This question may be better suited for Cross Validated, but I'll do my best to answer.

A singular fit indicates a problem with the model--overparameterization or insufficient data. The model has not converged successfully and should not be used.

The estimate of 0 for the variance of Area.Code in the first model suggests that it is not contributing much to the model fit, possibly due to insufficient data; however, it is strange that the second model seems to converge. Are you certain that Area.Code should be treated as a random effect? Note that you can treat a covariate as both a fixed and random effect.

The scaling error indicates you have covariates with very different magnitudes. By rescaling the data and rerunning the model, this error should go away.