Difference between revisions of "C-band Receiver Station"

Revision as of 23:26, 12 May 2010

This page describes the work-in-progress design of an SDR based receiver station for the 5.8 GHz band, originally created for tracking the UNITEC-1 spacecraft on its interplanetary journey to Venus.

Event logbook is available on the talk page.

UNITEC-1 is the first interplanetary spacecraft built by university students. It will be transmitting telemetry using OOSK at 1 bps in the 5.7 GHz amateur radio band using a 10W RF into a pair of patch antennas. UNITEC-1 operators need the help of the global amateur radio community for tracking their spacecraft during its journey to Venus^[1]. Antenna pointing and Doppler shift measurements will be used for estimating the interplanetary trajectory of the craft.

System Overview

Link Budget for UNITEC-1

Summary

Assumptions

1. Attenuation due to rain, ionosphere and atmospheric gasses set to 2 dB^[2]
2. TX power is 4.8 watts/antenna^[3].
3. TX antenna is microstrip patch, linear, assuming 5 dBi gain^[3].
4. Our beam width is 0.5° and I assumed a pointing error no greater than 0.1° (very optimistic) giving a 0.4 dB pointing loss.

TBDs and TBCs

1. Measure the receiver noise floor, i.e. local interference contribution to sky temperature (important)
2. Re-assess sky noise taking expected solar noise, etc. into account^[2]

Conclusions

1. An optimistic estimate suggests that we should be able to receive UNITEC-1 up to 10-15 million km distance.
2. There is plenty of margin in the beginning and we can use a low gain antenna for initial acquisition. The IKEA dish has ~25 dBi gain.

Antenna

Parabolic Dish

Wide-band Feed

Low Noise Down-converter

Kuhne Electronic had several options for 5.8 GHz LNC that covers the whole 5.65...5.85 GHz range:

• MKU LNC 57 — converting 5650...5850 MHz to 1450...1650 MHz, 1 dB NF and 50 dB gain
• MKU LNC 57-3 — converting 5600...5900 MHz to 1400...1700 MHz, 1 dB NF and 40 dB gain
• KU LNC 5659 C PRO — converting 5600...5900 MHz to 400...700 MHz, 1 dB NF and 40 dB gain

We have about 40 meters of H1000-class (TBC) cable running from the Dish to the control room and using 1.4-1.7 GHz as first IF would result in too much loss. Therefore, we chose the 5659 PRO version which has output in the 400-700 MHz range where the cable loss is limited to 6 dB (TBC).

The LNC is supplied via the coax cable. DC viltage is injected into the coax using KU BT 271 N bias-T from Kuhne. We also had something called MSTTR001 from Snec but that is believed to work up to 100 mA, while the LNC typically requires 180 mA.

Receiver

The USRP equipped with a WBX transceiver board and the TVRX receiver.

The receiver is a software defined radio and has two parts:

1. The hardware part — Converts the 400-700 MHz IF to baseband and sends it to a computer
2. The software part — Takes the baseband data from the hardware and performs filtering and demodulation in software

Receiver Hardware

The receiver hardware is based on the Universal Software Radio Peripheral (USRP) equipped with a WBX transceiver board. On the receiver side, it is a direct conversion software defined radio architecture where the RF is converted to baseband using a quadrature demodulator (ADL5387), digitized using 12 bit A/D converters (AD9862) and down-sampled using an FPGA. The resulting digital data is 16 bit signed I/Q that is sent to the host computer via USB2 interface.

The USRP Architecture

The USRP can host 2 receivers and 2 transmitters that can work at the same time sharing a total bandwidth of 8 MHz. Note that the ADCs are clocked at 64 MHz but the effective bandwidth is limited by the USB 2.0 interface to the host computer.

When we take all the protocol and other overhead away, USB 2.0 gives us 32 Mbytes/sec data rate. The USRP1 uses complex 16 bit signed integers (4 bytes/sample) and therefore we get 8 Msps. Since we use complex processing this gives a maximum effective total bandwidth of 8 MHz.

The WBX Receiver

The WBX is a full duplex transceiver board covering 50 MHz – 2.2 GHz. For this project we are only concerned about the receiver.

WBX receiver specifications
Rev	2
Frequency	50 MHz - 2.2 GHz
Noise Figure	5-6 db[4]
Sensitivy (CW)	better than -130 dBm[5]
IIP2	40-55 dBm[4]
IIP3	5-10 dBm[4]
AGC Range	70 dB[6]
Antenna	TX/RX and RX2

A block diagram of the WBX receiver is shown below. The detailed schematics are available from Ettus Research website.

• Two HMC174MS8 GaAs MMIC T/R switches are used to configure the connection between antenna connectors and receiver/transmitter. We will use the RX2 input so that we only have one switch in the loop (estimated 0.5 dB improvement).
• MGA-62563 GaAs MMIC low noise amplifier gives 22dB gain at 0.9 dB noise figure.
• HMC472LP4 is a broadband 6-bit GaAs IC digital attenuator programmable in 0.5 dB steps.
• MGA-82563 GaAs MMIC driver amplifier gives additional 13 dB gain at 2.2 dB noise figure.
• ADL5387 50 MHz to 2 GHz quadrature I/Q demodulator converts the RF to complex baseband signal.
• ADF4350 wideband synthesizer provides local oscillator signal for the I/Q demodulator.
• ADA4937-2 low distortion differential ADC driver brings the signal up to level suitable for the ADC. The ADC full scale is 2 Vpp / 50Ω differential but there is also a 20dB PGA reducing the required input level to 0.2 Vpp.

