Sumbandila Mission Blog

System limitations

2010-03-30T16:11:00.000+02:00

Although a hugely successful mission thus far, the past six months have unfortunately not been without incident. We suffered a permanent loss of the Z-axis reaction wheel early on during commissioning, where after the ADCS algorithms have been adapted so as to still allow for controlled imaging with the remaining two wheels - Yaw-control was then slightly impeded and the off-Nadir Roll angle limited to ±20°.

Due to a power system anomaly, we have sadly lost access to one of the two CCD control boards. Each control board interfaces to one three-colour CCD. The board we lost contains the Green, Xantrophyl and Blue spectral bands, and the remaining board the Red, Red-Edge and Near-Infrared bands.

We have recently also lost the Y-axis reaction wheel (most likely also due to a power distribution failure), this meant the loss of three-axis control, but amazingly we can still take high-resolution images as before: the ADCS algorithms have been updated very innovatively whereby the satellite now “flies” in a derived Y-Thompson mode (the satellite tumbles “head-over-heals”). The spin rate in the Pitch axis is now 0.959°/s which results in an Forward Motion Compensation (FMC) rate of 4. This allows the imager to scan a target from back-to-front from the point where the satellite is pitched 15° forward over the target, to the point where it points 15° backwards. During the period that the image is scanned, the Roll angle is controlled by the remaining X-axis wheel, where the off-Nadir Roll angle is now limited to ±5°. The magtneto-torquers are used to phase the Pitch-rate such that the sub satellite point overlaps with the desired target at the center of the image.

We have obtained very good imaging results in this new control mode (see Miami and Dubai images in a separate post). The same controller is also used during data download sessions, but with a Pitch rate of 0.6°/s, and a Roll angle close to zero. During three download sessions last week on the same day, we managed to capture 6.3GB of image data via the S-band transmitter link, running at 72 Mbit/s!

High-resolution images

2010-03-30T15:26:00.000+02:00

Here are some of the long-awaited images from the main on-board payload, a F#4 linescan imager (refractive design) with 400mm focal length, and resultant GSD of 6.25m from an orbit height of 505km. See separate posting describing the system limitation as applicable to these images.

Buenos Aires (Argentina)

Date: 21 Dec 2009

Scene size: 60km x 52km

Colour bands: Red, Red-Edge, NIR

Post-Processing: SAC (CSIR)

Cropped from main image (100% zoom factor)

Nkandla (Kwazulu, Natal, RSA)

Date: 6 Dec 2009
Scene size: 40km x 52km
Colour bands: RGB
Post-Processing: SAC (CSIR)

East London (RSA)

Date: 3 Feb 2010
Scene size: 12.5km x 7.5km
Colour bands: Red, Red-Edge, NIR
Post-Processing: SAC (CSIR)

Sossusvlei (Namibia)

Date: 14 Feb 2010
Scene size: 16km x 10km
Colour bands: Red, Red-Edge, NIR
Post-Processing: quick view (WHSteyn)

Stellenbosch (RSA)

Date: 20 Feb 2010
Scene size: 6.5km x 7.1km
Colour bands: Red, Red-Edge, NIR
Post-Processing: quick view (WHSteyn)

Cape Town (RSA)

Date: 24 Feb 2010
Scene size: 17km x 19.6km
Colour bands: Red, Red-Edge, NIR
Post-Processing: quick view (WHSteyn)

Miami (Florida, USA)

Date: 16 Mar 2010
Scene size: 8.4km x 7.8km
Colour bands: NIR
Post-Processing: none

Dubai (UAE)

Date: 24 Mar 2010
Scene size: 8.5km x 7.7km
Colour bands: NIR
Post-Processing: none

Note shadow of Burj Khalifa

6 months in space today!

2010-03-17T12:40:00.000+02:00

Time flies when you're having fun! We are indeed grateful for this milestone and look forward to many more happy anniversaries. Some mission stats to date:

Number of orbits: 2,750
Months in space: 6 (today!)
Distance travelled: 0.74 AU
Activities uploaded: 1900
Images taken successfully: 93
Firmware uploads: 25
ADCS SW updates: 12

Striking image taken by Horizon Sensor

2010-02-10T15:08:00.002+02:00

Star camera commissioning

2010-02-03T13:43:00.000+02:00

The star camera is currently being commissioned. Below is a picture taken by the camera with four stars indicated that were extracted by the on-board algorithm used to detect valid stars.

GPS commissioned

2010-02-01T21:28:00.001+02:00

The graph below shows the telemetry from the on-board GPS over three orbits. The bottom trace indicates a valid "lock"; the yellow trace the number of satellites tracked; and the top three traces the X,Y,Z position vectors.

In a perfectly circular orbit, the satellite velocity would be constant. In a near-circular orbit (as is the case with Sumbandila), the velocity fluctuates slightly between a minimum at Apogee and a maximum at Perigee. This effect is clearly visible in the measured velocity data from the GPS (see below). It shows a difference of 30m/s between Apogee and Perigee. Also note the very low level of "noise" on the measurement.

Orbit correction

2010-01-28T09:58:00.000+02:00

The graph below shows the effect on the Semi-Major Axis (SMA) of the Sumbandila orbit after the successful propulsion events.

Propulsion system commissioned

2010-01-27T14:43:00.001+02:00

Sumbandila has a small (2.5l) Butane propulsion system which we have "fired" twice over the last couple of days: first for only 15s and the second time for 60s.

The telemetry of the second event can be seen below. The top trace (red) is the pressure in the propulsion tank. The second trace (purple) is the nozzle pressure. The green and yellow traces represent the temperature of the pipe leading to the nozzle (thruster) and the nozzle temperature respectively. Lastly the blue trace shows the pitch angle of the satellite, which has been maneuvered to 90° prior to the event in order to align the thruster with the satellite velocity vector.

The image above shows the propulsion system being integrated into the satellite (bottom compartment).

Commissioning timeline

2010-01-19T08:45:00.000+02:00

There remains just under two months available to complete the outstanding commissioning activities. Thereafter the plan is to transfer operations of the satellite to the Satellite Application Center (SAC) at Hartebeeshoek (Pretoria).

The following is a list of the main outstanding commissioning activities to complete:

Star camera
Propulsion system
GPS
Resistive magnetometer
Experimental payloads
Viewfinder functionality

The graph below shows the summarized timeline since launch:

Imager refocusing

2010-01-19T08:16:00.003+02:00

Initial images taken by the main payload imager, indicated that a focus shift had occurred, probably due to a combination of launch vibrations and an elevated thermal environment (higher than expected temperatures). The images furthermore indicated that this shift was towards the blue image band (away from red).

The imager is designed with a built-in mechanical refocusing mechanism, so we attempted the refocusing of the imager by making use of a live PAL video feed (9 Dec 2009). This feed originates from the matrix sensor that is co-located on the focal plane with two line sensors. The focusing was done on four steps as shown below with the snapshot images (cropped) from the matrix sensor. These images have a resolution of 8.4m compared to the resolution of the main imager of 6.25m. A marked improvement was visible on both the video feed and the snapshot images that were downloaded afterwards. The focus was shifted by 300um towards the red band.

High-resolution images taken by the line sensors after the focus effort confirmed that the imager is now again in focus.

First image from the Horizon Sensor

2009-11-19T13:50:00.000+02:00

We have downloaded the first image from the on-board Horizon Sensor, which forms part of the sensor suite of the Attitude Determination and Control System (ADCS). Data from the sensor is used to improve the accuracy of the attitude estimators. The sensor automatically extracts the Earth horizon profile from the image and then passes the coordinates thereof on to the ADCS SW.

First high-resolution images

2009-11-17T14:59:00.000+02:00

We have received the first images from the main payload (6-band imager) at the capture station at the Overberg Test Range (OTB). With this significant milestone we have confirmed that the imager electronics and mass memory storage are working as expected. Over the next couple of weeks we will perform various system calibration functions to further optimize the image quality.

The first officially released image will be sent to the Department of Science and Technology when available. The Department may publish the particular image at its own convenience. Further images on this blog will be published with the consent of the Department.

SA-AMSAT payload activation over the USA

2009-11-08T20:37:00.003+02:00

The SA-AMSAT payload was again activated over several areas around the globe, with keen participation from many hams. The two passes over the USA were particularly crowded.

Shown below is the satellite footprint over the USA for the 14:18 UTC pass (left) and 15:52 pass (right).

From what I could decypher from the recordings, the following callsigns were heard (callsigns from the 14:18 pass are shown in yellow and those from the 15:52 pass in red):

You can listen to a recording of the first pass activity here and of the second pass here (thank you Johan Papay, K8YSE EN91, for making the recordings available).

Data transmitter activated

2009-11-06T18:39:00.001+02:00

We have for the first time activated the S-band data transmitter on the satellite and captured the transmitter-generated test pattern at the Overberg Test Range (OTB) station near Bredasdorp. During this exercise the satellite was in a target-tracking control mode, with the S-band waveguide antenna "tracking" the OTB station.

The image below shows the captured test pattern.

Fine Sun Sensor calibrated

2009-11-06T17:38:00.000+02:00

Below is an image taken by the Fine Sun Sensor before calibration (right) and after the sensor output levels have been calibrated (left)

Three-axis stabilized

2009-11-05T22:18:00.000+02:00

The satellite is now in a full three-axis control mode, using the three orthogonally placed reaction wheels, data from the calibrated fine sun sensor and the fiber-optic gyroscopes. This is a major step in the direction of taking the first images with the main payload camera.

SA-AMSAT payload status

2009-11-03T16:37:00.000+02:00

For those interested in the progress made on the SA-AMSAT payload commissioning, please visit / join the Google Group that was specifically established for this purpose (follow link at the bottom of the right-hand column on this page). The discussions on this group are of a more technical nature. The required CTCSS tone to activate the transponder has for instance been published on this group.

During the commissioning of Sumbandila we try to work in the scheduling of the SA-AMSAT payload as often as we can, but the service can by no means be guaranteed at this stage. The commissioning of the satellite is expected to last more or less six months (from date of launch), during which time our primary focus is on demonstrating success with the high-resolution imager - the main payload on the satellite.

The activation schedule for the SA-AMSAT payload will be published on the AMSAT-BB, in addition to being visible on the Google Group, as and when our satellite commissioning schedule permits activation.

We appreciate the eager participation of all Hams worldwide, and would like to encourage making your feedback visible on the Google Group.

ADCS code updated

2009-11-03T10:10:00.000+02:00

We have completed a major Attitude Determination and Control System (ADCS) software code upload to the satellite. As a result the satellite is once again in a Nadir-pointing attitude.

In addition, we continued with periodic scheduling of the SA-AMSAT payload and recently had contacts through the satellite from the USA and Australia.

We are now working towards obtaining the first high-resolution images from the on-board linescan and matrix detectors.

Fine Sun Sensor image

2009-10-24T17:29:00.001+02:00

This image of the sun was taken by die Fine Sun Sensor today. The sun centroid is then extracted by SW using a region-growing algorithm. The centroid is required by the more advanced attitude determination estimators in order to 3-axis stabilize the satellite.

Below is a graph depicting the extracted X and Y coordinates of the centroid, calculated 1s apart, for a duration of 1 hour and 4 minutes (total sunlit part of one complete orbit).

OSCAR number (SO-67) allocated to Sumbandila

2009-10-21T08:42:00.003+02:00

Given the recent success with the SA-AMSAT payload on the satellite, we have been officially assigned the OSCAR number of SO-67 by AMSAT (USA). See message of Bill Tynan (AMSAT OSCAR number coordinator) below. For more background on assignment of OSCAR number, see http://www.amsat.org/amsat-new/information/faqs/numbers.php. SUNSAT was allocated the OSCAR number SO-35 in 1999.

You can also listen to an edited version of the latest ARNewsline report here (see www.arnewsline.org for more detail).

I am in receipt of an e-mail sent to the AMSAT-NA Board of Directors by Hans van de Groenendaal, Secretary SA AMSAT.

The e-mail states that the amateur payload on SumbandilaSat was developed by SA AMSAT and incorporated by the University of Stellenbosch into the main payload and that the SA AMSAT payload was officially coordinated through IARU Satellite Frequency Coordination Panel with an uplink of 145,880 and downlink of 435,350 MHz

The e-mail further states that the amateur radio transponder on SumbandilaSat was successfully switched on from the ground by ZS6BPZ during the test phase on Sunday 18 October 2009 and that several QSOs were made through the transponder.

