computation errors of cdp mechanisms

[TZKuo]

Hello,
I want to add active conductance in a detailed-reconstructed passive cable model (Schmidt-Hieber 2007),
(modelDB http://senselab.med.yale.edu/modeldb/Sh ... odel=95960)
I add many calcium channels using NMODL by method cnexp.
However, the cdp mechanism uses method sparse to solve the states and often report "errors of not converging in compartment X"
when local calcium flow is large.

On the other hand, when appling these mechanisms in a single compartment model, errors of not converging seldom occurred. But I need to adjust
the rate constant of calcium pump to avoid over-pumping of calcium, which leads to a exaggerating hyperpolarization, is that needed for many general cases?

How can I fix the computation erros? By reducing geometrical conplexity (not preferred) or rewriting all the mechanism by method sparse?
Any help will be very apreciated!

[ted]

Without actually trying your code I can only make general suggestions. You might try reducing dt, increasing nseg, or increasing the number of radial compartments.