I think I can help but too bad because I don't use qq.
Here are some free tips:
1. Before carry out the MCMC, check your Likelihood function again,
2. Check your proposal distribution and rejecting ratio again,
3. If you are using Hybrid Gibbs sampling, you'd better choose your tuning parameter in the proposal distribution very careful.
4. Also, you may want to doing the sampling in groups if you have too many parameter or the rejection ration will be either too large or too small.
5. Good Luck
