Space Environment Center
SEC home
|
Space Environment Center |
SEC home |
| GOES Solar X-ray Imager |
|
|
|
|
|
|
This web page and its links contain an overview of the SXI instrument and its operations. These links include:
Further information on the engineering aspects of the instrument can be found in the GOES-12 Handbook section on SXI.
The SXI instrument consists of a telescope assembly (Figure 1) and its
associated electronics boxes. The telescope itself can be thought
of in three basic physical segments: The optics section, the metering
assembly, and the camera section.
Figure 1: Telescope Schematic
Soft X-rays in the spectral region SXI 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 taking one or more bounces at very shallow
angles. SXI uses a Wolter I design with a single piece of glass fashioned
into two sections, one parabolic and one hyperbolic. The SXI mirror design
has substantially poorer spatial resolution near the edges of its field
of view. Similarly, it suffers vignetting of about 20%.
Figure 2: Optics Summary
Figure 3: Detector Summary
The SXI detector stack consists of a microchannel plate (MCP), a phosphor, a fiber optic taper (FOT), and a CCD detector. This complex detection system allows substantial sensitivity to emission temperatures less than 2 MK, whereas a bare CCD would be biased towards hotter plasmas. The MCP serves multiple purposes. First, it is not sensitive to visible light which eliminates the need for extra filters in the optical train. Second, it provides amplification for detected photons. Each photon detected causes the emission of at least one electron from the MCP. A large potential (500-1000V) across the MCP provides energy to the electron as it travels down a pore on its way to the phosphor. Each time the electron hits the wall of the pore, its energy causes the release of more electrons. This cascade results in the creation of 1000-10000 electrons per detected photon. The electron shower impacts the phosphor which gives off visible light. This visible light is routed to a CCD via the FOT. The taper is used to match the image scale at the MCP to the CCD pixel scale.
Figure 4: Detector Stack
This detector has a number of important features and artifacts. The primary feature is the fact that the MCP gain can be adjusted. Ideally, one would want to match the signal from a single detected photon to 1 DN output. This appropriately balances the photon noise and the digitization noise of the system and occurs for a setting of about 650 V. However, as discussed in the spacecraft considerations section, exposure times are limited to about 3 seconds. Thus, for cases of low signal, for example with the beryllium filters, one can limit the digitization noise by increasing the gain. In that case, photon noise dominates and dynamic range is decreased. A plot of MCP gain and dynamic range shows the 'sweet spot' for maximizing dynamic range around 550 V. It also shows that as we go below 1 DN per detected photon we are sacrificing sensitivity for dynamic range.
The detector creates several image artifacts. Two of the most noticeable are the wide angle 'halo', seen most clearly during flares, and the so-called 'line-noise'. Although the exact origin of the halo scattering is unknown, it is clear that it is introduced downstream of the optics. This is because even at very high gains, where individual photons are evident as points and can be counted, the halo remains diffuse. This can only happen if the single quantized X-ray photons have been converted to multitudes of electrons or visible-light photons. Even though the halo is clear when there is a flare, it also causes a general 'hazyness' of the image. The net effect is predominantly on detecting coronal hole boundaries.
The line noise results in a usually subtle striping of the raw images. The noise is actually due to the use of a strip of reference pixels, one for each CCD line. These reference pixels are used by the amplifier to subtract off a background level, but the reference pixel values are not sent down in telemetry. So, residual noise from dark current, etc. in the reference pixels affects the whole line for which it is the reference. Our background subtraction technique attempts to minimize the line noise that shows up in level one processed images.
Figure 5: Electronics Summary
The data rate for the SXI is 100 kilo-bits per second (kbps). This rate allows for the transfer of one image in about 36 seconds. Since SXI normally takes an image once per minute, this allows about 24 seconds for other telemetry. Image parameters such as exposure duration, gain, and filter are controlled by a set of tables in the ROM and RAM memory of the SXI computer. Typically, we copy parameters out of the ROM into the RAM, and then modify them for what we believe is optimal imaging.
Figure 6: The GOES-12 Spacecraft
The overall performance of the GOES-12 SXI is similar in spatial resolution to the Skylab S-054 telescope. Its spectral sensitivity bridges the gap between Yohkoh's SXT and SOHO's EIT. Performance is covered in detail in the Instrument Performance Summary.
