I encountered exactly this problem years ago when using a $50,000 high-end spectrum analyzer (at least it was back then - HP 8560 series) to automatically characterize a wideband free-running VCO's tuning range, tuning slope, harmonics, etc. across its entire tuning range (before the days of the fancy VCO testers and signal source analyzers). The specan has a limited number of display points, and the wider the span and resolution bandwidth (RBW) got, the less accurate the frequency readings were (the 8560 series has a real built-in freq counter).
The trick was to start with a 100 MHz span centered near the expected tune freq and a mid-range resolution bandwidth (1 MHz or so) to get a fast sweep, peak search to find the center of the hump, marker to center freq, narrow the span and RBW, repeat the peak search & centering routine, then keep narrowing and centering until the span was as narrow as a 1 Hz RBW would let me go. With the final peak in the display center, and the marker set on top, the 8560's counter was used to accurately read the frequency. This became the reference point for all the other measurements. Takes forever, but it was accurate. Why didn't I just use a stand-alone counter to make it go faster? Well, I had to do all that stuff to set the fundamental freq marker accurately in order to measure all the other parameters, anyway. This technique could be used in PowerSDR to improve frequency calibration accuracy, although I suspect it might be a wee bit faster than what I had to do since our calibration frequency is known and we shouldn't be that far off to start with. The limitation with the SDR-1000 will be LO cleanliness and stability. It's hard to use a 1Hz or 10Hz RBW with a slightly "rough" LO (spurs and temperature stability). 73, Dan KB5MY/6 DM13nc > It's definitely not just your problem. The display code changed in the > latest versions to enable the zoom/pan features that came along with the > wider display. Because of this, we have to use a more flexible manner > of converting a pixel on the display to a frequency. Unfortunately the > resolution of these pixels is very poor. Consider that when running > 96kHz on the 1x zoom, the display is showing 40kHz of data. This data > is spread over 704 pixels. This means that each pixel represents 40,000 > / 704 = 56Hz. What this means is that the accuracy of those frequency > readouts is only going to be ~56Hz in that setting. It gets worse when > you go to 192kHz and zoom out further than 1x. > > The long and short of it is that this needs to be reworked in the code > to be calculated more accurately. I'm not sure how we will do this, but > I'm sure we can improve on the current situation. > > > Eric Wachsmann > FlexRadio Systems > >> -----Original Message----- >> From: Larry W8ER [mailto:[EMAIL PROTECTED] >> Sent: Wednesday, January 17, 2007 2:20 PM >> To: FlexRadio - Eric; flexradio@flex-radio.biz >> Subject: Frequency Calibration Problem >> >> I have three versions of Power SDR available. They are PowerSDR 1.6.3 >> (with K6JCA mods), PowerSDR 1.8.0, and PowerSDR latest SVN. >> >> There are three groups of numbers below the panadapter. First is hz >> deviation of the frequency of the strongest signal in the filter >> passband. Second is that signal's strenght in dbm. Third is the >> frequency of the strongest signal in the filter passband. >> >> I start PowerSDR 1.6.3 and tune the receiver to WWV (15 mhz). Going to >> the setup/calibration tab, I put 15.000.000 in the Frequency >> calibration Window and hit calibrate. The main screen comes back and >> shows me [0.0 Hz -73dBm 15.000.000 Mhz]. This is what I would expect >> it to show. >> >> I start PowerSDR 1.8.0 and do the same thing. Now the display gives me >> [-20.2 Hz -73dBm 14.999.980 Mhz]. It appear to be telling me that WWV >> is 20 Hz low. The same is true of the latest SVN code. >> >> I have tried reloading the program. I did not import the database but >> started with a new database and fresh calibration each time. Each of >> the PowerSDR versions is in it's own directory with it's own database. >> >> I have checked the KB. I have checked the reflector and seem to find >> nothing on it so it seems to be only my problem. Any ideas??? >> >> --Larry W8ER > > > > _______________________________________________ > FlexRadio mailing list > FlexRadio@flex-radio.biz > http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz > Archive Link: http://www.mail-archive.com/flexradio%40flex-radio.biz/ > FlexRadio Homepage: http://www.flex-radio.com/ > > FlexRadio Knowledge Base: http://kb.flex-radio.com/ _______________________________________________ FlexRadio mailing list FlexRadio@flex-radio.biz http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz Archive Link: http://www.mail-archive.com/flexradio%40flex-radio.biz/ FlexRadio Homepage: http://www.flex-radio.com/ FlexRadio Knowledge Base: http://kb.flex-radio.com/