The e-mail also requests that an OSCAR number be allocated to this spacecraft.

Therefore, by the authority vested in me by the AMSAT-NA President, I hereby designate this latest amateur radio satellite as SumbandilaSat Oscar 67 or SO- 67.

On behalf of AMSAT-NA and the world's amateur radio satellite community I congratulate those responsible for building, testing and launching this new satellite. May it have a long and successful life.

73,

William A. (Bill) Tynan, W3XO

SA AMSAT payload tested successfully

2009-10-19T13:12:00.004+02:00

During the first morning pass of Sun 18 Oct we have activated the voice beacon (435.350 MHz) at 07:48 UTC and switched on the bent-pipe transponder (uplink 145.880 MHz, downlink 435.350 MHz) during the first evening pass at 19:13 UTC, with several Radio Amateurs then operating through the satellite transponder.

You can listen to a recording made during the evening pass here.

Below is a graphic depicting the location of the six Radio Amateurs that partook in the first test.

Feedback from some of the Hams that participated in the first test:

Congratulations to all part of the success of the first Sumbandilasat comms. Well done! (Eddie Leighton, ZS6BNE)
When I was on freq the Sat's audio was clear with a S7 signal which sounds just like a local repeater. Congrats to the TEAM ! (Andre van Deventer, ZS2BK)
The down link was pretty strong - It's at least 10dB stronger than AO51/SO50 given the same setup. Good news indeed. Suggest operation from Quadrifilar Omni with a low noise preamp on 70cm might be quite achievable (Allan Saul, ZS1LS)

First month in space

2009-10-17T23:28:00.001+02:00

Sumbandila completed 457 orbits during its first month in space which corresponds to 0.1316 AU (1 AU = distance between Earth and Sun). We managed to make successful contact with the satellite during all the available* passes over the Stellenbosch ground station, namely 107 (or roughly 23% of all orbits). During these 107 ground station passes, we had a total contact time with the satellite of roughly 15 hours.

* Passes with a maximum elevation angle of at least 4°.

First video received from the satellite

2009-10-13T16:55:00.000+02:00

We are blessed - have received the first real-time video images from the satellite during the two morning passes! The signal was simultaneously received by the ground station at Stellenbosch University and the one at SAC (Hartebeeshoek) - the one in Stellenbosch with a 4.2m dish and the one at SAC with a 10m dish. The signal-to-noise ratio of the latter was obviously much better.

You can see an edited version of the video below, captured in Stellenbosch. At the time the satellite was moving over the North-Western corner of Namibia in a South-Western direction. In the second half of the video you will notice the top part of a VHF antenna visible at the top of the frame.

Effects of radiation

2009-10-12T14:57:00.000+02:00

We have noticed that some of the SRAM devices in use on the satellite are sensitive to Single Event Effects (due to radiation). This typically manifests as increased current.

All our electronics are protected by latch-up circuitry, but some of the increases in currents observed are below the set thresholds of the built-in trip functionality of these circuits.

We have now implemented a further protection mechanism in software which necessitated code uploads to the main on-board computer, some of the power distribution units, as well as the attitude computer.

In parallel we continue to perform minor commissioning activities and regular health checks. We plan to resume commissioning of the attitude control systems this week.

Satellite Nadir pointing

2009-10-03T11:50:00.000+02:00

The ADCS computer now runs a Y-momentum wheel controller that stabilizes the satellite with near-zero roll and yaw angles using data from the coarse sun sensors and a magnetorquer nutation damping controller. The main solar panel faces the sun in this control mode. The Y-wheel is controlled to track a reference pitch angle (as close as possible to zero).

With the satellite now Nadir pointing (ie the imager aperture and antennas continuously faces in the direction of Earth), our communication with the satellite has improved further. The next step is to start commissioning the various attitude sensors so that we can activate full three-axis control on the satellite.

All other systems continue to operate within nominal limits.

New software for ADCS processor

2009-09-30T16:20:00.002+02:00

We have successfully uploaded new software for the ADCS processor (T800), containing the following changes:

New coefficients for the 2009 IGRF model
Latest NORAD elements
Calibration values for the fluxgate magnetometer
Minor modifications to code functionality

After verifying that the code is executing correctly, we have switched on a Y-Thompson attitude controller, for the first time transferring attitude control of the satellite to the ADCS processor (from the magnetic subsystem on which a basic controller was run up to now). From here we can proceed with the commissioning of the more advanced attitude controllers, which will finally lead to a 3-axis stabilised satellite.

