Flex Radio Systems SDR-1000

[WB2WIK]

Well, I see in the October issue of QST there's a pretty thorough review of the new Flex Radio Systems SDR-1000 software defined transceiver. Emphasis on, "software," as in "ever changing, never finished, don't look now here comes version 102.7."

With a modest PC and sound card to run it, plus the optional automatic antenna tuner, it now costs $3002 and still doesn't have QSK or VOX, nor is it very usable on CW even for semi-QSK operators unless you have at least a P4 CPU computer and a lot of go-fasters. Then, it sorta works CW.

I've already bumped into some SDR-1000 advocates on the air, and all I can think of is, "Are they kidding?"

Maybe next year, or the one after that.

Those interested in the review, it's on line in the "Members Only" section of the ARRL website.

WB2WIK/6

[n8yx]

Are you referring to the "Flex Radio" shown in the "Products" list of THIS site (http://www.flex-radio.com/) ?

$1375 for the 100w transceiver - no ATU included.

But that's cheap compared to the offering from DZKit - last I heard, they're expected to top the $6K mark.

[WB2WIK]

Yep, that's the one. But it's not really $1375. Check out the QST review, where they reveal what you really need, to operate the rig. They now sell an accessory Dell P4 computer with flat screen monitor, a high-end sound card and other stuff as part of the "okay, now we can claim it actually works" package.

The QST review totalled $3002, for the "slower" CPU version. An upgrade to a faster CPU, which will allow improved CW operation, costs more.

WB2WIK/6

[n8yx]

Quote[/b] (WB2WIK @ Aug. 30 2005,15:11)]Yep, that's the one. But it's not really $1375. Check out the QST review, where they reveal what you really need, to operate the rig. They now sell an accessory Dell P4 computer with flat screen monitor, a high-end sound card and other stuff as part of the "okay, now we can claim it actually works" package.
Ahhhh...

Sorta like me selling you a car with no engine, eh? http://www.qrz.com/iB_html/non-cgi/emoticons/wink.gif

I'll build my own PC to control the thing; thankyouverymuch.

The funny thing is that just today I was checking out their product line, contemplating a purchase. If an SDR is the order of the day, methinks I'll be forking out the bux for an Orion II when they become available...

[AC0H]

Although there's a huge difference in implementation, the same kind of idea gave us the Ten-Tec Pegasus, Jupiter, and Orion. Those three radios are all "software" defined. The SDR-1000 tries to put all the processing power on the PC CPU through windows whereas the Jupiter and especially the Orion use onboard processors and Flash ROM.

I've kinda been wondering what kind of performance you could get out of the Orion DSP system if it used and off the shelf Intel (P4) or AMD (A64) processor running tight machine level code.

[w5alt]

Well, I can guarantee I didn't pay anywhere near $3000 for my SDR-1000! Of course I assembled it all myself, already had the PC and power supply, and don't have the 100 watt amplfier. And it works just fine with my HP laptop computer and an external Soundblaster. (Most laptops these days don't have a stereo input for the I and Q channels.) That said, it certainly is not for everyone.

Early versions had a thermal drift problem, that I think has been solved. Also performnce is sometimes pretty bad due to the way Windows screws with the system, but normally it works fine. The performance depends more on the sound card quality than anything else, but the receiver in the thing is pretty darn good for everything from SWBC to weak CW. I particularly like listening to nice SWBC from my computer speakers while I'm working on the computer or putting it into the 22 kHz passband mode and listen to and "see" what's going on a fair portion of the band. Running the thing with 25 Hz filters or narrower for weak CW copy is amazing.

It's pretty nice if you are a fairly hard core computer geek that wants to play with real RF. I enjoy tinkering with it, but it's all software tinkering. If you're looking for a plug-and-play, forget it - it isn't, and your money would be better spent elsewhere.

But, if it's so overpriced, why doesn't anyone else come up with a nice software defined radio that can be built from a kit, has the specifications of the SDR-1000 and costs less?

WB2WIK: Steve, you've said you have experience doing that sort of thing. How much would it cost you to put something equivalent together for us computer geeky people to tinker with?

