Hi, thanks for sharing this issue.
Let's see if we can get a more descriptive error.
In your global NCrunch configuration, set the 'Log Verbosity' to 'Detailed', then try running one of the failing tests.
Open up the processing queue, then find the test task responsible for the test run you just executed.
Take a look at the trace output at the bottom of the processing queue when you have this task selected. Do you see anything interesting? Exception information perhaps?
Something else worth checking is to make sure you're not experiencing a failure on the analysis step. TestCaseSource executes code during the 'Analysis Task' for the respective project. If this code fails, you can receive an analysis error. This will usually be visible against the project itself (in much the same way as a build error).