Today PI9CAM´s QSL for the first Satellite Bounce QSO via an unmanned spacecraft done by radio amateurs arrived by mail. As we know, there have been previous commercial attempts for Satellite Bounce in the early 60s using ECHO 1 and ECHO 2 which were inflated balloons with diameters of 30 and 41 m. The initial orbits were at heights of 1500 km and 1200 km.
The theoretical radar cross section (RCS) of ECHO 1 was 700 m², but measurements by military radar stations resulted in 900 to 1000 m² in the beginning. Later, the satellite deformed and shrunk. OKEAN-O, the one we used, has a radar cross section of 18 to 20 m² but is in a much lower orbit at a height of 650 km. This leads to quite similar unit power budgets, regardless the difference in size,
Enjoy the movie “The Big Bounce” about our predecessors 55 years ago!
While ISS Bounce took Jan, PA3FXB, and me 2 months of testing and improving to succeed, Satellite Bounce was a much bigger challenge. Despite the fact, Jan and the team of PI9CAM are operating the 25 m dish of the Dwingeloo radio telescope, it took us nearly 2 years, enormous patience and scores of tests until we finally managed to receive “Rs” to complete a QSO today (December 8th, 2015). As far as we know, it is the first time ever, a two way amateur radio contact could be completed by using an unmanned spacecraft as a reflector.Above screenshot shows the position of the satellite at the end of the QSO. The Satellite rose in SSE and set in NNW. A calculative common window opened at the point, marked “O”. Local obstructions were not considered. Due to safety reasons transmissions in Dwingeloo are limited to elevations above 10°. So the AOS (acquisition of signal) happend shortly before the groundtrack of the Satellite crossed the 40th degree of latitude northwards, as soon as PI9CAM started transmissions. Sum of slant ranges (distance between ground station and satellite) was 3400 km at the beginning and 2000 km at the end of the contact.
Much of the reflections remained below the noise floor, but this one of PI9CAM, right at the beginning (14:10:10 UTC), is a nice example, of what can be received:
And vice versa DJ5AR as to be heard in Dwingeloo (14:11:00 UTC):
The used object OKEAN-O (NORAD #25860) is a joint Russian-Ukrainian Earth observation satellite, launched on July 17th, 1999 by an Ukrainian Zenit-2 carrier rocket. The satellite is in a polar orbit of about 650 km height with an inclination of 98°. The mass is 6.2 tons and the RCS (radar cross section) is figured between18 and 20 m². It has been used for research of natural resources, ecological monitoring and hazards prevention. Designed for a life time of 3 years, it is out of service now.
QSO in WSJT-X
In use by the ground stations were the 25 m radio telescope in Dwingeloo by PI9CAM with 120 W and a 3 m dish with 150 W at the feed by DJ5AR in Mainz. The mode used was digital JT9H that comes with the new WSJT-X software by Joe Taylor, K1JT. The transmit/receive periods were set to 10 seconds, working around a center frequency of 1296.300 MHz. The automated Doppler tracking (+/- 60 kHz) has been performed for the complete path on DJ5AR´s side with a homebrew tracking software. The calculative power budget during the QSO was about -154 dBm. This value is very optimistic, as it presumes the optimum reflectivity of the satellite, which depends on its orientation.
Conclusion: The main difficulties in this game are:
Selection of suitable satellites, depending on radar cross sections and slant ranges.
Compensation of the Doppler shift with a maximum rate of 600 Hz/second.
Following Jan´s (PA3FXB) suggestion, we tried the new experimental WSJT-X software. The mode, we chose was JT9 H. Also we agreed in trying full doppler compensation to be used on my side. Everything worked fine, as can be seen in the screenshot below.
Back from our holidays Jan, PA3FXB, and I had another test via ISS Bounce on 23 cm today. As I located a bug in my Doppler correction software, causing unwanted steps, it could be fixed by finding a workaround for the malfunction in compilers NOW() routine, for returning the correct time in milliseconds. So the improved Doppler correction is working smoothly as can be seen and heard in the signals received.
PA3FXB in JTMS received by DJ5AR via ISS Bounce. Center frequency was 1296.300 MHz.
We even had some kind of conversation at the end 😉
As further tests showed, the full doppler correction on my side is working very well now. This enables potential sked partners to work on a fixed frequency by just tracking the International Space Station with the antenna.
Sked requests are welcome: dj5ar (at) darc.de
Modes, successfully being used so far: CW, SSB, ISCAT, JTMS
After many previous attempts Ronny, SM7FWZ and Jan, PA3FXB, managed to complete an ISS Bounce QSO on 23 cm today. The rigs used, were a 3 m mesh dish with 375 W on the dutch and a 4 m solid dish with 300 W on the swedish side.
Ground track of the ISS versus the direct path between SM7FWZ and PA3FXB
The window opened from 2015-07-13 08:07 to 08:17 UTC. While the antenna tracking was done automatically by both stations, only Jan could perform Doppler compensation. So they used a kind of mixed mode in tracking the frequencies.
Ronny transmitted on 1296.300 MHz and tried to catch Jans signal manually. So I monitored the frequency, Ronny should be heard on my side. It is nice to see in the video, how Jan appeared on the very left, getting closer and closer to the frequency, Ronny could be heard.
Since QSOs via ISS Bounce are quite easy to perform, Jan, PA3FXB, and I evaluated the possibilities, to use other objects in orbits around the Earth. The table below shows a selection of objects in earth orbits. There are some quite big ones in geostationary orbits, but the distance is the most limiting factor, not the size, as can be seen in the predicted maximum reception level provided by the radar equation. So I had an intense survey on the objects in low orbits and determined the maximum linkbudgets, to filter the most promising ones. My PC had to work a couple of hours to perform that. On the end it has been a little surprise, that COSMOS 1823, an old soviet geodetic satellite, wasn´t among them. We used it for previous tests and were able to detect faint reflections.
