N4HY would probably be the a good expert to answer the questions in your PS. Let me clear up a few things from the body. The IF we use is 11,025Hz. However, I'm fairly certain the extra 600Hz can be attributed to being in CW mode where the DDS is tuned "CW Pitch" amount up or down from the VFO depending on whether you're in CWL or CWU.
I would have to put debug in the console to see what the actual tuning words are that are being sent, but I am confident that the spur reduction technique works well. Note that the algorithm requires clearing some low order bits and setting at least one bit. This could be why you don't get exactly 0x400000000000. I don't remember where we got the description for the spur reduction, but I'm sure either K5SDR or N4HY could help us dig it up for your reference. Now, as stated before, there are multiple kinds of spurs, and you have just found the worst lot of them. This is because tuning to 50MHz on the DDS (VFO = 50.011025 in SSB -- 50.011625 in CWU) is exactly 1/4 the 200MHz oscillator and mixing products result. I'll let the experts explain why this is the case. :) Eric Wachsmann FlexRadio Systems > -----Original Message----- > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] > radio.biz] On Behalf Of Jeff Anderson > Sent: Wednesday, December 07, 2005 1:07 PM > To: flexradio@flex-radio.biz > Subject: Re: [Flexradio] Interesting behavior when connected to a dummy > load > > I'm seeing something strange which has me puzzled... > > If, when on 6 meters, I turn Spur Reduction OFF (and set my Clock Offset > to > '0'), I see a picket-fence of spurs whose spacing becomes smaller and > smaller as I tune until, at 50.011625 MHz, the picket-fence disappears > altogether and the spectrum on the panadapter looks spur-free (except for > an > itsy-bitsy one at -108dB/-8200 Hz). > > I'm assuming the IF frequency is 11625 Hz, so my tuning frequency at this > point would actually correspond to the DDS IC set to exactly 50 MHz, (i.e. > an FTW of 0x400000000000 for the 9854). > > What's strange is that, if I turn Spur Reduction ON, as I tune around this > frequency I *never* see a Spectrum that looks as clean as the one that I > see > at 50.011625 MHz with Spur Reduction OFF. In other words, it almosts > seems > that the DDS's FTW is never set to 0x400000000000, otherwise, I would > expect > to see a "tuning block" that's about 3052 Hz wide that has this clean > spectrum, but...I never seem to see it. > > Just wondering if anyone can help explain what I'm seeing (or not seeing). > > Thanks & 73, > > - Jeff, WA6AHL > > P.S. Even with spur reduction ON, there are quite a few spurs in this > region of 6 meters (ditto on 10). Per my understanding of DDS chips, even > though the phase-accumulator is hitting its marks with no phase > "remainder" > bits (i.e. only the top 4 nibbles of the 9854's 48-bit word are non-zero), > quantization introduced by the phase-to-amplitude converter (e.g. the 12- > bit > approximation of Sin(0.0054931640625 degrees) ) also creates spurs. Is > this > amplitude-quantization the source of the spurs that I see around 50.011625 > MHz when I tune with Spur Reduction ON (rather than mixing-products or DAC > non-linearities)? > > > > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] Behalf Of Robert McGwier > Sent: Tuesday, December 06, 2005 10:23 AM > To: Jiri Sanda > Cc: flexradio@flex-radio.biz > Subject: Re: [Flexradio] Interesting behavior when connected to a dummy > load > > > > There is a lot of confusion here. Let me attempt to straighten it out. > > 1) The extra noise occurs (as the original note said) when SPUR > REDUCTION IS <<<OFF>>>. This is because the hardware is tuned EVERY > time the dial is changed. > 2) Spur reduction ON, does NOT move the spur out of band. It does > something quite different and clever. There are "good frequencies" > where the spur generation which is caused by truncation of the phase > word and by the number of bits that can be applied to the DAC are > minimized. If you are not on the good frequencies the phase > accumulator has fractional parts that are not exact values to give the > DAC in the synthesizer. This leads to "walking on and off" an exact > DAC value in the DDS. This "walking on and off" of the good points is a > periodic process and and because the amplitude and phase are just a > little off when you are not exactly on a table phase (DAC bits are > nonzero but the rest are zero), this little bit of amplitude and phase > distortion generated spurs. Spur reduction recognizes this process and > limits the HARDWARE DDS oscillator settings to these good frequencies. > The remainder of the tuning is then done in the "perfect" software > oscillator inside the code. So when spur reduction is ON, and the > frequency request changes by 3051.7578125 Hz from one of these good > frequencies, we move the hardware frequency only then. This approach > has pluses and minuses but it is felt the pusses outweigh the minuses. > > You can still hear spurs from the DDS but these are due to DAC > nonlinearities and clock leakage and mixing in the AD9854. These are > almost gone in the AD9954 and really gone in the AD9958. > > Analog is learning along with the rest of us. > > Bob > N4HY > > > _______________________________________________ > FlexRadio mailing list > FlexRadio@flex-radio.biz > http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz
