First 100 Days

Page TOC


5th June 2013 100 days

AAUSAT3 has now been in space for 100 days.

AAUSAT3 was launched 25. February 2013 on PSLV C20 and has 5th June 2013 14:37 CET been in orbit for 100 days.

BTW 5. June is also the Danish Constitution Day (Grundlovsdag) from 1849.

On the evening af June 5th, when AAUSAT3 passes the AAU ground station, with an AOS at 19:02 (CEST), the satellite will have been around the earth 1434 times, equal to 64.5 million km with an average speed of around 7.5 km/s in an altitude of 800km - and weights only 800 gram.

In that period our main payload - our in-house developed SDR AIS receiver - has received more than 700.000 AIS messages

so ...


our biggest space success ever

We are today in a good condition and our experiments work better than anticipated:

  • Everything is working ok. Lots of AIS messages received on our two AIS receivers
  • Communication (COM), power system (EPS) and log system (LOG) has been working perfect.
  • Ground station is running perfect and we are preparing another ground station at Thule in Greenland at 72 degrees North, which will triple download capacity

We do receive and download 10.000 AIS messages a day to our ground station in Aalborg, and statistics from all subsystems are pulled down on regular basis.

In the following status for all subsystems are given and two major crisis during the first 100 days will be explained.

We believe in sharing knowledge - both the good things as well as the not so good :-)


Communication subsystem

Overall, the COM and groundstation subsystems works beyond our expectations. Based on experience from AAUSAT-II, the communication system on AAUSAT3 uses an integrated transceiver chip and a custom packet format with multi-layer forward error correction (FEC). As a compromise between packet length and symbol energy, we started the mission at 2400 bps and with all FEC enabled.

We had two-way communication on the first pass over Aalborg (8 degrees max elevation) at a range of over 2500 km.

Because of the good link quality we moved to 4800 bps within the first 10 hours, and to 9600 bps after less than two weeks. We have briefly tested the communication link on 19200 bps, but experienced increased packet loss and decided to stay on 9600 bps as it gives us the highest throughput.

During the last 100 days, we have received 254963 packets from the satellite of which 46% has been received without any bit errors and close to 54 % received with correctable bit errors due to our FEC protocols (instead of AX.25). In total more than 20.6 MB of data has been downloaded from the satellite.

The COM subsystem works as a Cubesat Space Protocol (CSP) router, and forwards packets between the space link and the satellite's internal CAN-bus. CSP has proved to be very stable and has worked well as a platform for distributed satellite subsystems.

The Cubesat Space Protocol - CSP is an open source spinn off from AAUSAT3. We invite everybody use it.

During the 100 days, the COM subsystem has had one unintended reboot due to a watchdog timer reset. We have not examined the cause of this in further detail. See further down on this page for the situation.


the most experimental subsystem in AAUSAT3

AIS2 is the first SDR system from Aalborg University to fly in space. It is sampling a low-IF (intermediate Frequency 162 MHz down to 200 kHZ) and demodulating the VHF ship position reporting system, AIS.

Unique ships detected during a 12 hour period primo June 2013.
Unique ships detected during a 12 hour period primo June 2013.

For the first time in history, Aalborg University uses a processor at more than 400 MHz together with high speed SD-ram in space. Hence, we where excited at the very first pass above Aalborg ground station. Would our small Linux computer even boot in the space environment, without inches of shielding? Yes, indeed it would.

At the beacon received at 16:43:49 UTC+1, first pass ever after launch on 25th Feb 2013, we where very busy looking at EPS and COM information - to see if the spacecraft was okay - so we almost didn't notice the status from our AIS receivers; AIS1 502 CRC OK and AIS2 1181 CRC OK. Hence, the very first steps toward receiving AIS in space with small deployable antenna and low cost electronics and launch was achieved.

Our biggest concern during development was the low signal strength and colliding AIS messages, due to the highly increased line of sights, when spotting ships in 800 km height.

Hence, the SDR uses a lot of time to filter, demodulate and analyze the gathered data using a part scientific and part intuitive algorithm. This makes AIS2 work far from real time; it uses 10-30 sec. to process 1.2 sec of raw data. Today, this algorithm has received more than 700.000 AIS msg. with correct CRC. The following density plot shows what areas the current algorithms is performing in, and in what areas it does not (e.g. EU).

Ship density detected by AAUSAT3.
Ship density detected by AAUSAT3.

This plot is generated by dividing earth into 2.5 by 2.5 degrees steps, and counting the number of ships received in that area during a fixed time period. If more than 50 ships have been counted the number is limited to 50 to produce a nice looking plot. Also be aware of the polar dilution, as this is a Mercantor projection.

