I was reading amateur radio blogs and for the third time or so in a month, I was treated to what amounted to a diatribe against software defined radio. When I read these, I can’t help but sigh in frustration.
I understand nostalgia. We all like the cars we wish we could have had when we were teenagers. We like the music that we learned to like as young people. Heck, I still have the first computer that I bought at age 14. In moderation, nostalgia can be a good thing. It reminds us where we came from. It gives us context. It gives us history.
But I think we always have to temper a sense of nostalgia with perspective and vision. Our first cars were our loves, but they were (let’s face it) often gas guzzling, unreliable death traps. Some of our music was probably brilliant, but a lot of it was (let’s face it) crap. And my first computer, while enabling me to explore a world of computation, which eventually led a rich and rewarding career, was by any modern standards less powerful than an alarm clock.
Which brings me to software-defined radio.
I’ve seen lots of people write negatively about software defined radio. Often it’s just pure nostalgia. “Real radios glow.” “Real radios have KNOBS.” “Real radios don’t require your laptop.” “My old radios just sound better.”
I think we as hams should be more visionary than these statements. Software-defined radio is enabled by the remarkable evolution of computing hardware. The speed of computation has in the time that I’ve owned computers increased by a factor of about 200,000. The cycle time of most desktop computers is now low enough that an inexpensive computer can execute dozens of instructions for every cycle of an HF signal. This enables the traditional features of radios like mixing and filtering to be done in software, instead of being cast in hardware. This means that we can have unprecedented control over these processes. And, potentially we can even change many aspects of operation even after the soldering of our radio is complete. Software isn’t a panacea, but it is the source of such great power it can’t and should not be ignored.
Most of the criticisms I see about radio interface seem curiously misplaced. I agree that using your laptop as an interface is often sub-optimal. But if you’ve played with any radios at all, I bet you had some criticism about the conventional radios that you had as well. Modern radios are complex, and this complexity is not often controlled by careful user interface design, whether in software or in conventional hardware front panels. Even if you thought that some radio you had in the past was perfect, I would submit that it would still be beneficial to use software defined radio technology inside.
But even beyond that, I have been playing with an SDR-IQ for the last couple of months. Its ability to monitor all signals on an HF band is enormously compelling. Using digital displays, we can give the user a perspective on the entire band which is undeniably useful. Technology like CW Skimmer can track dozens of Morse signals in real time. Programs like HRD and fldigi can do the same with multiple PSK signals.
I’ve seen a few people abandon the idea of software-defined radio simply because they aren’t adept at software design or have no knowledge of the mathematics and algorithms that underly software-defined radio. They fatalistically claim that they will never or can never understand or learn how software-defined radios work. I am kind of flabbergasted by this attitude. It’s not that it is easy: far from it. Studies show that to become expert in a subject takes the better part of a decade. But as hams, we are supposed to be about self-training and experimentation. Sure, there is no mandate for you to do so, but let’s not be so fatalistic and condemn it merely because it doesn’t coincide with our own abilities or interests.
There is no reason to forget the past, but let’s not worship it either. Let’s look at the best of what was, the best of what we have now, and the best that we can imagine, and experiment, design, build and use the radios that we have yet to invent.
The popularity of WSPR shows that despite our fascination with old technology, hams are still pushing the envelope of the new.
Software defined radio is the future. I think our phones are already using this technology.
The thing I think we should fear is that the age of home brew may come to an end when the only components are surface mount and too small for us to work with.
Nice post.
As one who designs software defined radios professionally (in FPGAs), I am always tempted to get my amateur radio license and do some new digital packet radio stuff. However, each time I look into it, I get disappointed by the unwilling community, the antiquated rules on usage, and the lack of interest. The unlicensed bands, even with their restrictions, are more attractive.
This wasn’t meant to be as negative as it sounds, now that I read over it. I just feel that a few small changes in different areas could help out ham radio quite a bit.
While the ‘innards’ of new equipment may all be based in software and onboard processing power, this doesn’t have to be apparent to the user. What’s wrong with an amateur transceiver just being another electronic gadget?
I don’t look at my toaster, for instance, and think to myself, I need to hack the software in that to make it a better toaster. I simply plug it in and use it to make toast.
The same goes for my DvD player, my television, the list goes on.
It is simply someone’s opinion that a ham transceiver is made better by connecting it to a personal computer and applying a lot more software. The person who believes this rarely takes into consideration the diminishing returns of having to use a specific hardware platform with a specific operating system to achieve optimal results.
Some day, when this is truly an ‘open’ field of endeavor, things might get a little more interesting. Until then, SDR in the world of amateur radio is a one-trick pony named Flex-Radio.
Not saying that’s bad, I’m just saying…
This is precisely the kind of comment that made me comment about “lack of vision”. It’s certainly true that the software component of an SDR need be no more noticeable than the software components that live inside your DVD player or your cell phone or your wrist watch. In fact, most high end and even moderate end amateur radios already have a very significant software component inside them. This usually takes the form of DSP units to provide filtering, or modems to supply modes like RTTY or packet radio. To most users, the fact that there is software inside the box is something that they choose to mostly ignore.
In software defined radios, software is responsible for additional functionality such as mixing or modulation. You could choose to ignore that of course. If the functionality that you got is equivalent to a radio you already use and like, and presents the same interface, maybe that’s cool. But to me, software represents an upgradeable component that is inexpensive to modify (no hardware changes) and even better, can be redistributed cheaply (bits are cheap to mail).