73,
Walt, W5ALT

[AK7V]

Check this out for a 26 dollar entry into the SDR world:

http://www.amqrp.org/kits/softrock40/index.html

[WB2WIK]

Quote[/b] (w5alt @ Aug. 30 2005,15:37)]WB2WIK: Steve, you've said you have experience doing that sort of thing. How much would it cost you to put something equivalent together for us computer geeky people to tinker with?
Beats me, I wouldn't even try. I have considerable experience "productizing" ideas and making real products, but it's all hardware. And for markets that are ramping in identifiable growth niches. I wouldn't define the amateur radio market this way, and I don't know that the "future" lies in software defined radios.

You commented about the SDR's receiver, but not much about its transmitter. Its transmitter characteristics leave a lot to be desired, based on everything I've heard on the air and substantiated by the QST review, who proclaimed in the nicest possible way that the rig is barely a CW transmitter, at all. A rig whose CW transmission characteristics are not only software defined, but defined by the host computer it's attached to, and that computer's specific CPU. Hmmm. Never had that problem with any other transmitter I've ever owned, and I'm not going to start having one now.

One of the "original" SDRs, to some extent, was the Kachina. A company no longer manufacturing anything for the amateur radio market, with no available service or accessories. The Pegasus followed and it was a market flop, but followed by the Jupiter, which is nearly the same rig but with a built-in user interface that greatly improved its popularity.

I bought a Jupiter, and sold the Jupiter. Not much of a rig, compared with my 27 year-old Drake TR-7 that literally runs rings around it in every measurable respect. And I indeed made the measurements.

I think I'll wait a bit longer, but unfortunately I'd be very gun-shy of any product made by a new company who might be gone a couple of years down the road. Luckily, there are those who will take that chance, so with any luck they will survive and prosper. I hope so. But I won't be an early investor.

WB2WIK/6

[n8yx]

Quote[/b] (WB2WIK @ Aug. 30 2005,16:26)]Not much of a rig, compared with my 27 year-old Drake TR-7 that literally runs rings around it in every measurable respect. And I indeed made the measurements.
This statement makes a "new" Drake fan feel good about the amount of money I've poured into my 7-line over the past 6 months...

A TR7, R7, RV7, 2x MS7, CW75, SP75, P75 and assorted other bits later, I have a station which is really enjoyable to use. No, they don't "scan da bands" like my computer-controllable rigs do...and they drift a little bit, but not much...they're not as flexible as 99% of the microprocessor/synthesizer-based rigs on the market...but they WORK. Very quiet RX (especially the R7), and excellent transmitted audio reports. As configured, the station is an excellent DX platform - and I can easily fix the rigs if need be.

A difficult act to follow.

[w5alt]

WB2WIK: Steve I haven't messed much with the TX part of the SDR. I do know that the rig sends CW very nicely and since the stuff is created in software, key shaping can be very precise.

The main problem is people generally don't like to work CW totally on a keyboard and interfacing a CW key or paddle to the computer and writing software that handles that and does all the other stuff is a challenge. In fact any digital mode should be nicely done on the SDR-1000, once the software stuff gets worked out. It's certainly come a long way from the first version and it's all pretty much open source so anyone can tinker.

Will SDR's be the future? I have no idea! I think they're neat and an ideal place to tinker with DSP and similar stuff without having to use dedicated chips, etc. I think they could handle some things very nicely.

On the one hand, I like knobs and buttons and meters and dials. I think the user interface will be a serious challenge and it may well turn out that they aren't accepted in situations where people expect a typical user interface. On the other hand, it should be possible to stick the computer in the same box with the rest and add buttons and dials, but why? I think that would only work if users accepted it and manufacturers found a nice advantage in simplifying designs and costs.

But I think we're still finding out what the problems are. Not to speak of the regulatory issues! Imagine that someone could buy a public service band radio - for example - then change the software by downloading something from the internet and end up with a SW pirate TX. The possibilities are interesting - and the potential for problems is scary!

At any rate, I do like to play with my SDR-1000 and mostly I've used it as a receiver. If the $26 version were general coverage and available or if the RX only version of the SDR-1000 were available at the time, I probably would have started with those, but they weren't. But I can promise you I would never pay $3000 for one! I probably have about 1/10th of that in it and to me the receiver is probably worth close to that much for general SWLing.

I just checked their web site again and it looks like they've stopped selling just the unassembled cards, or I missed them. Too bad.

