Single tone signal generation at microwave frequency with best performance
What is the best way to generate a single tone signal with best phase noise and spurious performance?
Is frequency multiplication the best way? single stage multiplication or multi stage multiplication?
I want to produce a single tone 6 GHz signal with a 100 MHz Sine Wave OCXO by multi stage multiplication.
For best phase noise and spurious performance what issues are important to notice?
Best Regards.
Thanks.
If you multiply the crystal oscillator then the far out phase noise floor will be higher than if you used a PLL which will have higher PN closer in (100kHZ or so) but be better further out.
I would use multistage multiplication, it is easier to filter out the spurs should be able to get them at or below the spectrum analyser noise floor (at least you know where to look for them), and for most applications the PN will be more than adequate, with care and a good 100MHz source you should be able to get around -147dBc/Hz or so from 100kHz out, below that it should be very close to 35.6dB worse than 100MHz crystal. If PN in several MHz range offset is a problem then a norrowband filter close to the output will help with both wide band PN and spurs.
You do not give nearly enough info to comment seriously.
But if it were me, i would look at a dielectric resonator oscillator at 6 GHz (possibly sapphire resonator in whispering gallery mode), and phase lock it to the 100 MHz crystal with a phase sampler
Multiplication is the worst method to obtain Pure signal spectrum.I recommend you to use this
http://www.ti.com/product/lmx2594
that is a PLL+VCO which is known the best in industry.
If you don't want to use a PLL, Coaxial Resonators @ 6GHz are available for Discrete Oscillators.They exhibit also very low Phase Noise.
This is my specification: Output freq.: 6GHz, PN. better than: -116dBc/Hz @ 1KHz, -124dBc/Hz @ 10KHz, -132dBc/Hz @ 100KHz, -142dBc/Hz @ 1MHz, -156dBc/Hz @ 10MHz, Spur: < -70 dBc. these are most important parameters that need. Thanks
Maybe this reference with some DRO design considerations is useful
http://www.rupptronik.de/rupptronikd...nal_201804.pdf
Thanks BigBoss. The lmx2594 is so useful IC. but with my specification that I quoted, lmx2594 is not useful for me.
May I ask why ?
Phase noise is too bad, the integrated VCO hasn't high enough Q factor. It's 40dB worse than required at 10kHz offset. You can clean up noise inside the loop to some extent (if reference oscillator permits), but outside the loop bandwidth you see the bare oscillator performance.
This is my specification: Output freq.: 6GHz, PN. better than: -116dBc/Hz @ 1KHz, -124dBc/Hz @ 10KHz, -132dBc/Hz @ 100KHz, -142dBc/Hz @ 1MHz, -156dBc/Hz @ 10MHz, Spur: < -70 dBc. these are most important parameters that need. I don't think that this specifications be reachable by lmx2594. Best regards.
Volker,
Look at what they have done..I don't agree that IC has bad Phase Noise since it's been very carefully designed.
http://erainstruments.com/
But that is only -100dBc/Hz @ 100kHz offset
meryem_eda needs -132dBc/Hz @ 100KHz
The 3.9GHz DRO from my link above (designed by my friend Andreas Gronefeld) has -177dBc/Hz @ 100KHz
OK, I didn't see these specifications.
This can only be realized a High Quality Factor DRO or Coaxial Rasonator Oscillator but take the drifting and aging into account.
YIG Resonator Oscillator will probably the best option but It's really hard to get them..
Dear volker@muehlhaus, yes, PLDRO is one of the best way for my need, and is my answer. but is there any alternative for PLDRO?
I have an alternative: at first generate 1 GHz signal by PLL with 100MHz OCXO reference and a ceramic resonator as VCO at 1GHz (I have a ceramic resonator with -103dBc/Hz@1KHz, -129dBc/Hz@10KHz, -147dBc/Hz@100KHz) and then multiply by 6 (by 3 and by 2) to have 6 GHz signal. What do you think?, is this alternative good enough?
(My PFD phase noise floor at 100MHz is -150dBc/Hz).