Censored Binomial GLM: Sampling fails

I don’t think binomial glms with random N are very common. Have you considered a Poisson or Negative Binomial parametrized in terms of the mean and standard deviation instead?