To help forecast such events, the following SXI requirements have been set:
Individual images are characterized by six primary parameters. The first three, integration time, filter and MCP gain are rather straightforward and are very relevant to image interpretation. The last three, SADA offset, special features field, and phosphor gain affect the images, but are often not of as relevant 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 electronically shuttered. That is, only during the period in which high-voltage is applied to the MCP are photons detected. The exposure duration can be set from 1 ms to approximately 65 seconds. The effective lower limit is 1.5 ms because the rise time of the high voltage supply is about 0.5 ms. The effective upper limit is ~3 seconds because the solar array drive assembly (SADA) steps this often. Longer exposures would be substantially blurred. Nominally, images are synchronized with the SADA such that images are not taken during SADA steps.
The SXI provides a set of analysis filters to help discriminate between solar sources of emission with difference spectral signatures. Each filter has a bandpass designed to be sensitive to a certain temperature range. There are three basic types of filters: the so-called 'Open' filter, which is actually no filter at all, the polyimide filters, and the beryllium filters. There are three polyimide filters categorized by thickness: thin, medium, and thick. Similarly, the beryllium filters are thin, medium, and thick. A comparison of several filter bands to solar plasma spectraat different temperatures shows that 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. Similarly the open filter blocks no additional shortwave photons, but increases relative transmission of cooler features.
The MCP gain is the voltage placed across the plate and controls the amplification of incomming photons as they are converted into electrons. The higher the gain setting, the greater the number of electrons for a given photon. Typical settings are 550-700V.
The SADA is the Solar Array Drive Assembly. It takes two discrete steps, back-to-back, once every three seconds to track the Sun. The offset determines when the image will be taken between double steps. If there was a clear period that was particularly quiet, we would set this parameter such that our images were integrated during that time. Generally, we take our images as late as possible in the 3 second interval between SADA steps.
The special features field is a bit field that allows control of various options in the images. For example, it allows images to be taken with or without the line-advance that tracks the Sun on the CCD. It is sometimes desireable for test images to be taken without line advance.
The Phosphor gain sets the potential of the phosphor. This gain may have a small effect on image focus, but no significant changes with voltage were observed during post-launch testing. We typically operate at 5000 V.
Figure 7: Imaging Sequence Summary
Sequences of images are controlled by the Image Control Table or ICT stored in ROM and RAM in the SXI computer. This table contains sets of sequences of images. Each sequence is constructed something like a computer program with a set of pointers, counters, and loops. Each entry in a sequence contains a pointer to an entry in the Exposure Setting Table (EST). Each entry in the EST contains pointers to the six image Parameter Tables (PTs). The PTs contain settings for the image parameters discussed above in section 3. By controlling this suite pointers and counters from ground command, NOAA is able to set up different sequences.
Typical sequences are constructed from pairs of images taken with the same filter, same gain, but different integration durations. Thus, a short exposure - long exposure image pair can be 'composited.' In the case of a bright flare, the long exposure will be saturated. This degrades locating the centroid of the flare and any examination of the flare morphology. The short exposure is intended not be be saturated and so is better suited for flare location and morphology.
We can take images without a gap for this housekeeping data, however, we don't do this for extended periods, because the housekeeping data is important. If we operate in a so-called 'high-cadence' mode, we would typically take a series of several images at ~36 second intervals then resynchronize to a one-minute cadence to allow a gap for housekeeping data. This would repeat.
To simplify the selection of images, 'intent categories' have been assigned to different EST indices. These intent categories include:
The basic concept is that a range of exposure times and gains provide images that best represent certain types of solar features in certain filters. Thus, an OPEN, 3 sec, 550 V image is categorized a Coronal Structure. It provides high sensitivity and good dynamic range for seeing fainter features. Alternatively, a Beryllium, 10 ms, 700 V image is better suited to the hot, bright plasma of flares.
SXI images are downlinked as part of the telemetry stream to NOAA's Space Environment Center (SEC) in Boulder, Colorado. This stream is assembled into images, processed to remove defects, and used to generate products for forecasters. In addition, the images and products are archived to preserve the historical record of solar activity and to allow researchers to improve forecasting models. The ground data system implemented at SEC for the SXI provides a robust conduit for data and products with significant redundancy and graceful degradation/recovery in case of systems failures. The system is depicted in the figure below.
Figure 8. Ground Data System
SXI telemetry data is received at the NOAA Boulder ground station and assembled into raw images by the ingest and preprocessing system. This real-time process also generates other instrument related non-image products. Both product types are created as files, which are transferred to a circular file buffer on the image processing platform for further processing and distribution. The raw telemetry stream can be replayed, for data recovery or product improvements, simultaneously with real-time telemetry.
The image processor generates cleaned and calibrated images, archives meta data in a relational database, archives image files on a mass storage system, and forwards images to the real-time operational server. This processor is composed of control code that oversees a set of image algorithms and support software and provides the interfaces to other elements of the ground data system. The image algorithms are written in a high-level language to take advantage of massive image processing libraries.