At our meeting on the Dorsten GHz Convention last Saturday, we discussed latest details. On Sunday, Jan and the crew of PI9CAM activated the Dwingeloo Telescope for the 23 cm EME SSB contest, but they had some spare time for other experiments. Just to warm up we had a nice SSB QSO via the moon first, right after their first QSO PA/JW: Congratulations!
Some of the top objects would pass on suitable elevations and we decided to start with NORAD 39771 which is the second stage of a japanese H2A202 rocket, launched on May, 24th 2014 with a radar cross section of 20.7 m². DJ5AR was to transmit CW continously, while PI9CAM checked for reflections. After some trouble with the tracking they caught it shortly before set.
The used power at DJ5AR was 150 W into a 3 m dish (28 dBi) and the PI9CAM team used their 25 m dish (48 dBi) for reception.
The predicted signal level was about -158 dBm, but, as can be heard, it´s clearly audible.
Titan 4B (26474)
Next one on the list was NORAD 26474, the 2nd stage of an american Titan 4B rocket, launched on August 17th, 2000. With a radar cross section of 15.4 m² it is a little smaller than H-2B. But the predicted reception level for Dwingeloo looked very promising anyway:
My signal could be picked up by PI9CAM when the rocket body rised up to 5 degrees over the horizon and they tracked it for about 4 minutes until I lost the object near the culmination point, when the angle velocity became too fast for my tracking system.
DJ5AR as received by PI9CAM. Picture is upside down for better matching to the table.
The signal has been detected at an predicted signal level of -160 dBm and became clearly audible soon for serveral minutes. So we can think of trying a CW QSO next time. The drift, which can be seen, has it´s cause in my transmitter, running for the duration of the whole pass without any pause. I will try to reduce the growing spread of the signal by using shorter intervals in the doppler correction.
Marten, PA3EKM, documented this (historical) moment on video:
Many thanks to the team of PI9CAM / CAMRAS: Cor, PE0SHF, Eene, PA3CEG, Marten, PA3EKM, Gert-Jan, PE1GJV, and Jan, PA3FXB
Jan, PA3FXB, and I skeduled a very special ISCAT test on 23 cm for today. The CAMRAS team had planned to operate PI9CAM with the 25 m radiotelescope in Dwingeloo for some astronomical experiments this afternoon. Before starting with that, we used an ISS pass with low elevation to try ISS bounce with the big dish. Such a pass ensures slow variation of azimuth end elevation angles, which is essential to track moving object like the ISS (or other spacecrafts in low orbits) with an antenna of a weight of 120 tons.
We had very strong reflections right from the beginning, but no decodes. Maybe the signal level was to high? This has to be investigated. Then we changed to aircraft scatter, although the dish had to be kept elevated at 10° by safety reasons, I got strong reflections from airplanes quite close to PI9CAM. Not as strong as from the ISS, but decodable now and it was possible to work in CW too.
Despite the fact, we had no QSO via ISS bounce, we learnt, that it is possible to track objects in low orbits with the 25 m radiotelescope, as long as the elevation keeps low. This opens up a perspective to make use of other spacecrafts as reflectors.
Jan, PA3FXB, and I were discussing the use of digimodes on ISS bounce for quite a while. Today we tried it with ISCAT-B and were successful with the first shot. As in the tests with Ronald, ON7FLY, on AS, we used 15 seconds periods.
From the moment on, traces could be seen in the waterfall diagram, decodes were possible.
When the ISS culmiated and the variation of the dopplershift was fastest, the frequency correction came to its limits. But while I am used to ISCAT in the meantime, I needed no decodes to hear, that Jan was transmitting RRRRs.
At least I decoded a 73 from him. The experiences with ISCAT in ISS bounce are very promising, although the 15 seconds periods are too long. So we have to discuss it and try with shorter ones.
After getting familiar in working us in CW via ISS Bounce, Jan, PA3FXB and I were looking for a new challenge. So we decided to try it in SSB. After some tests in the last days and finding a workaround for a bug in my PowerSDR, we completed a two way contact in SSB on a center frequency of 1296.300 MHz tonight. The doppler compensation worked well on both sides. Just minor corrections with the RIT were neccessary. The signals weren´t as strong as they were in the days before, but strong enough to copy even some 73s at the end. It seems to be the first SSB QSO via reflections on a spacecraft at all.
The audio recording starts with my recording of PA3FXB on one channel synchronized with Jans recording of my signal on the other channel.
Tonight I had a sked with Jan to try ISS Bounce once again. We agreed in trying SSB after getting in contact in CW. Jan appeared shortly after rise of the ISS and it was no problem to exchange reports and confirmation with strong signals. Then we switched to SSB and I could hear Jan clearly, but my tracking software refused to stay in SSB mode. So I have to fix this bug for our next test.
Later we tried to detect reflections from Iridium 4. We expected to see something about -19 dB, but nothing was to be seen in the waterfall diagram. Maybe that the choosen pass wasn´t too high in elevation and we will try again. When Jan was operating PI9CAM on February 2nd, faint traces of reflections on COSMOS 1823 could be detected by Hannes, OE5JFL, Dan, HB9Q and myself. As it is quite time-consuming, to find reasonable passes of satellites, I wrote a small Excel tool get an overview over satellites, passes and elevations. It can be found for free download here: Visibility
Macros have to be enabled in your Excel. Just enter the date, time, period and the locators of the two stations in the approriate cells, click “Calculate” in the “ADDINS/Satellite Overview” menu and the worksheet will be updated. Don´t forget to fill the TLE worksheet with the latest data of your choice.