Thanks to all who have contributed to our discussion of phase noise and the Si5351 chip. Let me throw out some ideas -- some technical, others philosophical.
1) We may be worrying about this too much. In all of the homebrew or kit rigs we've built over the years, I never recall much concern about the phase noise specs of the LC or crystal oscillator circuit that we were using. What were the phase noise stats on a Heath VF-1? How about the phase noise stats for the little Hartley oscillator in those DC receivers we made? No one even checked. Our rigs usually worked just fine. We would have noticed if they were extremely noisy, but if they were good enough, we left well enough alone. It doesn't really make much sense for us to now be suddenly very concerned about the phase noise stats of the various DDS and PLL chips that are replacing those LC and crystal circuits, especially when the measurements show that they are usually in the same range as our old familiar oscillators.
2) The perfect can be the enemy of the good, and the "good enough." We have a long tradition in ham radio of tolerating less-than-perfect or less-than-optimum parts. Remember, the NE-602 has some shortcomings, but we use it. We use it a lot. The IRF-510 wasn't even designed to be an RF amplifiers, but we have pressed it into service for our PAs.
3) We should be willing to give a new part a try, and we should be pleased if it proves useful. We should be wary of untested claims re the unsuitability of a component. We have to avoid the "works in practice, but not in theory" situation. If something works well, doesn't create additional QRM, is inexpensive, and fosters experimentation and homebrewing, we should be happy about being able to use it.
4) All electronic components -- not just the Si5351! -- produce noise. Resistors produce noise. Look at this:
" We can infer... that if we install phase-quiet oscillators in transmitter and receiver, we ought to be able to tune our receiver to a frequency closely adjacent to a very strong signal from the transmitter without encountering anything like phase-noise hiss. Yet, after an exhaustive phase-noise cleanup at transmitting and receiving sites, we test our communication system only to discover that the transmitter still emits broadband hiss! The culprit is transmitted amplifier noise. Just about every modern transmitter or transceiver consists of a high-gain, linear amplifier strip that amplifies the low-level output of oscillators, mixers and phase-locked loops to hundreds of watts or a few kilowatts. Because amplifier circuitry is not perfectly quiet, the output of the transmitter contains noise (hiss) in addition to the amplified signal. Transmitted along with the desired signal, this hiss can degrade the noise floor of nearby receivers-just as transmitted phase noise can. Where does amplifier noise come from? Thermal noise, for one thing. Electronic components operated at temperatures greater than absolute zero generate random electrical noise. This noise is broadband in nature. Greatly amplified in an audio amplifier-or greatly amplified in a radio transmitter, transmitted as broadband radio noise, received and converted to audio-it sounds like hiss. Random variations in electron flow within active amplifier components (transistors and vacuum tubes) are another source of amplifier noise. Transmitted as broadband radio noise, received and converted to audio, it also sounds like hiss." Source: http://www.robkalmeijer.nl/techniek/electronica/radiotechniek/hambladen/qst/1988/03/page14/index.html
5) It seems that whenever a new technology or part comes along there will be those who issue dire warnings about how we can't or shouldn't use it. When transistors came along, there were those who said that hams shouldn't homebrew with them because -- it was argued -- without spectrum analyzers we couldn't possibly come up with spectrally pure signals.
6) We have to be careful lest this obsession with perfection and extremely high tech standards be used as a rationale for not homebrewing, or (much worse) as an argument against homebrew rigs on the ham bands. There is a bit of this going around. Get on 40 meters with rig that drifts a bit or that is not "on frequency" to within 10 Hz and you will find out what I mean.
7) The Si5351 is a good part for our purposes. It does something new and VERY useful for us: It can put out BOTH our VFO and BFO frequencies. It makes it much easier for us to change bands and-or switch between USB and LSB. Its phase noise figures are fine. LA3PNA (citing measurements by KE5FX) notes: "The phase noise of the Si5351 is around -130dBc/Hz at 10KHz. This is quite decent, If compared to a Hartley or Collpits you would see little or no difference. Some of my measurements of published free running oscilators show phase noise in the -110dBc/Hz range!"
-130 dBc/Hz at 10 kHz puts this part on the "good" curve of this chart. From (http://www.robkalmeijer.nl/techniek/electronica/radiotechniek/hambladen/qst/1988/03/page14/index.html
We should give this little chip a chance! Give it a try!