convergence problem in dc sweep curve at 1.5000

i am trying to simulate my circuit in hspice... it is showing these errors......

____________________________________________________________________________________________________________

vin 1 0 dc 5v
G12 1 2 poly(2) 1 0 2 0 0 1.4783676e-009 -1.4783676e-009 -3.9434654e-009 7.8869308e-009 -3.9434654e-009 3.6185472e-009 -10.8556416e-009 -10.8556416e-009 -3.6185472e-009 -1.1063002e-009 4.4252008e-009 -6.6378012e-009 4.4252008e-009 -1.1063002e-009
G23 2 3 poly(2) 2 0 3 0 0 1.4783676e-009 -1.4783676e-009 -3.9434654e-009 7.8869308e-009 -3.9434654e-009 3.6185472e-009 -10.8556416e-009 -10.8556416e-009 -3.6185472e-009 -1.1063002e-009 4.4252008e-009 -6.6378012e-009 4.4252008e-009 -1.1063002e-009
G36 3 6 poly(2) 3 0 6 0 0 1.4783676e-009 -1.4783676e-009 -3.9434654e-009 7.8869308e-009 -3.9434654e-009 3.6185472e-009 -10.8556416e-009 -10.8556416e-009 -3.6185472e-009 -1.1063002e-009 4.4252008e-009 -6.6378012e-009 4.4252008e-009 -1.1063002e-009
G14 1 4 poly(2) 1 0 4 0 0 1.4783676e-009 -1.4783676e-009 -3.9434654e-009 7.8869308e-009 -3.9434654e-009 3.6185472e-009 -10.8556416e-009 -10.8556416e-009 -3.6185472e-009 -1.1063002e-009 4.4252008e-009 -6.6378012e-009 4.4252008e-009 -1.1063002e-009
G45 4 5 poly(2) 4 0 5 0 0 1.4783676e-009 -1.4783676e-009 -3.9434654e-009 7.8869308e-009 -3.9434654e-009 3.6185472e-009 -10.8556416e-009 -10.8556416e-009 -3.6185472e-009 -1.1063002e-009 4.4252008e-009 -6.6378012e-009 4.4252008e-009 -1.1063002e-009
G56 5 6 poly(2) 5 0 6 0 0 1.4783676e-009 -1.4783676e-009 -3.9434654e-009 7.8869308e-009 -3.9434654e-009 3.6185472e-009 -10.8556416e-009 -10.8556416e-009 -3.6185472e-009 -1.1063002e-009 4.4252008e-009 -6.6378012e-009 4.4252008e-009 -1.1063002e-009

Vx 6 0 dc 0v
.dc vin 0v 5v 0.1v
.plot dc i(vx)
.op
.probe
.OPTION GDCPATH
.OPTION CONVERGE
.option post
.end
-----------------------------------------------------------------------------------------------------------------------------------------

**diagnostic** dc convergence failure,
resetting dcon option to 2 and retrying.

convergence problem in dc sweep curve at 1.5000
pseudo tran is tried

*** initial damped pseudo transient completed. ***
*** final try started ***

**diagnostic** dc convergence failure,
resetting dcon option to 1 and retrying

**diagnostic** remove cshunt failed
**diagnostic** try to load final result like nodeset

**diagnostic** gmin ramping for pseudo-tran
final try failed.

**diagnostic** no convergence in the final try

convergence problem in dc sweep curve at 1.5000
gmath ramping is tried


**diagnostic** no convergence in gmath method gmath= 2.00



**error** no convergence in dc sweep curves at 1.5000



1 ****** HSPICE - Y-2006.09-SP1 32-BIT (Nov 24 2006) 22:13:11 04/21/2010 linux
******

****** operating point information tnom= 25.000 temp= 25.000
******
***** operating point status is debug simulation time is 0.
node =voltage node =voltage node =voltage

+0:1 = 1.5000 0:2 =-251.2509 0:3 = -16.6533
+0:4 =-251.2509 0:5 = -16.6533 0:6 = 0.




hspice diagnostic for nonconvergent nodes and elements


node subcircuit old new error
name definition voltage voltage tolerance
(3) main_ckt -16.653 -15.941 42.625
(5) main_ckt -16.653 -15.941 42.625
______________________________________________________________________________________________

how to get rid of this error..

What are you trying to do? The G sources have two inputs but the output depends on the inputs so there is a real possibility the iteration will just shoot off to +/-infinity. I guess that is what is happening. I think you need to go back to what you are trying to achieve - I think the theory is flawed.

Keith.

ya thats my problem... i have to simulate a negative differentail resistance .... i have the I-V data points...... but when i try to simulate a circuit containing only these resistances, i fail because the resistance of the resistances depends on the potential difference between the nodes it is connected......

how do i get rid of tyhe problem...

Could you post the curve you are trying to achieve? Can the characteristics not simply be broken down to a single polynomial? I am also not sure why you are using so many identical G sources (although just using one doesn't solve the problem - you cannot have an uncontrollable feedback).

Keith.