What is the easiest way of generating a low-phase noise RF carrier at about 1V into 1K Ohms?
It used to be I would call up International Crystal Manufacturing (RIP) and order a custom crystal.
Who does that nowadays?
Would any of the "DDS" modules do it?
Other methods?
I can re-purpose some GE ICOM channel elements (TXCO modules) or build an Colpitts oscillator from scratch if I have to...
I don't know how accurate is accurate or how low is low, but the Digikey will program an SiT8208 (a 20ppm MEMS oscillator, not a crystal) and you can specify the frequency to 6 decimal places.
Would an LC oscillator be good enough. Or an LC oscillator under control
of a micro which counted it against a crystal and adjusted it
every 1 mS to 1 S?
15.186667MHz
15186667 * 3 is 45560001, is that 1 Hz important?
Looking at the frequencies (and standard crystals)
It is a replacement for receive channel crystal in an old GE MASTR II VHF repeater.
In years past, for Ham radio use, I would order a crystal, take apart the ICOM channel element module, install it in the ICOM, tweak to the center of the channel, and we were good to go...
In application, this oscillator is multiplied by 9 and used for low-side injection into a mixer in a receiver with a 11.2MHz first IF. The channel center is 147.880 MHz.
A nice cheap AD9850 DDS module would make the frequency you want, no problem.
But, and it is a big but, after multiplication there will be significant phase noise and low level spurious frequencies.
For a simple solution, just suck up the cost of a one-off crystal.
The part chosen by dknguyen is better in performance , but is a 3.3V programmable XO. ( internal fractional N PLL)
MEMS OSC have higher Q's than typical AT cut XTALs. They are superior mechanical, non-piezo effects, better thermal, EMI, Phase noise. Field Programmable.
Consult with Digikey Tech support about getting either the XO or VCXO for your frequency.
Since you have a 9th order harmonic oscillator, , the XO needs to have harmonics ( e.g square wave)
It has far superior phase noise, more than adequate tolerance and stability.
If the freq you want is not an integer division or multiplication of the xtal in the Dds then you'll have some phase noise.
One way I can think of is to go ahead and use a Dds, then construct a Pll with a reasonably stable Vco running at your required freq, and lock it to the Dds with a slow Pll loop filter, the Pll will reduce the noise.