months, at which time GOES-N will be put into on-orbit storage. When one of the two currently operational GOES satellites fails, GOES-N will be revived as GOES-13 and it will enter full operations following a short reactivation test period. GOES-13 has a design lifetime of 5 years.
Once GOES-13 is in routine operation, the spacecraft scheduling cycle dictates that sequence changes must be planned and approved two weeks before they are implemented.
Bruner, M. E., R. C. Catura, J. E. Harvey, P. L. Thompson, and P. B. Reid, "Design and performance predictions for the GOES SXI telescope, in Proc. SPIE, Vol. 3442, 'Missions to the Sun II', Clarence M. Korendyke; Ed., p. 192-202, 1998.
Harvey, J. E., P. L. Thompson, and A. Krywonos, "Hyperboloid-hyperboloid grazing incidence x-ray telescope designs for wide-field imaging applications," in Proc. SPIE Vol. 4012, 'X-Ray Optics, Instruments, and Missions III', Joachim E. Truemper; Bernd Aschenbach; Eds., p.328-341, 2000.