effective channel length in hspice

Thread Starter

lili94

Joined Nov 6, 2015
4
Hi,
I have a basic question. I want to simulate a 14 stages ring oscillator under all possible values of effective channel length (from 35nm to 55nm) and temperature(from 20 C to 150 C) via Hspice, with the 45nm high-performance predictive technology model. But I don't know how simulate with different effective channel length and temperature! Please help me. This is the what I write with 90nm, but it doesn't work:

*CMOS Ring Oscillator

.op
.include 'e:\PTM90.txt'

V1 vdd 0 5V
V2 vss 0 0V

.PARAM Le=90nm
.PARAM Wi=180nm


.subckt inv Vout Vin Vdd Vss
* D G S SS Model W L *
M1 Vout Vin Vdd Vdd PMOS W=Le L=2Wi
M2 Vout Vin Vss Vss NMOS W=Le L=2Wi
CL1 Vout Vss .1pF
.ends INV

*14 stage ring
x1 2 1 vdd vss inv
x2 3 2 vdd vss inv
x3 4 3 vdd vss inv
x4 5 4 vdd vss inv
x5 6 5 vdd vss inv
x6 7 6 vdd vss inv
x7 8 7 vdd vss inv
x8 9 8 vdd vss inv
x9 10 9 vdd vss inv
x10 11 10 vdd vss inv
x11 12 11 vdd vss inv
x12 13 12 vdd vss inv
x13 14 13 vdd vss inv
x14 1 14 vdd vss inv

.MODEL nch NMOS
.MODEL pch PMOS

.tran 10ps 64us
.prob
.END
 

eetech00

Joined Jun 8, 2013
3,946
If HSpice supports a ".step" function, you could step through the value.

* step Le value from 35n to 55n in 1n increments
.step param Le 35e-9 55e-9 1e-9
 

Thread Starter

lili94

Joined Nov 6, 2015
4
If HSpice supports a ".step" function, you could step through the value.

* step Le value from 35n to 55n in 1n increments
.step param Le 35e-9 55e-9 1e-9


Thanks eetech00, but my main code for 14 stages ring oscillator doesn't work, I mean without multiple length. Just the main code.
 

eetech00

Joined Jun 8, 2013
3,946
Sometimes you have to "kick start" oscillation.
Try setting the initial value at the input node of the first inverter to "0" volts.

I think that is node "1".

.IC V(1)=0
 
Top