It looks like you got the test to work ! The need for the collimated light is caused by the thickness of the water film. When you test glass on glass, the interference takes place in the air gap between the two surfaces. This gap is very thin. With the water test, the film of water is many times thicker.
Ed Jones' Sky and Telescope article is an excellent reference on doing the test and he goes into the exact details of why the collimated light is critical. As Ed has stressed you can't make a critical judgment of flatness unless you use collimated light. Without it thou you can tell if the surface is smooth and what the condition of the edge is like.
You might try testing your small flat since you have a better idea of what it's surface quality is and compare them to what your seeing now.

- Dave