What happens to the spectral purity of a ring oscillator with different delays for each stage? My understanding is that even for a highly linearized SS model staggered poles will lead to a much wider frequency spread to achieve the avg. 360 phase shift; hence higher Jitter/Pnoise. This can also be explained if we start looking at the non-linear limit cycle behavior; the trajectory might still be stable e.g. Demir; but the boundary will be spread (chaotic) even if we ignore internal noise of the devices. If I'm correct is there a way to demonstrate this using HB-PSS?