I wanted to have a few working PRR type signals on the layout so I decided to use very scale looking brass signal poles and lighted signal heads from Alkem Scale Models. A circuit board from Heathcote Electronics provides the between-the-rails infrared train detection and signal animation.

You have to replace the solid pole that comes with the signal pole kit with a hollow one so you can run the signal head wires inside it.  I tried an 18 gage medical needle and it seemed go be a good size.  I cut the bottom plastic off and was surprised to see that it looked like a pole support piece.  

The pieces were then dry fitted before glueing.  I cut one of the unneeded wires off at the head as I wanted a seven position head instead of nine.

The signal head was glued to the back of target and I put a small strip of paper between the two to keep stray LED light from shinning out the sides.  The pole (needle) was put through the hole in the assembled brass base and the supports and ladder were glued to it. The large brass rod sticking out the bottom was for assembly support but will be used to get the very small wires to the bottom of the layout. 

The typical PRR signal head now has seven lights.   The Alkem Scale Models PRR signals look really GREAT!

The signal has been painted and is in position. At this point I haven't decided on how to activate it yet, but I manually applied some reduced voltage so it could have its picture taken. I have noticed some light bleeding from one light to another and a little light leakage through the sides at places. On the next signal, I think I will paint the four edges of each LED in the signal head with flat black paint before attaching the head it to the front target.
p.s. Looks like I have to call the signal maintainers as the pole seems to be leaning a little.

After some research into train detection and signal animation methods, I decided on the MAS-SEQUENCER-4 board from Heathcote Electronics in Great Britain.   I ordered the board configured for 3 signal aspects, common anode, no internal limiting resistors and the optional 18-inch leads for the between-the-rails IR detectors.  The LEDs on the left are for testing. 

I now have a working day-or-night 3 aspect simulation signal on the layout thanks to the MAS-Sequencer-4 board.  When a train just passes the signal it turns from green to red.  A yellow signal will appear after the train passes the signal and a time delay has occurred.  After another time delay, the signal will return to green to await the next train that comes.   I use the term red/yellow/green for clarity, but signals used on the Pennsylvania Railroad really show three different positions of amber lights. The Sequencer board can also be ordered for 2 aspect (red/green) or 4 aspect (red/yellow/double yellow/green) signaling if so desired. 

My layout is in a low light room and I sometimes run trains at night (e.g. after the wife goes to bed) so I wanted a solution to the simulated signaling that would work in that environment. Logic Rail makes a very nice board that does the same type of signal simulation but it relies on light sensitive photocells. The board from Heathcote is the only one I could find that includes small infrared (IR) devices.

The Sequencer has two 3mm infrared detectors that go between-the-rail (under the space between two ties) and "looks up" to sense the presence of train.  I ordered the optional 18" leads on the detectors as I had to drill up through the hollow-core door base of the layout, 1" of foam, the cork roadbed, and have enough lead left to route it to the location under the layout where I wanted to mount the circuit board.

The installed IR detectors have worked perfectly in the detecting of n scale engines and rolling stock.  I have not seen any false starts or "flickering" as a train goes by.  The Sequencer board(s) can also be linked together as shown on the Heathcote web site but I don't plan on doing that on my small layout.

 Some installation notes:

1. Put a 1/4" length of shrink tubing on the end of the detectors to hold them together at a very slight angle to themselves.  Wrapped an inch or so of electrical tape to give the end of the detector wires some rigidity. 

2. Drilled a 1/2" hole up through the bottom of the hollow-core door. I managed to hit a solid part of the door interior where the knob would go.

3. Very carefully finished the hole up through the extruded foam layer and cork roadbed using various shape instruments.  The space between two ties (sleepers to our English friends) was cut-out from above with a sharp knife.

4. The detector leads were sufficiently stiff enough that I could push them up from the bottom until the top of the outer edges of the detector hoods touched the bottom of the ties, and they looked up through the space.  A dab of glue on the outside top of the detectors hold them in place. 

5. Final tweaking with include varying the 3.2k ohm limiting resistor for signal light intensity and adjusting the variable time delay for aspect changes on the Sequencer board. 

For my application of signal simulation, the MAS-Sequencer-4 board from Heathcote has been a good product that performs flawlessly.