Difference between revisions of "Receiving LRO and LCROSS"
|Line 423:||Line 423:|
== System Tests ==
== System Tests ==
== OTA Results ==
== OTA Results ==
Revision as of 14:19, 16 September 2009
This page documents my attempts to receive the S-band downlink of LRO and LCROSS using mostly what I already have in stock. This includes using the USRP and GNU Radio as software receiver. It is important to note that we are only talking about detecting the presence of the signal from the spacecraft and not decoding the signal. Decoding would require a much higher G/T than what can be achieved with small antennas.
- 1 System Overview
- 2 Link Budget
- 3 Detailed Component Descriptions
- 4 Wiring
- 5 Detecting the Signal
- 6 System Tests
- 7 OTA Results
- 8 References
|Budget / BOM|
|60cm Dish + patch feed||165 €|
|Photo rack||60 €|
|Various mounting hardware||30 €|
|HEMT LNA||250 €|
|Coax cables & connectors||50 €|
|USRP v1||500 €|
Complete link budget is available here: File:AMSAT-IARU Link Model Rev2.5.2-LRO-1.ods.
A summary and some notes can be found below.
|Frequency||2271.2 MHz||2248.5 MHz|
|TX power||37 dBm (5W)||38 dBm (7W)|
|Antenna gain||0 dBi||22 dBi||0 dBi||10 dBi|
|EIRP||37 dBm||59 dBm||38 dBm||48 dBm|
|Free space loss||-211 dB|
|Atmospheric, pointing and polarisation losses||-5 ... -1 dB|
|Signal at receiver antenna||-177 dBm||-155 dBm||-176 dBm||-166 dBm|
|Receiver G/T||2.6 dB/K|
|S/N0||22 dBHz||44 dBHz||23 dBHz||33 dBHz|
|SNR @ 1 MHz BW||-37 dB||-15 dB||-35 dB||-15 dB|
|SNR @ 500 Hz BW||-4 dB||18 dB||-2 dB||17 dB|
- For the S-band, LRO has omni (assumed 0 dBi) and a HGA. The HGA is shared with the Ka-band transmitter and is a 75 cm dish → 22 dBi gain assumed.
- LRO transmission modes for S-band downlink: 3M40G2D, 4M57G2D, 5M00G1D.
- LRO transmission modes for Ka-band downlink on 25.650 GHz: 57M25G1D, 114M50G1D, 229M00G1D.
- LRO downlink data volume is 461 Gb per day @ 100 Mbps
- LCROSS transmission modes for S-band downlink: 3M41G2D, 5M00G1D
- LCROSS Telemetry: Spacecraft communications are provided through two medium gain antennas operating at 1.5 Mbps (nominal), two omnidirectional antennas operating at 40 Kbps (nominal), and a 7-watt S-band radio frequency transponder. 
- SNR @ 1 MHz is for USRP+DBSRX, which has a programmable channel filter that can be as narrow as 1 MHz.
- SNR @ 100 Hz would be for a standard ham radio receiver in narrow CW mode.
System Noise Temperature
I do not have any inline devices between the antenna feed and the LNA. In fact, I will mount the LNA directly on the antenna feed via an adaptor that has a loss of ~0.1 dB.
The LNA is a HEMT-based amplifier with gain 30 dB and typical noise figure 0.5 dB corresponding to an LNA temperature of 35.4 K.
From the LNA I will have a few meters of low loss coax running to the receiver. Between the receiver and the cable coming from the LNA there is a bias-T I will use for injecting DC voltage into the coax to feed the LNA. This has an insertion loss of 1 dB.
The tuner is a wideband receiver covering 800 MHz to 2.4 GHz and has a typical noise figure between 3-5 dB. I used 4 dB which is roughly 400 K.
The total loss between LNA and receiver, including all connectors and bias-T is estimated to 3.2 dB.
The sky temperature is the noise coming from the sky. Empty regions of the sky have a temperature of 2.73 K (the cosmic background radiation), but since I will be pointing my antenna at the moon I will have to include that too.
