to the Technology Task Force
Rev B-- January 8, 2003
1. Progress Toward Goals
The Software-Defined Working Group (SDRWG) has made significant progress toward our stated purposes to A) encourage SDR use, experimentation and the exchange of related ideas and information among radio amateurs; B) organize, document and disseminate interesting approaches to SDR design and implementation; C) enhance awareness among radio amateurs of what SDRs are and what they can do; and D) advocate Amateur Radio designs offering access to digital signals and algorithms. In the last year, members and associates have published at least eight software-radio articles in periodicals, a few in conference proceedings and several on the Web. Feedback has been encouraging and enlightening.
What we are discovering is that many hams are anxious to learn more about, experiment with and use software radios; but that most would like to do so without having to learn special software skills. Bob Larkin, W7PUA's SDR-10 kit evidently sold out its first run and a second run is intended. Its popularity among weak-signal fans and others is well documented on the Web. A kit of Gerald Youngblood, AC5OG's SDR-1000 is in the works. Release is scheduled for 2003. Read about a new third project below.
Since software is the key element in software radios, Group members have focused much of their attention on it. They have developed innovative algorithms for noise and interference reduction, simulated stereo reception and fast-convolution filtering, to name just a few. They have placed the algorithms in the public domain through their writings.
In addition, at least three public forums are in the planning stages for 2003: at the ARRL National Convention, Dayton and the Georgia Institute of Technology. We look forward to bringing what we know about the technology to as many as possible.
2. Our Immediate Plans
A. Standard Application Programming Interface (API)
We learned about two efforts to standardize APIs for Amateur Radio. For our purposes, an API is a bit of software that translates between a set of standard commands, controls and data and a radio-specific set.
The first is called hamlib, championed by a group found on the Web at sourceforge.net. Our contact is Chris Karpinsky, AA1VL. It is written in C++ and versions have been compiled to run on Linux systems. Several off-the-shelf ham transceivers are currently supported to various degrees. We see no great difficulty in porting it to Windows.
The second effort is that of GNURadio. Our contact there is Eric Blossom. Those folks' mission is quite similar to ours and it makes sense for us to work together where possible.
At the time of this writing, we are discussing with those parties the prospect of writing an API definition that we can publish. By virtue of his QEX article on the subject, Larry Dobranski, VA3LGD, is being included in the discussions. Finally, the SDRWG will coordinate with the Industry Advisory Council to gather suggestions and build support from equipment manufacturers and software vendors.
B. Group Software Radio Project
The group would like to embark on a new project embodying next-generation software-radio principles, including a high-speed digital interface providing access to digital samples, algorithms and configuration. To that end, we sought the assistance of Analog Devices, Inc. (ADI) with selection of state-of-the-art hardware. Jim Scarlett, KD7O and Brad Brannon of ADI were kind enough to arrange the donation of hardware to the project that gets us started on the receiver-- the toughest part.
We now have in hand evaluation modules for a AD6645 14-bit, 80 Msample/s analog-to-digital converter and an AD6634 receive signal processor, also at 80 Msample/s. With appropriate front-end filtering, this hardware is capable of multi-channel operation at up to 100 dB of spurious-free dynamic range over the sampling bandwidth. With certain ancillary circuits, that is enough to attempt a digital direct-conversion (DDC) receiver without performance compromise.
We would select either Firewire (IEEE 1394, 800 Mbits/s) or USB 2.0 (480 Mbits/s) for the interface. It may be possible to place a module for either in the final design. The actual data rate supported by the radio may be quite a bit less, depending on what we think we can handle. We will try to adapt software developed by the Group (Linrad and SDR-1000) to perform all traditional functions, such as demodulation, noise reduction, and so forth on a PC.
It would be nice to develop a graphical user interface that would allow users to perform change signal-processing configurations by drawing a block diagram. The user would be able to individually configure each block and connect them together. Eventually, an exciter (transmitter) would be added with similar capabilities. The transceiver's API would include all its functionality.
We see some issues with regard to giving configurations a sort of "sanity check" for expected results and against the capabilities of the target PC. The sanity check would be especially important for the exciter, so as not to generate illegal emissions.
As for a timetable, we would like to integrate hardware concurrently with development of the API in the next 12 months. It seems reasonable to put a working model on the table within 18 months.