73,
Walt, W5ALT

[WA9SVD]

I, too looked at the review in QST, and my impression, to plagiarize Saturday Night Live, "Not Ready for Prime Time..."
Certainly the concept of an SDR is interesting, but in the current incarnation, there are too many compromises to consider the radio as an only HF rig; as an expensive addition to an already equipped shack, it may have possibilities.
But the cost is another matter. The radio really DOES require a REALLY FAST computer, and an expensive sound card to properly utilize it's features. THAT has to be figured into the cost, and when all is said and done, without the computer, any SDR (not just this one) could end up as a very expensive door stop. (And there's no telling how future Operating Systems (Windoze, or Vista, or whatever comes from that company) will muck up the operation with changes in the way it addresses the computer hardware.)
I think for the time being, I'd prefer to stick to the knobs and switches, and straight key. Anybody designing a USB Vibroplex yet? Or would it have to be FireWire?

[w5alt]

Quote[/b] (wa9svd @ Aug. 31 2005,02:16)]# #But the cost is another matter. #The radio really DOES require a REALLY FAST computer, and an expensive sound card to properly utilize it's features. #THAT has to be figured into the cost, and when all is said and done, without the computer, any SDR (not just this one) could end up as a very expensive door stop. #(And there's no telling how future Operating Systems (Windoze, or Vista, or whatever comes from that company) will muck up the operation with changes in the way it addresses the computer hardware.)
It works just fine with my laptops. My old 1.2 GHz runs at about 60% load (Win ME) and my normal 2.8 GHz laptop runs at about 25% CPU load (Win XP). Interestingly, it isn't the DSP that loads it down - it's the graphics for displaying all the spectrums, etc. I expect that will improve, since other software handles that much better already. My homebrew software doesn't have that problem and runs at about 15% CPU load.

As far as sound cards, the Soundblaster MP3 external USB is fully supported and that's what I use. It cost me $39, but I didn't go around looking for the best deal - I just went to the first computer store I found and got it. It also works fine with the stock soundcard in my laptop, but since there's no stereo input, you sacrifice some DSP power because the I and Q signals are not separate.

I can't foresee that other operating systems would muck up the stuff too much. It uses a normal old everyday soundcard for all audio I/O and needs a few control lines to switch filters, set frequency, etc. Right now that's done using the standard parallel port, but there's work on switching that to USB. So as long as you can run audio software and access a USB port, things shouldn't get messed up much beyond minor adjustments. There's also a version for Linux, so that's covered, too. As long as the PC can play CD's and listen to WAVE files, they should work in the future. Of course if the underying architecture changes, then all bets are off - but no other software would work, either.

Like I said, it isn't for everyone. If you don't already have a computer capable of running it, then you're not into computers enough to even consider an SDR! These things are still experimental in my book.

73,
Walt, W5ALT

[WB2WIK]

Walt, if you haven't read the QST review (October 2005, not on the stands yet, but on line), it's very interesting. You haven't even begun to find the "problems" with the SDR because you haven't been exercising it the way the League's lab did, especially with regard to transmitting. They found only very high-end cound cards actually support full operation (including transmitting), and Flex Radio is now selling "compatible" sound cards just to make sure they work when the user gets a system. To take full advantage of the SDR1000 features, it takes two sound cards, so there goes two PCI/PCIx slots (oh, and another $139). If you don't have an extra slot as a spare, you might need a new computer, or at least a new mother board.

The automotive industry, one of the most mature industries in the world, went through this process already in much larger numbers to determine the User Interface is a powerful selling tool and drives consumer acceptance heavily. There's absolutely zero reason we need things like "speedometers," "fuel gauges" and so forth. It would be cheaper to make these fully digital, or even have a voice annunciator tell us what's going on. The electronics are much cheaper, and more accurate, than stuff with "moving parts."

But they tried that, and consumer acceptance was low.

So, they went back to "gauges." It's what works, and it's what's acceptable. Also, interestingly, as the wristwatch and clock industry also found many years ago, a quick glance at an analog indicator often reveals more meaningful data than a study of a digital one. Analog instruments show an immediate trend, whereas digital ones don't. Rolex still sells $25,000 analog watches -- that have mainsprings and no electronics inside. They outsell all the other high-end electronic watch companies, in sheer quantities.

WB2WIK/6

[w5alt]

Steve, I agree with you 100%. As I mentioned I haven't done much transmitting, just a couple of contacts using the bare rig at about 1/2 watt. The RX part is actually what intrigues me more than anything else - I've found that computer generating stuff to TX is pretty straight forward - receiving and processing is the tougher problem, especially on HF.

