GlmmTMB giving error realted to log-likelihood

Question

I tried to run the following glmmTMB model to include spatial autocorrelation:

glmmTMB(response ~ var1 + var2 + var3 + var4 + var5 + exp(coord + 0|random).

It gives this error:

Error in fitTMB(TMBStruc) : negative log-likelihood is NaN at starting parameter values

I tried to include this term:

start = list(psi = c(-1, 0)

and update the package "Matrix" to the version 1.6.4, but nothing seems to work.

I anyone can help, I really appreciate it.