The OBC is running more reliably after modifying some EDAC parameters. We have identified a couple of improvements to the flight code which will be uploaded soon. This will affect only some of the software modules running on the QNX operating system, so that we do not have to upload and replace all the flight software.

Whole Orbit Data (WOD) collected

2009-09-26T11:02:00.000+02:00

We have made a further improvement to our ground station setup by replacing the existing UHF yagi antenna with one which has 4dB more gain. It is now mounted on the dish which eliminates some obstruction we experienced previously. During last night's passes we had a marked improvement in communications.

We have now downloaded almost 500kB of WOD, including fluxgate magnetometer readings which will enable us to calibrate the magnetometer. The ADCS system is continuing to perform well, as are the thermal and power subsystems.

There seems to be a high incidence of Single Event Effects (SEE) due to radiation, especially on the On-Board Computer (OBC). We have however designed all the electronics with latch-up protection, so none of the SEE occurrences to date have caused any permanent damage. Sometimes an SEE causes the OBC to reset and we have recently noticed that an auto restart of the processor then does not fully complete. With the WOD now available to us it seems that the lengthening of a timing constraint in software will be able to alleviate this problem. Note that we could not test for SEE on the ground so were not able to observe this timing effect prior to launch.

Sumbandila Two-Line-Element (TLE) set

2009-09-24T14:39:00.007+02:00

Peterson AFB in America informed us that they have incorrectly labeled Sumbandila as Fregat-IRIS. They are part of the radar tracking network responsible for uniquely identifying over 20,000 objects in Earth orbit! We have therefore up to now tracked the Fregat-Iris (35867) in stead of Sumbandila (35870).

After correcting the TLE information, our communication with the satellite is more consistent and reliable.

We intend to affect the improvements that we made to our ground station equipment, to the SAC ground station in the coming week.

Sumbandilasat Technical Status Report No. 1

2009-09-22T12:35:00.013+02:00

Sumbandilasat has completed more than 50 Earth orbits in space and we are now in full swing with the various commissioning activities.

The two autonomous Attitude Determination and Control System (ADCS) controllers (Bdot and Y-Thompson; Bdot-Sun) running on a small micro processor on the magnetic subsystem (no floating point processing) are perfectly pitching the satellite at 2°/sec with the main body-mount solar panel facing the sun. The power system is therefore running optimally. We are satisfied with the thermal state of the satellite in the current basic attitude control mode.

As expected, the 500km orbit of the satellite causes incidences of short contact opportunities with the satellite due to the fact that our ground station is located at the Engineering Faculty (in Stellenbosch) where it is surrounded by various high mountains. We have identified certain aspects in our communication setup on the ground which can be improved to make the communications with the satellite more robust and reliable.

In addition, we are re-evaluating the possibility to extend the contact time with the satellite by also employing the Sumbandilasat TT&C ground station at SAC (Hartebeeshoek), by remotely controlling passes from Stellenbosch.

At the moment we are focusing on the commissioning of the main On-Board Computer that will allow us to collect whole-orbit telemetry data, calibrate the magnetometer and proceed to the more advanced ADCS control modes on the main Attitude Computer.

17/09/2009 20:48 Our first pass

2009-09-22T11:37:00.003+02:00

We are extremely pleased with our baby in space!! Our first pass last night at 20:48 local time was a very low elevation Eastern pass with some mountains obscuring the “view” as well, but the satellite responded beautifully almost immediately after sending the first command from the Stellenbosch ground station!

We received real-time telemetry which suggested that the satellite was already executing an autonomous attitude stabilization maneuver. This was confirmed during the second pass where we also had good communications. This morning the satellite was stable in a sun-Thompson mode.

Celebration on our first Contact with Sumbandila

Estimated angular rates

2009-09-18T13:57:00.002+02:00

The graph below shows the estimated angular rates 2 hours after launcher separation, already close to Y-Thompson spin at -2 deg/sec. The Y and Z angular rates are less than 0.5 deg/sec.