I will use a 60 cm dish, which has a field of view (FOV) of 15.4°. The size of the Moon is ~0.52° meaning that it covers 3.4% of the antenna FOV. I will assume 250 K for the part of the sky covered by the Moon and 2.73 K for the rest:
T(sky) = 0.034 * 250 + .966 * 2.73 = 11.14K
Of course, this is just an approximation and proper calculation would use integral calculus and take the antenna pattern and temperature variations over the sky into account. For this experiment, the above approximation is good enough. Also note that I assumed that the Sun is not in the field of view of the antenna. If it was, the performance would be degraded by ~2.5 dB (measured value).
The result above is not complete - we should add about 12 K for noise coming from the atmosphere :
T(sky) = 23 K
An additional contribution to the sky temperature that we have to take into account is the terrestrial noise coming into the system, for example via the side lobes of the antenna. I decided to leave that contribution at 0 K for now because I do not expect any interference at the receiving site.
A system noise temperature of 65 K is excellent for amateur equipment. Interestingly, the assembly instructions of the antenna mention that when using a good LNA one can expect a system noise temperature around 90 K that includes about 25 dB "spillover" that I didn't take into account. The 25 K difference is not much considering that the system noise temperature can quickly increase by several hundred degrees if using a lossy coax cable or bad connectors.
Detailed Component Descriptions
To be written.
The antenna consists of a 60 cm dish equipped with a patch feed. It is the S-band antenna system by James Miller. The dish and patch have been specially manufactured for satellite operations and has better performance and e.g. standard satellite TV dishes in term of side-lobes. The gain of the antenna is 21 dBic. Comparing with the formula for a parabolic dish, we can see that the aperture efficiency is around 60%, which is very good.
|-10 dB beamwidth||28°|
|Connector||N-male [Option: N-female]|
|Overall diameter||590 mm (23")|
|Weight||1.4 kg (3 lb)|
The dish has 590 mm diameter, 119 mm deep, 1.2 mm thick (18swg). This gives an f/d ratio of 0.31, and is virtually identical to the dish described in Oscar News issue No. 100, Amsat Journal, Amsat-DL Journal, and many others as A 60 cm S Band Dish Antenna. The dish can be ordered from G3RUH. Package deals include the RHCP/LHCP patch feed and mounting kit.
The maximum usable frequency of the dish is TBD.
The feed can be used as an antenna on its own having a gain of 8.5 dBic. It is suitable for illuminating a dish with f/d ration between 0.3 and 0.5, therefore it works very well as a feed for the 60 cm dish. The feed polarization is LHCP (RHCP as option), so that the antenna becomes RHCP when the patch is mounted as a feed for a dish.
|Frequency||2250-2450 MHz [Option 2150-2350 MHz]|
|-10 dB beamwidth||125°|
|Feed polarisation||LHCP [Option: RHCP]|
|Suitable dish f/d||0.3 to 0.5|
|Connector||N-male [Option: N-female]|
|Overall diameter||120 mm|
|Depth||17 mm excl. connector|
Note that the gain of the patch feed can not be added directly to the gain of the dish. The gain of the dish is determined by its size. The efficiency of the feed – in particular how it illuminates the dish – has influence on the antenna efficiency, which of course has influence on the effective gain of the antenna.
Mast and mount kit
The antenna is not heavy so it can be mounted on a photo tripod.
To be written.
The local hardware store didn't exactly have antenna mount kits so I had to improvise with what I could find.
The plate shown on the photo turned out to be made of something that I could not work with (steel?). Eventually I got an aluminum plate instead.
Low Noise Amplifier
Kuhne KU LNA 222 AH HEMT super low noise amplifier is mounted close to the antenna feed and is used to improve the receiver performance by increasing the figure of merit (G/T).
Expected performance gain TBD.
|Type||KU LNA 222 AH-HEMT|
|Frequency range||2200 ... 2400 MHz|
|Noise figure @ 18 °C||typ. 0.5 dB @ 2250 MHz|
|Gain||typ. 30 dB|
|Output IP3||typ. 27 dBm|
|Supply voltage||+9 ... 15 V DC|
|Current consumption||typ. 80 mA|
|Input connector||SMA-female, 50 ohms|
|Output connector||SMA-female, 50 ohms|
|Dimensions (mm)||73 x 30 x 20|
The LNA will be mounted directly onto the dish feed.
|Type||KU BT 271 N|
|Frequency range||10 ... 3000 MHz|
typ. 0.1 dB @ 150 MHz
|Voltage range||0 ... +15 V DC|
|Current||max. 1 A|
|DC connector||DC socket 2.1 mm|
|Input connector (DC output)||N-female, 50 ohms|
|Output connector||N-female, 50 ohms|
|Dimensions (mm)||37 x 37 x 30|
RF Front-end (tuner)
This component is the RF daughter-board that plugs onto the USRP. It converts the high frequency RF signal to I/Q baseband that is is passed to the USRP ADCs. The options for this include the RFX2400 and DBSRX.