C. Publishing
We are happy to report that our documentation goals will be solidly satisfied for 2003. QEX has several series to finish. Gerald's final segment is in hand and scheduled for Mar/Apr 2003, as is Jim Scarlett's Part 2; Leif's Part 2 is in the current issue and Part 3 is on the way. In addition, several other authors have indicated interest in writing more on fast-convolution filtering and the overlap-add method. We have one under review on cognitive signal processing (recognition of signal types). We learned of a DDC project out of Germany that claims excellent performance. An invitation to publish has been extended. More articles seem to be coming down the line as momentum builds.
D. IMD reduction through DSP predistortion
We would like to undertake a project to demonstrate transmitter IMD reduction through DSP predistortion. The technique is widespread in cellular-phone systems and the like, but has seen no use in Amateur Radio.
We shall begin by measuring the amplitude and phase distortion of an HF power amplifier. We will attempt to describe and implement a transfer function that partially cancels low-order IMD based on our measurements. The bandwidth necessary for such predistortion will be taken into account, as will the likelihood of deleterious effects on high-order IMD products. Results of actual experiments will be published.
E. Group expansion
Our work would be facilitated if Brad Brannon, Jim Scarlett and Eric Blossom were added to this Group. They are all doing outstanding work in software radio. We recommend the TTF forward our wish for approval.
F. Further outreach
We have certainly been aware of the SDR Forum, as mentioned in the TTF's report to the ARRL Board of January, 2001. The dues of $2000 seem a bit steep. We could try to enlist members to collaborate with us; but otherwise, we could think of better things on which to spend the money.
Our further outreach efforts shall concentrate on those who are willing to contribute their software efforts for the good of our service. We shall participate in forums and on-line discussions of future possibilities.
3. Interim Recommendations
We ask that the TTF support our efforts to conduct planned public forums. Members have expressed interest in higher visibility on the ARRL Web site than is currently enjoyed by purely technical endeavors such as ours. We acknowledge that national security places special importance on our emergency-services role, but that it also involves a significant investment in our technical roles. As an important part of the basis and purpose of Amateur Radio, we ask that experimentation and higher technical education be given equal billing with other public-service activities.
Respectfully submitted,
Leif
Åsbrink, SM5BSZ
Gary Barbour, AC4DL
Paul Rinaldo, W4RI
Doug Smith, KF6DX, Chair
Gerald Youngblood, AC5OG
References:
1. L. Åsbrink, SM5BSZ, "Linrad: New Possibilities for the Communications Experimenter, Part 1," QEX, Nov/Dec 2002; Part 2, Jan/Feb 2003.
2. B. Brannon, "Basics of Digital Receiver Design," QEX, Sep/Oct 1999.
3. L. Dobranski, VA3LGD, "The Need for Application-Programming Interfaces in Amateur Radio," QEX, Jan/Feb 1999.
4. R. Green, VK6KRG, "The Dirodyne: A New Radio Architecture?" QEX, Jul/Aug 2002.
5. M. Marcus, N3JMM, "Linux, Software Radio and the Radio Amateur," QST, Oct 2002.
6. P. Rinaldo, W4RI and J. Champa, K8OCL, "On Amateur Radio Use of IEEE 802.11b Radio Local Area Networks," Proceedings of the 21st ARRL and TAPR Digital Communications Conference, 2002.
7. J. Scarlett, KD7O, "A High-Performance Digital Transceiver Design, Part 1," QEX, Jul/Aug 2002.
8. D. Smith, KF6DX, "ARRL's Active Role in Software Radio," sidebar to "Linux, Software Radio and the Radio Amateur," QST, Oct 2002.
9. D. Smith, "Time, Space and Frequency Diversity Above 50 MHz," Proceedings of the 2002 Southeastern VHF Society Conference, Apr. 2002.
10. G. Youngblood, AC5OG, "A Software-Defined Radio for the Masses, Part 1," QEX, Jul/Aug 2002; Part 2, Sep/Oct 2002; Part 3, Nov/Dec 2002. Parts 1 and 2 are also found in Proceedings of the 21st ARRL and TAPR Digital Communications Conference, Sept. 2002.
11. G. Youngblood, "Title here ????" CQ VHF, issue information here????.