[R] nls() and "singular grandient"

[Guillaume STORCHI]

Hello,

I have a problem with nls() when I want to fit a model with more than two
parameters to be fitted which are written as start=list(a1=,a2=,a3=,...).
Everytime, it displays:'error... singular gradient'
it's a real pain!!! 

Hope sb knows something about this

Thanks

GS