If you only support 1 aspect ratio, then just using any ortho camera should be fine - you won't need any special logic for normal orthographic cameras to scale to fill the screen, and especially so since the aspect ratio is constant.
The tk2dCamera will work, but will need the default override added.
To make it work with multiple cameras -
1. Set up depth on the cameras - the values should increase in the order the cameras draw.
2. Make sure your 2nd and 3rd cameras don't clear color - just depth, or nothing.
3. Set up layers on objects, and then culling masks to ensure each camera can only see what its meant to.
That should be it really.
If you're using tk2dCamera, you can link the 2nd and 3rd camera to the first, so you will only need to set up overrides on one.