Initially, this option was considered; however, since the RFX only covers 2.3 to 2.9 GHz it is not suitable for this experiment. Even if it was possible to go down to 2.25 GHz, the noise figure of this receiver is worse than the DBSRX.
The DBSRX is a 800 MHz to 2.4 GHz receiver with a 3-5 dB noise figure and a software controllable channel filter that can be programmed between 1 MHz and 60 MHz.
It contains an MGA82563 wide band LNA followed by a MAX2118 DBS direct conversion tuner chip, followed by an AD818x (TBC) VGA. Note that according to the MAX2118 data sheet, the tuner is specified to work in the 850-2175 MHz range.
- DBSRX block diagram (incl gains, AGC, dyn range)
- DBSRX schematics
- DBSRX PCB
- MGA82563 data sheet
- MAX2118 data sheet
- AD818x data sheet
Also see DBSRX Reference.
The software receiver is implemented using GNU Radio. Since we only want to detect the signal (but not decode), the signal processing can be very sinmple and consist of some basic filtering, down-sampling and display of the baseband data coming from the USRP.
Coax cables, connectors and adapters were purchased from Wimo. DC and USB cables were available from stock. Following wiring is needed:
- N-female ↔ SMA-male adaptor for mounting LNA on the patch feed
- SMA-male ↔ 2m AIRCELL 5 ↔ N-male for connecting LNA and bias-T
- N-male ↔ 0.2m AIRCELL 5 ↔ SMA-male for connecting bias-T to USRP
- Alternatively, N-male ↔ SMA-male adapter
A simplified wiring is being considered, where the bias-T is replaced with a 9V rechargeable battery (life time 2-3 hours).
Detecting the Signal
Without any detector, we can only observe the spectrum of the passband. How do we know that we are actually receiving the spacecraft and not just some background noise?
There are two possibilities:
- Indirect method 1 — Point the antenna towards the moon; we should see increased noise level in the passband. to check whether it is satellite signal or moon-noise, tune to a frequency where the satellite is not transmitting.
- Direct method — We can try to demodulate (maybe even decode) one of the sub-carriers.
- For LRO we can simply listen as the crafts orbits the Moon and compare received signal with http://lroc.sese.asu.edu/whereislro/ – The signal should disappear when LRO is behind the Moon.
Will it be transmitting while over Europe? Maybe...
These tests were carried out on 2009/09/15 using calibrated instruments. No DSP software was prepared for this test. Only a simple FFT spectrum analyser and waterfall displays were used.
- LNA specs confirmed: G = 29-31 dB, NF = 0.5 dB 2.2 - 2.3 GHz. We found out that the LNA has no built-in bandpass filter so the useful range is much wider. Gain stays above 28 dB between 700 MHz and 2.4 GHz but with increased noise figure ~ 1.0 - 1.5 dB.
- USRP+DBSRX @ 2.25 GHz, 250 kHz spectrum, FFT scope avg. α = 0.07 could detect a -132 dBm signal (~ 5 dB SNR). Inserting LNA increased this to -138 dBm.
2009/09/22 – TBC
To be written.
- NASA Solicitation: Lunar Reconnaissance Orbiter Mission Communication System, see Talk Page
- K-Band TWTA for the NASA Lunar Reconnaissance Orbiter, http://gltrs.grc.nasa.gov/Citations.aspx?id=3385
- NASA Long Range EM Spectrum Forecast, October 2007
- Ippolito, Radiowave Propagation in Satellite Communications, Figure 7-7
- Datasheet for the BasicRX, BasicTX, LFRX, LFTX, TVRX, and DBSRX daughterboards http://www.ettus.com/download