Really long fixed step simulation times using powertrain blockset

Hello,

I'm trying to create an engine model using the powertrain blockset which will need to be implemented in real time or faster and with fixed step. Currently running the simulation and while it runs super smoothly and quickly on variable step, it requires an incredibly small sample time step to run without errors such as:

"Derivative of state '1' in block 'BLOCK' at time 0.01 is not finite. The simulation will be stopped. There may be a singularity in the solution. If not, try reducing the step size (either by reducing the fixed step size or by tightening the error tolerances)"

Obviously, since in our application requires real time, we cannot afford to reduce time step much. I'm thinking that since there are likely errors with my model, are there any rules, trends or approaches people take to make the simulation happier and more stable? An example of an observation I've made is using larger volume control volume blocks does this, however these values are fixed for me so I cannot change this. If anyone has found anything else or similar to this, would be great to know.

Or could this be a powertrain blockset issue as a result of using a high fidelity model?

Thanks for any help,

David