If you look at the Template Editor in this area I can almost guarantee that the model is correct, it is just the Surface Model output which is incorrect Pat - this is because the Surface Model is created using Surface Model Rules based on data that is interpolated from the template Model. Where you get Hard Stops of elements at a specific station, and then you get Interval Locations e.g. every 5' you will see this sometimes in the Corridor Surface Model. If you eg reduce the interval you will see it improves the situation (never entirely fixes it),
I have some requests in the system for this as I agree that it is an issue that the Surface which is what you will ultimately use to run the machines etc. is important, and while it is an interpolated model from the template Model, these locations where things start and stop or where you define Tables need to be generated more accurately. If you drop a copy template into the model at that station you will likely see further improvements because the Template Drops get forced into the model as a priority.
If I see these types of issue that the Surface Model looks wrong, I will just create a Reference Template Drop at the station where I want to see the change and you should see that it fixes the problem. Currently Tables have a tolerance of I believe 0.33' longitudinally and they do not 100% honor the Table Contents if there are lateral, vertical or slope steps in the tables that are closer than that tolerance of 0.33' (that creates issues in e.g. MSE Wall or Gabian Wall footings and that is where I see it most and that is where using additional Template Drops at the start and end of the steps gets me what I need.
I have development looking at this currently
Hope this helps
Alan