The offset frequency is not characterized, but usually specified and a strong function of the modulation or commnication system. As an example, RADAR versus narrow band FM. Each would have it's phase noise specification versus offset from carrier dictated by the nature of the detection system.

Thanks for your gentle response Alan Victor ,

I am trying to find offset frequency for 148GHz, so what would be my approach to select the best offset frequency.

...

Anuj

I understand. Was not aware of your carrier frequency. In any case is the application specific about desired offset and phase noise required? If not, then I would consider scaling your oscillator from known specified types, usually drven by the type of resonator and active device and their offset to yours. For example, a microstrip oscillator operarating at 10 GHz would specify offsets at 10, 100 kHz and catch the floor at 1 MHz. This is driven partly by loaded Q and partly by circuit application. In any case, scale those values to your operating carrier. Also, consider that the noise floor is intersected at about fo/2Ql, hence walk back and establish offsets that are in line with your noise sources of the active device and the resonator shaping of the noise. You will want to capture those as well. Is this source locked or free runing, optical or ??

I should add, you should look at the IEEE transactions on Terahertz technology and check out typical offsets reported by others. I would expect you will fnd that they share a common value, say 0.1 % of the carrier frequency or less! And move from that offset to capture the shape of the noise down to the noise floor.

Many of these sources are phase locked to minimize spectral width increase about the carrier. The amount of loop gain required by the PLL is directly tagged to the free runnning phase noise at some offset and that offset dictates the loop bandwidth. Hence loop bandwidth and loop gain will drive the desire to know oscillator phase noise and offset. What is a reasonable PLL gain, bandwith is one point to start.

Hi Alan Victor,

Actually i am designing a oscillator for the W band application that oscillator will be used in PLL. In PLL gain, bandwidth these points are unknown.

Those points can be determined. They are design parameters and you are free to choose. You may start with an assumption. Assume you wish to correct the PN at 100 kHz offset from the carrier and 30 dB of correction is requred. Then find the desired loop BW, from the loop parameters. Of course it would be nice to measure your oscillator and fine tune this estimate. This may not be possible without either a FLL or PLL. So consider some off the shelf hardware that woud bootstrap the process.