AIS2 has the ability to sample and save raw IF samples on a fault tolerant SD card. A few of these samples has been downloaded for further analysis on ground, to have a known basis for the development of an enhanced algorithm. This raw sample, shown on out homepage, has helped our algorithm developer team to fine-tune our automatic gain control, channel filters and demodulating method. This algorithm will be uploaded and taken through initial tests on AAUSAT3 during the next 14 days. Early testing on ground shows that this new algorithm, thanks to three sixth semester signal processing students, will be able to receive real time.

The following figure shows an eye diagram of a AIS msg. received in space more that 2800 km away from the ship. This eye diagram is possible to show due to the raw sample downloaded from space. Basically it shows the soft-output from the demodulator, with the x-axis corresponding to time in 80 samples per symbol, and the y-axis being positive equals a '1' transmitted and negative corresponding to '0' in the AIS transmission. The left figure below shows a histogram of the soft-output after timing-recovery (center of eye diagram) - here the algorithm identifies one bit, that could just as well be a '0' or a '1'. Hence, if you get a wrong CRC, this single bit is most likely at fault.

Histogram of soft-decision from the SDR demodulator and eye diagram
Histogram and Eye diagram from the SDR demodulator.


standard AIS receiver

AIS1 is an real time AIS receiver based on a adf7021 as radio and 9600 baud fsk modem and an avr90can128 8MHz as processing unit. From very first orbit AIS1 has collected AIS messages from ships. It has and are worked beyond expectations.

All operations has been without any problems. In average 1-2 AIS messages has been received every second of operation. Power consumption is around 150 mW.

After the first weeks of operations we have switched to only operate AIS2 due to the higher performance and AIS1 now works a cold backup for AIS2.


B-DOT detumbling attitude control system

Our detumbling system is performing well. It has taken our satellite from a spin of almost 1.5 Hz in the early stages of operation down to approximately 720 degrees per orbit. The remaining spin is due to the change in the Earths magnetic field experienced in our polar orbit and the spin ensures that the satellite antennas are not pointing in the same direction all the time. A calibration test have also confirmed that all 6 magnetic coils are working. However, to save power, we are only operating with half of our coils taking an average power of approximately 20 mW.

The first days of operation were quite hectic as we saw that the tumble rate gradually increased and our detumble algorithm seemed to make things worse. After reevaluation of all parameters we discovered that the feedback sign for our z-axis coil was incorrect. We did find the correct sign during a calibration test in Canada, but we forgot to save the value in flash memory and thus the incorrect setting was back after a power cycle. If the satellite was spinning more than approximately 1.8 Hz we would not have been able to recover as our magnetometer is not sampled fast enough for such a high spin rate. However, with new control parameters we luckily managed to detumble the satellite and stabilize our radio communication.

ADCS is based on a 8MHz avr90can128 with additional 32kByte RAM and is coded in C and running freertos.


advanced attitude control system

All our 6 coarse two-axis (azimuth and elevation) sun sensors are working as expected. These have been actively used to validate the EPS system and the power intake from solar cells. A magnetometer has also been used to verify the magnetometer on ADCS1 and makes it possible to do attitude estimation.

Our rate gyroscopes worked well during the first couple of days in space. However, they do not seem to work currently as they return very unrealistic values. They do give values that change and they are not either min/max values. Further experiments are needed to validate if we need to buy better gyroscopes for our next launch. Fortunately, they are not mission critical for AAUSAT3 and we also have a sun synchronous orbit which means that we can determine the satellite attitude and potential magnetometer bias without gyro data.

ADCS is running on a 40 MHz Arm7


Power supply and supervision system

After 100 days in space, the EPS is healthy and running well. In the start of the mission, the power input varied drastically. Around March 1st the power input was very low, and AAUSAT3 was powered down most of the time. During the very low power days, the EPS experienced around 3240 reboots. Around March 15th the average power input was around 2W. The power input has stabilized a bit, and is now in the range 1W - 1.2W. This is lower than expected, but enough to operate the AIS2 receiver around half of the time.

All sensors connected to the EPS are running normally. Sensors include temperatures for all subsystems, voltages on the supply rails, power consumption of subsystems, battery conditions. Temperatures inside AAUSAT3 are in the range 25-35 degrees Celsius. A temperature sensor close to the power amplifier of the UHF transmitter peaks in the mid 50s during heavy communication.

