Originally Posted by: Danlake 
Wasn’t part of the advantage of using the Marklin 3 rail system that you always have 2 rails for ground contact to ensure smooth running?
I believe this to be correct - and a very desirable backup for warping track especially any transition in gradient and/or curves.
Quote:
Just wondering if you isolate long section of rails, if you have observed any performance issues, especially at slow speed when trains are pulling into the shadow station?
It will !! (performance issue = stall)
But risk management has to weigh this against the benefits of using the isolated rail for sensing.
Suggestions
- "Work" at laying the track correctly (physically smooth to mitigate rocking locos from loosing contact)
- Don't uses abrasive cleaners that remove the rail protection once rust gets through it will always be an issue.
- If you have to isolate the rail for sensing on a a curve , do it on the inside rail - the loco will tend to tilt towards the outside and that is where you really need the power circuit to be complete
- for small 0-4-0 type devices, maybe with tires (e.g. glastkasten), an additional wire to the carriage chassis may help
In my ideal world I would leverage the current detection method like the 2-Rail guys do(They have no option). but this doesn't work if there is no power to the track. Both methods have there place as does the actual "tracking" of trains around the layout (i.e. prototypical block control and CTC).
Personally I prefer an "out of band" sensing solution where there is no connection between the track power system. e.g. reed sensing, LISSY, etc. But this is never an either/or discussion