Looking for some suggestions on engineering... I'm building my Granville Bridge on the workbench. It's a custom configuration of the Model Railroad Stoneworks Rockville Bridge kit. Amazingly, the specs are nearly identical (See The Granville Bridge Project).

Half of LEWIS interlocking is situated on top of the bridge deck. Specifically, three sets of crossovers. The bridge is made of Hydrocal. The roadbed is a half inch of Homabed. The issue is control of the turnouts.

On a former N scale layout I constructed a beautiful (if I may so so) Sherman's Creek Bridge. On that model, I ran control rods out the back and painted them black so they weren't very visible. The bridge was at the back of the model railroad, so it worked. Not so on the new model railroad. The bridge will be in the foreground and that won't work.

Though it is possible -- but risky -- to drill bores down through the piers, there are two issues with this approach. First, only one end of a crossover would be above a pier. There is no way to align the paired turnout above a pier. Second, the height is such that the control rod would be driven against the edge of the bore and the results would be unpredictable.

The seemingly obvious solution might be to use an "old school" snap switch or something. The Homabed could be hollowed out below the turnout to accept the switch machine. Obvious problem here is maintenance. Tortoises are rock solid, but maintenance on a snap switch would involve removing the turnout and the ballast that has been installed over it. The second issue is DCC control. I know there are some stationary decoders that can drive snap switches. Any recommendations? Needs to work over LocoNet.

When I did the Shermans Creek Bridge, some had suggested control mechanisms as used in radio control planes/boats/etc. I have no experience with these. Thoughts?

Your feedback is appreciated!