SXI images are stored as files on a RAID (Redundant Array of Inexpensive Disks) system. A relational database is used to index and track the image files and contains all ‘metadata’ about the images. Stored data has two applications: operations and archives. SEC maintains an operational store of 60 days worth of image files, but also stores all the metadata. NGDC archives all images and metadata. The images are archived at NGDC in real-time as they are received at SEC. Both SEC and NGDC maintain redundant database/file servers with the SEC system having automatic fail-over capability. The metadata is used to create 'on-the-fly' products such as static movie loops and real-time movie loops that update as new images arrive.
Operational users of SXI data include NOAA, USAF, and NASA. Distribution of the SXI products to these users is done using Common Object Request Broker Architecture (CORBA). A server provides access to all real-time and archived products to registered clients over intranet or internet. This operational server receives products in near real-time from the image processor and provides rapid and flexible access to the operational database/file server for archived products.
For further analysis, forecasters and other operational users need quantitative analysis and visualization tools. As part of the ground data system, SEC has developed a single image analysis client for quantifying solar features. It has also developed a flexible 'movie' player for visualization and assessment of dynamic changes in the corona.
For researchers, the public, and other non-operational data users, distribution is through the World Wide Web. While SEC maintains a cache of recent images in compressed (*.png) format on its web site (http://www.sec.noaa.gov), as the archive of SXI data, NGDC will be the main source of data for the public (http://www.ngdc.noaa.gov). At this web site, users are able to browse data products and select from various formats. The default display shows the most recent sequence of images as a table of thumbnail images. The browse capability is flexible and allows users to step though the archive with a programmable interval, i.e., 1-minute, daily, solar rotation, etc. A significant search capability is built into the web site and a wide range of data retrieval options are provided.
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. These changes must be planned and approved two weeks before they are implemented, because this is the duration of the spacecraft scheduling cycle.
During post-launch test, SWO issued special detector protection messages to SOCC for flares greater than M5. This may continue into operations.
Figure 9: SXI Operations
Balch, C.C., S.M. Hill, V.J. Pizzo, and D.C. Wilkinson, First Forecast Products from The GOES-12 Solar X-ray Imager, in 82nd AMS Annual Meeting, Sixth Symposium on Integrated Observing Systems, Orlando, Florida, 2002.
Berthiaume, G.D., B. E. Burke, J. A. Gregory, and P. L. Bornmann, Improving the Performance of the GOES Solar X-ray Imager (SXI) with a Back-illuminated X-ray Sensitive CCD, SPIE, 2812, 552-548, 1996.
Bornmann, P.L., D. Speich, J. Hirman, V. Pizzo, R.N. Grubb, C. Balch, and G. Heckman, The GOES Solar X-ray Imager: Overview and Operational Goals, in GOES-8 and Beyond, Proc. SPIE Vol. 2812, pp. 309-319, 1996.
Bornmann, P.L., V. J. Pizzo, D.Speich, S. Cauffman, R. Hooker, K. Russell, S. Wallace, J. Davis, S. Buschmann, R. Beranek, The First and Following Solar X-ray Imagers (SXI), in AAS Solar Physics Division Meeting, 1997.
Corder, E.L., Development of the Solar X-ray Imager (SXI) Camera, SPIE, 2214, 294-300, 1994.
Davis, J., D. Bagdigian, S. Buschmann, K. Russell, and K. Wallace, The Solar X-ray Imager for the Geostationalry Operational Environmental Satellite (GOES), in AIAA Space Programs and Technologies Conference, Huntsville, AL, USA, 1994.
Hill, S.M., S. Barsness, L.D. Lewis, J. Vickroy, C.C. Balch, A. Muckle, V.J. Pizzo, D.C. Wilkinson, and A.T. McClinton, GOES M Solar X-ray Image Availability, in 17th Conference on Interactive Information and Processing Systems (IIPS), 81st AMS Annual Meeting, Albuquerque, New Mexico, 2001.
Russell, K., J. Briscoe, E. Corder, S. Wallace, and J. H. Chappell, The Solar X-ray Imager (SXI) Detector Calibration and Characterization, SPIE, 2812, 638-650, 1996.
Smithers, M.E., and D.E. Zissa, Solar X-ray Imager (SXI) Optical Performance Analysis, SPIE, 2805, 115-120, 1996.
Wallace, K.S., T. A. Brown, and K. A. Freestone, A Table-Driven Control Method to Meet Continuous, Near-Real-Time Observation Requirements for the Solar X-ray Imager, in 17th Digital Avioncis Conference, Seattle, Washington, 1998.
