Jeff, As in all technologies, there are tradeoffs. In this case the tradeoff is phase noise vs. spurs at discrete frequencies. If we had used a traditional PLL type LO, my observations indicate that we would probably have 20dB lower close spacing dynamic range due to phase noise that is present at ALL frequencies. This is opposed to spurious output of a DDS, which is present only at predictable discrete frequencies. When you are trying to achieve almost 100dB of 2KHz spacing IMD DR3, a DDS is the only way I know to do that. There are very expensive and well known PLL based rigs on the market that only get 80dB of third order dynamic range at 2KHz spacing due to phase noise limitations.
The worst case spurs on a DDS occur at integer divisors of the clock, which in our case is 200 MHz. As you go down in frequency (larger divisors) the spurs will reduce in amplitude according to the formula sin(x)/x. This means that as you go to lower frequencies the spurs become more closely spaced but also much lower in amplitude. Fortunately, atmospheric band noise covers the vast majority on the lower bands. In our case, we offset the DDS by 11.025 KHz +/- about 1.5 KHz. This number must be added to the display frequency to find the spurs. There are singular worst case spurs in only one location on 10m and on 6m respectively. They will show up around 28.582 (200/7 + .011) and 50.011 (200/4 +.011). The simplest method will be to just move the software IF down 22 KHz when we are near integer multiples. This would not involve shrinking the panadapter bandwidth. The IF frequency is arbitrarily chosen within the 48 KHz bandwidth. This is something that should be fairly easy to do so I will see that we move it up high in the software priority list. A longer term method is to map the spurs and linearly subtract them in the frequency domain. This subtraction would have NO effect on the signals of interest since power is additive in the frequency domain. The net-net is that I would rather implement software to do frequency planning at integer divisors of the clock than to sacrifice 20dB of dynamic range at ALL frequencies with a PLL type synthesizer. There is no way I know of to remove uncorrelated phase noise in software. 73, Gerald K5SDR FlexRadio Systems > -----Original Message----- > From: Jeff Anderson [mailto:[EMAIL PROTECTED] > Sent: Tuesday, December 13, 2005 10:36 AM > To: Gerald Youngblood; FlexRadio@flex-radio.biz > Subject: RE: [Flexradio] Spurs on 10 meters and 6 meters > > Gerald, I like the radio and I very much enjoyed meeting you > at Pacificon (I was talking with you when the "fire" broke > out on Sunday), but your remarks (and those of others) and > their implication that those who are concerned about the > spurs really ought not be (and the perceived denigration > implicit within that), beg for a reply. > > First, please, do not presume that the "locations of the > worst case spurs don't really matter." What may be "don't > care" locations to you might be important to me or someone > else. Fortunately, 10 meters isn't a happening-place at the > moment, but someday... > > Second, I paid in the ballpark of three thousand bucks for this radio. > Shouldn't I expect performance to be commensurate with this > amount, and, if there are short comings, haven't I paid for > the right to question these performance issues, if and when > they occur, and to wonder how they can be improved upon? > > Third, yes - one of the radio's virtues is that performance > can be improved with software updates. But if we're going to > "bring some reality" into the discussion, then I'd prefer not > to hear Marketing spin. Let's also discuss the negatives of > any proposed approaches (if negatives exist), and not just > the positives. I've been in the biz long enough to know that > software fixes to hardware problems are usually not a panacea > and often involve compromises. Are there trade-offs > associated with the proposed spur fixes? > E.g. shrinking of the panadapter display bandwidth (if > shifting the IF)? > Loss of a signal that might have been masked by a spur (if > "notch" the spurs)? > > The panadapter display is a great feature. It, and the > possibility of its width being expanded with a future s/w > rev, were the two main reasons why I purchased this radio. > And thus, to the extent that the display is anything other > than perfect (with perfection being max possible displayed > bandwidth & no spurs (neither masking signals nor in my audio > passband)), I would like to know about it. > > With that said, please recognize that, despite my grousing, I > like the radio. But, I would also like it to be better. > > Best regards from a Flex-Radio customer. > > - Jeff, WA6AHL > > P.S. While I'm at it - could I make a recommendation? I > would like to recommend that, when shipping the radio, you > include with the radio the following items (to minimize > customer grumbling and unanticipated trips to Radio Shack): > 1. A female-to-female stereo coupler. > 2. A DC Power cable (6-8 feet?). > > > > > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] Behalf Of Gerald > Youngblood > Sent: Tuesday, December 13, 2005 6:50 AM > To: 'Ross Stenberg'; FlexRadio@flex-radio.biz > Subject: Re: [Flexradio] Spurs on 10 meters and 6 meters > > > Ross, > > Thanks for bringing some reality back into the discussion. > The simple fact > is that the locations of the worst case spurs don't really > matter. Those > frequencies are rarely if ever used. Also, most spurs are > well below the atmospheric noise floor on HF so you don't > hear them in normal operation. I own a late model Japanese > radio (I paid more than the SDR-1000) that you can hear spurs > everywhere if you remove the antenna. With the antenna on, > most of them can't be heard. > > With that said, the nice thing about a software radio is that > there is more we can do to remove spurs in the software. > This we will do in future releases as we get time. Since > most spurs are well below the atmospheric noise floor, we > have not put a high priority on it yet. > > 73, > Gerald > K5SDR > FlexRadio Systems > > > -----Original Message----- > > From: [EMAIL PROTECTED] > > [mailto:[EMAIL PROTECTED] On Behalf Of Ross Stenberg > > Sent: Monday, December 12, 2005 3:34 PM > > To: FlexRadio@flex-radio.biz > > Subject: Re: [Flexradio] Spurs on 10 meters and 6 meters > > > > Sorry. but sometimes it takes a marketing person to bring the > > engineers back to reality; does it really matter in the scheme of > > normal amateur radio operation if there is a spur at > 50.011025 MHz and > > 28.582454 MHz? Please accept my sarcasm with the sincerest form of > > respect :) > > > > 73 Ross K9COX > > > > _______________________________________________ > > FlexRadio mailing list > > FlexRadio@flex-radio.biz > > http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz > > > > > _______________________________________________ > FlexRadio mailing list > FlexRadio@flex-radio.biz > http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz > >
