I need to generate an ultra-low phase noise sinewave at 10 hz. My idea is to take a pair of matched 5 MHz analog sinewave OCXOs, each having a phase noise spec of -130 dBc at 10 Hz. My plan is to offset them by 10 Hz and mix their outputs to get my desired frequency. Anyone know if this will actually work to get an even lower noise floor? Thanks!

 

Honestly ? use a simple MCU or DDS and maybe with a filter circuit : )

per wikipedia : phase noise is measured and expressed as a power obtained by integrating over a certain range of offset frequencies.

if you know your mcu can generate a good signal that has a stability in frequency of say 10% (10% of 10hz being 1hz), looking at offsets as far as 10hz or even 10Mhz as you usually see with phase noise measurements, you'll get an incredible specs. Which is normal, the variation in stability at high frequencies can be in the order of Khz or Mhz but at such low frequencies, that's a huge gap.

i'm sure you can find a cheap micro or device that you could program to deliver a 10hz signal with good stability : )

that should do you well.

EDIT : i say 10% stability as an exaggerated value, you will be able to get down to <5% easy

 

I wonder what phase noise specs you would need for the 10 Hz output signal?

As you know, phase noise is a power level at a certain distance from the carrier and combining two signals
the phase noise result (putting it simply) will be a sum of both phase noise levels.
For example, combining your two 5 MHz signals, each having the specs of the phase noise of -130 dBc/Hz at 10 Hz
from their carriers, will give a signal having a phase noise level of -127 dBc/Hz at 10 Hz from
the carrier i.e. 3dB higher, i.e. double the power, i.e. you would get a higher and not a lower phase noise.

OF course, you would need to consider the type of the mixer NF specs because it also contributes in the final
phase noise result (is it an active or a passive mixer, is it designed to cover the 10 Hz operating range at its output, etc)

Because you sound needing a sine wave output source at 10 Hz, the use of a crystal oscillator circuit followed by a
digital frequency divider chain (which would improve phase noise considerably) seems to be out of question, is this so?

I need to generate an ultra-low phase noise sinewave at 10 hz. My idea is to take a pair of matched 5 MHz analog sinewave OCXOs, each having a phase noise spec of -130 dBc at 10 Hz. My plan is to offset them by 10 Hz and mix their outputs to get my desired frequency. Anyone know if this will actually work to get an even lower noise floor? Thanks!

 

Phase noise addition does not work that way because it is linked to the frequency.
For a low phase noise 10 HZ sine wave an LC tuned circuit oscillator operating as a CLASS "A" amplifier, followed by a tuned circuit filter, powered from a very well regulated and filtered power supply will be the plan.

 

Phase noise can be quantified by measuring the time difference between zero crossings. It is often created by amplitude variations among cycles of sine waves. So keeping constant amplitudes and supply voltages helps to reduce phase noise.