At May 24 the EPS experienced what is believed to be a Single Event Upset, that meant the CAN bus was not performing normally. The EPS was only able to communicate with CSP packets that had 2 bytes of data. By exploiting a "feature" in the software, it was possible to reboot the EPS which returned the communication to normal. The EPS is running with uptimes around 30-40 days, and has been rebooted 3 times since the low power days.


logging system

AAUSAT3 has a common logging system which all subsystem can use and ground can pull information from. It has worked flawless

Ground Segment

and Mission Control Center

To support AAUSAT3, a ground segment with a simplified USB version of the COM subsystem, named BlueBox, was designed. The software consists of an embedded part on the BlueBox, a server part and clients to actually serve the users. Compared to the software in the satellite, the embedded software on BlueBox is highly simplified, as the forward error correction and routing functionality is implemented in the server. Apart from these functionalities, the server also handles permanent storage of the received and transmitted data, and acts as a CSP endpoint which repacks the data to the clients.

The client is a GUI-based application, written in Python, from which all parts of AAUSAT3 easily and failsafe can be monitored and configured. The most important telemetry data and status from AAUSAT3 continuous are showed on the dedicated beacon tab as the beacons are received. Furthermore, one tab exists per subsystem, and a tab with the world map, showing the decoded and downloaded AIS (Automatic Identification System) messages with possibility to show detailed information for single messages.

Apart from the GUI client, a subsystem of the ground station automatically redirects all downloaded AIS messages to the Danish Maritime Authority, to be included in their analysis and surveillance systems.


Not all has been fun...

Two critical situations has taken place during the first 100 days:

28. Feb 2013

28. February 2013 06:08 the very last beacon for some days was received. During the last 24 hours battery voltage was decreasing and temperature readings did indicate that AAUSAT3 was steady pointing the sixth side without solar cells towards the sun. The last passes over Aalborg it was observed that battery voltage was slowly decreasing.

At the same time tumbling was slowly increasing to 540 degrees/second (1.5Hz).

Not a funny situation.

So power and tumbling was major issues.

For the tumbling it was realized that the ADCS1/B-DOT detumbling was spinning up in one coil direction. So ADCS1 was switched off. It was realized that a minor error in last calibration had missed a direction bit (operator fault) so a new B-DOT was calculated for upload.

But it was not ready before the next situation...

Telemetry from the low power period
After contact was reestablished we downloaded telemetry from the EPS. This figure shows the battery voltage and the power input from the solar cells. During the period where very little input power is indicated, the EPS is subjected to a lot of reboots. This is because the satellite shuts completely off around 6.6V to protect the batteries.

Later we did conclude that the sidepanel without solarcells was pointing towards the sun. The fast rotation, approx 2Hz, was an excellent spin stabilisation. As a result a very low energy harvesting and therefore lack of energy (could not survive on albedo). So bankrupt - no power.

Every time there was just a small fraction of energy AAUSAT3 did emit a beacon and drained the batteries again. The radio uses approx 2W. So how to com back to normal normal business ?

The situation was discussed and analysed during the day and it was decided to instruct AAUSAT3 on all upcoming passes to switch off beacon mode because it was using power (radio).

So GND was programmed to send this command agian and again in every pass.

No further attempt was taken to gain contact during the weekend because if very slow charging was taken place it should have peace to regain to normal mode. All passes was supervised.

There were total silence until 4. March 17:52 where the groundstation in Aalborg did request and did receive a beacon. The beacon did show an up-time of 1 day 19 hours and that everything was working ok and more than 7000 AIS messages received. Everything was went back to normal. The requested beacon did also show a high number of reboots - which we think was due to low battery voltage.

The new B-DOT coefficients was uploaded and rotation was rapidly going down from 1.5Hz to 2 rotations pr orbit.

Not so bad for a engineering student attitude control system. Based on simulation and the experiences from this event it seems we can handle tumbling up to 1.8Hz or 650 degr/sec.

Due to this event AAUSAT3 is now by intention set to very tumble slow - 2 rotations for every orbit (approx 100 minutes) - to avoid a repetition of the situation.

24. May 2013

... we did receive an email from Japan from JA0CAW. He did report unormal beacon. We could track this down to that the power system (EPS) onboard AAUSAT3 rejected to handle multi frame communication packages from ground. During 4 passes (500 minutes) the students did analyze the situation and inspected the C source for the EPS server and a solution was found. One proper crafted package was send to AAUSAT3 and normal situation for the EPS was re established. During the event the rest of the satellite was - due to the decentralised design - working perfect. The team expect a bit flip to be the case.

During the first 100 days more than 700.000 AIS packages has been received and downloaded.


We are very satisfied and the mission continues

/The AAUSAT3 team