The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> PVT Simulation
https://designers-guide.org/forum/YaBB.pl?num=1198004686

Message started by SuperMentor on Dec 18^th, 2007, 11:04am

Title: PVT Simulation
Post by SuperMentor on Dec 18^th, 2007, 11:04am

I have deisgned a LNA, now I want to test it over PVT variations with add SP analysis. How do I go about it in Spectre? and I want to know how to choose the combinations of PVT Simulation (Process, Supply Voltage, Temperature) .Which is Worst Case Simulation?

If the Ocean Script is useful in this case, How do I go about writing the script and how do I run it later on.

Is there difference between PVT variations and mismatch analysis.

Title: Re: PVT Simulation
Post by sheldon on Dec 23^rd, 2007, 6:50pm

SuperMentor,

Responding to the question on the difference between PVT variations, corner analysis,
and mismatch analysis:

Normally, process variation of the PVT variations, corner analysis, effect all
devices in the simulation equally. This analysis is meant to model the effect of running
many wafers through the fab. These variations are sometimes called inter-die variations.
The analysis shows whether the circuit will meet specification for across all expected
process variations or not. By including the power supply voltage variations, V, and the
temperature variations, T, designers can verify that their circuit will meet specification
across all expected environmental conditions.

Mismatch analysis effects each device in the simulation differently and this
analysis is meant to model the effect of device mismatch on circuit performance.
These variations are sometimes called intra-die variations. The analysis shows the
sensitivity of the circuit design to the physical design, layout. As a result of the analysis,
the designer may need to specify layout constraints or modify the circuit architecture
to reduce the sensitivity of the design to layout. The good news is that most parameters
are relatively insensitive to mismatch: gain, bandwidth. Bad news, key parameters can
be highly sensitive to mismatch: IP2. Also in general, deterministic mismatch analysis
has limited applicability. For most RF circuits, designers need to perform Monte Carlo-
based mismatch analysis.

So corner analysis, PVT variations, models "macroscopic" process variations or
wafer-to-wafer process variations while mismatch analysis models "microscopic" process
variations or device-to-device process variations.

Best Regards,

Sheldon