To me, it's a neat toy and still has a ways to go. But someone has to start somewhere. If I had to pay $3000, I wouldn't have started at all! And I agree about the user interface, too - when I want to get on the air and enjoy myself chatting or working contests or DX, I use my other rigs - the ones with the big tuning knobs, meters, buttons and dials.

I find it interesting to note that the limitations and problems that have been noted generally fall into the category of user interface - which is very important - or into the need for extremely critical hardware - which isn't directly a software problem. As always, it looks like there's a tradeoff - you can't skimp on the hardware just because it a "software defined" radio. But who knows? Maybe someone will figure out how to make the software compensate.

Just a#funny sort of observation, too. People complain that hams aren't keeping up with technology, but here's a case where (as far as I know) there is no equivalent product available on the commercial market and people complain that the ham version isn't perfect already.

73,
Walt, W5ALT

[WA9SVD]

Walt (W5ALT:)

One of the problems with the SDR IS the software, the control software, that is, as well as the operating system.
And as you say, it uses the Parallel Port for some control functions.
But Microsoft (since the control software is apparently designed for Windoze) has declared the Parallel Port to be obsolete, or so says it's Hardware Design Guides. The parallel port (unfortunately) is no longer supported to much extent in WinXP, and will probably disappear in another year or two, when computers need the "compatible with MS Vista" (aka "Longhorn.") sticker. Just like the COM (RS-232) ports on many newer computers. (Or rather, lack thereof.)
As it stands now, it's impossible to find a new model printer with a parallel port; Mr. Gates wants ALL peripherals to be USB. But the problem is HOW LONG will the company support the product, or will it become an expensive paperweight if the company decides to either bow out of the market, or just not to support an "older" piece of equipment?
At least with "knobs and switches" we still stand a chance of using our equipment long after the company seeks more profitable pastures.
Innovation is great, and there MAY be a niche for SDR's, but the rapid obsolescence (similar to the obsolescence rate of computerrs!) makes many of us leery about making such a large investment for a "black box."

[w5alt]

Hi, Larry,

Well, if you have already decided to be tied to Microsoft for everything, then there's not much I can say. But USB is known and documented technology, so building a separate interface isn't a real big deal. In fact the $26 kit from AmQRP has a USB interface.

What would I do with my SDR-1000 if my computers in the future didn't have a way to use the parallel port? Simple, rip out the interface section and rebuild it to work with whatever I happen to need. Heck, I could even mount toggle switches on it and control it manually - now that would get some laughs! Hams historically haven't been afraid to dig into their rigs and I'm certainly not afraid yet. (More cautious with age, but not afraid.) Or maybe I'd toss it into the pile of TNC's, 300/1200/2400 baud modems, 8" and 5 1/4" disks and drives, handheld scanners, digitizing wands, color dot matrix printers, and other old stuff I tried and used and now no longer want to hassle with. Lord, I wonder how many dollars ended up in that pile!

Perhaps the problem is Gerald or QST trying to market the SDR-1000 as a final product to what appears to be mainly an "appliance operator" amateur service. Early amateur work trying to get SSB working was not necessarily easy or cheap - and there were already commercial applications to emulate. Neither was early digital stuff, but some hams made do. #But if you want a company to stand behind the entire system and solve problems that are mainly due to peripheral things like sound cards and graphics, then you pay for the equipment they know will absolutely work. And if you get the boards and aren't afraid to tinker, you can certainly put one together for much, much less - I know because I did.

Whether FlexRadio will be around in the future or not depends to a large extent on whether hams want to play and are willing to foot the bill. Maybe things will change in the future - and maybe they won't - but Gerald's market isn't appliance operators - yet.

I figure that I've had fun and learned a lot with my SDR-1000, so it was worth my investment. Some day I'll tell my great-grandkids (I already have grandkids!) either that "I was there and did that" or "forget it, we tried and it won't work." I don't know which it will be and it really doesn't concern me in the least.

73,
Walt, W5ALT

[WB2WIK]

I like W5ALT's attitude, it's actually just the right one for experimenting with new stuff like this.

One problem with software defined anything, for me, is that I'm not a software guy, don't do any programming, and was never interested in this aspect of technology other than as a "user" benefitting from the development work of others. I can play with hardware forever and understand everything I'm doing, but I can only install and use software. It's scary.

I have some specialized test programs which never work right, and the result is the hardware doesn't work right, either. So, one of our engineers here who's been developing software for years looked at the source code and said, "Oh, geesh, there's a bug in the control algorithm that makes a conflict with the xyz franistan, and that pushes the plunger into...yada yada <snore>." I have no idea what he's talking about. It doesn't work, and that part I do understand.

