Could some electronics guru please help ? I am trying to design a very high frequency ( >= 20GHz) ring oscillator with SPICE. As expected, I am using sub-micron MOSFETs. I have small (femto Farad) capacitors from each interconnect to ground, to include interconnect capacitance. Should these interconnects be modelled as transmission line, and if so what would be typical values for impedances to be used ? Thanks in advance for your help.
Interconnect capacitance or transmission line SPICE
Nov 03, 2010
3 Replies
Your "guess" is as good as mine ;-) ...Jim Thompson
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
Make the world a better place, dismember a Democrat.
Assuming one would lay out the circuit as compactly as possible to get maximum oscillation frequency, the interconnects between inverter stages will be short relative to wavelength and thus, modeling them as discrete capacitors is quite reasonable. IC interconnects are largely RC-dominated transmission lines. They slow down sharp edges so that receiving inverters or logic end up transitioning later relative to the driver. With short interconnects, R gets tiny and the transition time is dominated by how fast the driver FETs can charge/discharge the C.
-- Silvar Beitel
The output impedance of your transistors and the capacitance should be enough to give oscillatory behavior. The end effects of the very-short 'transmission line' are dominant in this kind of circuit, I'd imagine. Modeling a few microns of metallization as an infinite series of inductors and capacitors seems ... unnecessary.
Join the Discussion
Have something to add? Share your thoughts — no account required.
Didn't find your answer?
Ask the community — no account required