The USRP FPGA

The FPGA contains the digital down-converter that decimates the data to fit within the 8 MHz we can transfer over the USB. Actually, the decimation is variable between 8 and 256 allowing for bandwidth as low as 250 kHz (64MHz/256). The decimation factor is distributed between a four stage decimating Cascaded integrator-comb filter and a 31 tap halfband filter that decimates by 2.

USRP receiver specifications[7]
Rev	1.7?
Sample rate	64 Ms/s
Resolution	12 bits
SFDR	85 dB
Max Bandwidth	16 MHz
Host Interface	USB 2.0

Note that the FPGA design also includes a mixer and an oscillator (NCO) which allows the use of intermediate frequency input instead of baseband. This is very useful when we use an RF front end like the TVRX which outputs a 6 MHz wide spectrum centered around 5.75 MHz. Other RF boards output baseband signal centered around 0 Hz; however, the NCO is also useful for these board. The local oscillators on the RF boards have a limited resolution that does not always (read rarely) allows tuning to the exact frequency requested by the user. Using the NCO we can compensate for this difference. Fortunately, this is done automatically by the USRP and/or the GNU Radio driver and we don't have to worry about it.

For more technical details about the USRP I can recommend Exploring GNU Radio by Eric Blossom and The USRP under 1.5X Magnifying Lens! aka. USRP FAQ.

Receiver Software

To be written...

The receiver software is implemented using the GNU Radio framework.

Test Campaigns

2010.03.23

First time we powered up the LNC.

We were hoping to receive OZ7IGY beacon on 5760.930 MHz but in the OZ7SAT building we could only receive it while an airplane was passing by (reflection).

We will try again later on the roof, which will hopefully give clearer line of sight to OZ7IGY.

We could detect signal from an 5.8 GHz signal generator, but the generator was not suitable for quantitative measurements.

Detailed Report

2010.04.13

New session where we attempted reception of OZ7IGY. Tests were successful even though we only received a reflection and not the direct signal from OZ7IGY. Details are in blog post.

2010.04.24

European EME Contest on 5.8 GHz. We missed this opportunity because the antenna was not finished.

References

1. ↑ http://www.unisec.jp/unitec-1/en/cooperation.html
2. ↑ ^{2.0 2.1} Louis J. Ippoloto, Satellite Communications Systems Engineering, Wiley 2008, ISBN 978-0-470-72527-6
3. ↑ ^{3.0 3.1} Amateur Radio call for assistance for UNITEC-1 Venus-bound satellite
4. ↑ ^{4.0 4.1 4.2} WBX Receiver Performance Plots, http://code.ettus.com/redmine/ettus/documents/show/16
5. ↑ WBX receiver sensitivity in CW.
6. ↑ Transceiver Daughterboards brochure from Ettus Research.
7. ↑ USRP brochure from Ettus Research.

Blogs

• 2010.03.12: UNITEC-1: A New Deep Space Adventure
• 2010.03.16: UNITEC-1: The KU LNC 5659 C PRO has arrived
• 2010.03.17: Technical papers about UNITEC-1
• 2010.03.17: Bias-T for the UNITEC-1 Receiver
• 2010.03.21: Simple CW Receiver with GNU Radio
• 2010.03.23: Simple CW Receiver V0.3
• 2010.03.24: Successful first tests of the UNITEC-1 receiver setup
• 2010.03.24: WBX receiver sensitivity in CW
• 2010.04.01: From double side band to single side band reception
• 2010.04.06: Simple SSB Receiver in GNU Radio Companion
• 2010.04.08: Improved AGC for the simple SSB receiver
• 2010.04.09: Frequency xlating filter vs. complex multiplication
• 2010.04.09: Binaural CW Receiver with GNU Radio and USRP
• 2010.04.10: GNU Radio SSB/CW/AM/FM Receiver v0.6
• 2010.04.15: UNITEC-1 5.8 GHz Receiver Test Using OZ7IGY
• 2010.05.12: C-band Receiver Hardware for UNITEC-1

Videos

• 2010.03.15: Unpacking the LNC: YouTube or Ustream
• 2010.03.21: Simple CW Receiver with GNU Radio: YouTube
• 2010.04.05: GNU Radio Companion: Simple SSB Receiver: YouTube
• 2010.04.09: Binaural CW Receiver with GNU Radio and USRP: YouTube
• 2010.04.10: GNU Radio SSB/CW Receiver: YouTube
• 2010.04.14: 5.8 GHz Receiver Test using OZ7IGY: YouTube

In the news

• 2010.04.30: Receiver for Interplanetary Amateur Radio Satellite - The Southgate Amateur Radio Club.