As a practical example, consider making a phasing SSB exciter. Traditionally, what you would do is take your audio input, and pass it through some carefully constructed all pass phasing networks. These would then be mixed with the carrier frequency in quadrature and then either added or subtracted to generate the appropriate upper or lower sideband. Depending on the accuracy of the phasing network, you get good or not so good sideband cancellation.
But, let’s imagine something different: instead of using an all pass network, let’s compute the phase shift using a Hilbert transform. A simple dsPIC costs a few dollars, and has no difficulty whatsoever computing a Hilbert transform at audio signal rates in real time, and takes just a few ma of current. In addition, you can fold in all sorts of other functionality (filtering, compression, whatever) at no additional cost beyond just writing the software.
Once you begin integrating microprocessors into radios, then you can begin to do all sorts of fun stuff. Integrate SWR and power meters. Good battery indicators. Frequency counters. Time synchronization via GPS. Satellite tracking and Doppler tuning. Even provide streaming audio from the radio via Ethernet. The addition of each of these features is very cheap, because compute power is just about free inside of radios.
Having compute power also means that we can make better displays. Why does the average HT have such a crappy display compared to even the cheapest of cell-phones? If you surf over to sparkfun you can find nice graphic lcds for about $20. Instead of wiring in a bunch of 7 segment LEDS, why not create a generic interface that can display all kinds of information (I am always baffled why HT’s don’t include a time display that shows both local and UTC time).
Flex-Radio isn’t the only game in town either. At the low end, we see things like the Softrock kits. Moving up the ladder we see things like Genesis Radio. GNU radio. HPSDR from TAPR. RFSpace’s SDR-IQ, SDR-14 and SDR-IP.
Let’s not be so curmudgeonly that we draw a line between radios and computing. One enhances the other. I’m teaching myself RF design, I wonder why so few statesmen for ham radio work on teaching themself DSP and software design.
Necessity is the mother of invention and apparently, we have little need of more invention. If tomorrow the FCC said that our bandwidth on all modes had to be cut in half by a given date, we would no doubt innovate as required.
Given that we are under no such “gun”, radio is a hobby and many of us work on a PC all day and look forward to a break from it when we come home at night.
I don’t think it non-visionary to call a spade a spade. I think that root canals have served mankind well but I don’t want to celebrate them! 🙂
Other than the Softrock kits, I wasn’t aware that any of the others you mentioned included transceivers — I thought they were receive only.
I think the best place for the computer and software is embedded deeply inside the rig where I don’t have to deal with it. I like my computer for computer things and my radio for radio things and see no reason why the two should be shared.
But I do understand that desire makes me a curmudgeon and out of the mainstream of current thought. No matter, I also believe that those who believe in techno-triumphalism will eventually come around to my way of thinking. 🙂
73, Jeff
Hi All,
I agree with the post. I got into Ham Radio as a teenager in the early 70’s. At that time the inexpensive equipment was WW II surplus and used 1950’s rigs. My first station was an 40 Meter ARC5 and Hamarlund HQ-180 (#?).
I ended up in a Boy Scout Explorer Post sponsored by a local laboratory. In a town of 50 K, there was only 3 or 4 of us in this post at any one time. Our scoutmaster’s RTTY station was probably what resulted in my interest in Packet Radio many years later.
The best freindship I made then was with a guy who became a nostalgia radio afficianado. His livining room is made up to be identical to a 1940’s parlor, complete with restorted radios.
I could never understand this, myself.
My interest in ham radio has always been the technical and future technical aspects. And as with the other posters, I have run into the status-quo (youy know guys AM TV sucks, and Hams should have been on FM a long time ago) limitiations. Specifically with regards to Packet Radio transport network in the late 1980’s and early 90’s. This was so discouraging, I dropped out of general Ham Radio, and only used it for Internet access (9600 baud) for many years, until Internet access became generally available – thanks Al Gore.
Hi All,
SDR will have its unique applications. Specifically, I think digital HF voice may be the application that results in SDR being uniquely capable of outperforming older technologies.
I am listening to Radio Canada International, DRM (Digital Radio Mondiale) as I type this. DRM works well. Some people criticise it for not being generally DX able, but this is more a matter of the sub mode the broadcaster chooses. In its most robust sub-mode, DRM can easily outperform AM in the very high multipath HF DX paths. I have seen this many times in practice.
The signal is 10 KHz wide, watching the highly variable multipath signals, one sees the fading propagating through the signal in a very random fashion, rather quickly Yet the decode in such sitation at slower data rates can be very good. In such situations, the phase distortion on a AM signal would be horendous, and when the fade knocks down the carrier, envelope demodulation would cease.
So, my guess is that the nitch for digital voice will be these grey-line paths. A 10 KHz wide digital signal with random disposition of the voice data throughout the 10 KHz, will outperform SSB.
The problem with this is how can 3 or 4 voice channels be randomly multiplexed into the 10 KHz, to retain the same spectral efficiency as SSB? If it was 4 guys sending from the same area, but out of ground wave range, to 4 other guys in the same general area, bout out of ground wave range this could be done easily by interleaving carriers thoughout the 10 KHz spectrum. But, when there is a general mix of stations locations it becomes more difficult.