The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> High-Speed I/O Design >> Pulling between multiple VCOs on same substrate
https://designers-guide.org/forum/YaBB.pl?num=1221811084

Message started by kanan on Sep 19^th, 2008, 12:58am

Title: Pulling between multiple VCOs on same substrate
Post by kanan on Sep 19^th, 2008, 12:58am

Hi,
I am looking at a transceiver architecture with two VCOs, one for Tx and one for the CDR. I read about problems due to pulling and injection between VCOs.
How critical are these problems? Does keeping the VCOs differential and separating them with guard-rings do?
Has anybody tried such architectures and faced any particular trouble?

Title: Re: Pulling between multiple VCOs on same substrate
Post by ci on Sep 22^nd, 2008, 12:40pm

Differential circuits will help as long as the substrate and power supply noise does not exceed the small-signal analysis approximations. Otherwise, the common mode cancellation at the differential outputs is no longer efficient. Also, depending on the architecture, there might be single ended control signals which may capture noise from substrate or power supplies. Typical differential pair stages have tail current source transistors which still operate in “single ended” mode. For example in differential ring oscillators the noise coupled in the tail current source modulates the bias current, which then modulates the stage delay and consequently the oscillation frequency.

If you have access there is an example of substrate noise coupling in a differential CMOS ring oscillator and the implementation of a suppression technique on pages 233 – 238 of the book “Noise Coupling in Integrated Circuits: A Practical Approach to Analysis, Modeling, and Suppression” available on Amazon.com or on the web site www.noisecoupling.com .

Guard rings reduce only the noise coupled through the substrate, but this is only “reduction” not complete elimination. Part of the noise still propagates in the isolated region, and how much it passes through depends on technology. In bulk CMOS on lightly-doped substrate I’ve seen typically 50% substrate noise reduction when using shunting guard rings (p+ diffusion in p- substrate). Also, guard rings do not reduce the noise coupled through the power distribution. Shielding techniques are better, but those depend on the available options in the fabrication technology.

Regards,
ci