Censored Binomial GLM: Sampling fails

One issue I see os that N could be below the observations which would be impossible