Most of our shippable systems have software (actually, firmware) revisions every other week. Sometimes more often. Can hardly keep track of it. Is this version 48.1, or 48.2? What was the difference, again? Is it tested for complete backwards compatibility? Should I bother shipping it, or just wait for 48.3, since I'm sure that'll be along any moment, now...

WB2WIK/6 http://www.qrz.com/iB_html/non-cgi/emoticons/biggrin.gif

[w5alt]

Steve, we sound almost like opposites! I love tinkering, but readily admit I cannot think in terms of hardware, other than large-scale ideas. I've built plenty of things over the years and figured out to repair my rigs - often after I, myslef, let the smoke out. I've still got a lot to learn and I don't know if I ever will - but it's fun anyway.

Software and math, now that's easy and I find it a major challenge to do more with very little hardware. What we need is a way to combine interests. Software without hardware is nothing more than a useless intellectual exercise.

So if someone can figure out how to design and build the hardware for a better, cheaper, more reliable SDR, I'd have a blast trying to program it. Maybe it would work well - and maybe it wouldn't - but if all involved are having fun and learning, who cares? The next version could be better.

By the way, I'm trying to find out about ordering the 40m SDR kit to play with. With luck, I can pick it up next trip home and then that'll keep me occupied and tinkering a little while longer. I already am wondering if it can be easily modified to cover more band for 30m and SWLing on 25 - 49m. And where can we steal a signal to whip together a TX? Maybe I can and maybe I can't.

One can never have too many radios, even if it eventually ends up in my pile of retired equipment next to the 1200 baud modem that I paid $200 for way back when they were a luxury. #http://www.qrz.com/iB_html/non-cgi/emoticons/biggrin.gif

73,
Walt, W5ALT

[KF0RT]

Quote[/b] (WB2WIK @ Aug. 31 2005,15:49)]Most of our shippable systems have software (actually, firmware) revisions every other week. Sometimes more often. Can hardly keep track of it. Is this version 48.1, or 48.2? What was the difference, again? Is it tested for complete backwards compatibility? Should I bother shipping it, or just wait for 48.3, since I'm sure that'll be along any moment, now...
Laughing... I'm the guy who puts out those firmware releases. Every other week? Why, you guys are slackers (or more likely, are better organized).

I do all the code that touches the hardware, and do some of the hardware, too. Debug tool is an old HP logic analyzer that decodes assembly in real-time while the machine is running. I've done 80 "releases" since January, but most of these are for internal lab use only, and some are just to test an idea. On the other hand, I wrote both of the boot ROMs in that machine (it boots via LAN), and neither has been touched in over three years now.

Embedded programming can be a real hoot. I've taken over for an old friend who retired a few years ago (the people hired to replace him didn't work out). He's a model train buff and is busy automating his train stuff now, using CAN (Control Area Network) stuff.

You can't do the embedded stuff without a good grasp of the hardware. My problem is that my brain is always in digital mode. I can (and have) designed computers down to the component level, but I'm still weak on anything analog. It just doesn't "click" here. If I could get my Extra by passing a 20 WPM code test, I'd have it by now. CW is sooooo digital. http://www.qrz.com/iB_html/non-cgi/emoticons/smile.gif

73, Rob

[WA9SVD]

Walt (W5ALT):

I'm all for experimenting, and as you seem to be a software expert, I say more power to you! But not all of us as gifted in the programming dept.; I'm MUCH more comfortable with the hardware. But the two ARE tied together intimately.
And I'm not too fond of the boys from Redmond, nor do I like their strangle hold on the Operating System market, but that's a current reality. And when Micros*** publishes design requirements that manufacturers have to follow to get that little sticker that says "Designed for Micros*** Windoze," most manufacturers consider not following MS's guidelines and forgoing the sticker the "Kiss of Death." Because most purchasers look for that sticker (right or wrong.)

I''d love to have a SDR to use, play with, experiment with, hack into, etc. but I wouldn't want it to be my only radio.
And like you, I am more cautious about hacking into a radio (especially an expensive one) as the years go by.

[WB2WIK]

Yeah, I'm more like Larry than Walt.

My way of improving radio performance:

"I can't hear this guy."

"No problem, we need more aluminum. If that doesn't work, make the tower higher."

Problem solved.

Software?

WB2WIK/6