I'm not sure if I'm in the right sub-forum, so if necessary this can be moved.
I've designed a DIY coupler control device to suit my needs. I have a 5.1 metre plank which runs on DCC and has a John Allen Timesaver track plan. Control is by Lenz and eventually RailRoad & Co software.
https://ogrforum.com/t...port-elderley?page=1
So DCC uncoupling is the go.
I'm scratch building all of the rolling stock and building in my own design coupler control as I go.
My method is to use Muscle Wire.
Muscle Wire is an alloy of (among other things), titanium and nickel. As it says on the pack, it shrinks when heated. To heat it, simply pass a current through it.
I've chosen to use the DCC track power, which is nominally 14 Volts. If I ballast the circuit with a 22 Ohm resistor, the Voltage is low enough to heat the wire without destroying it.
To switch the track power on and off, I've used an ordinary function decoder; connected to the 16 Amp 30 Volt DC primary windings of a relay which is rated to switch 250 Volts 16 Amps. This means that the wire can be left on indefinitely without damage.
Using a box cars as an example, I've made the chassis stiff by adding longitudinal stiffeners which were spaced far enough apart to fit the 5 Watt resistors and the relays inside. At this stage, I added the anchors and pulleys for the muscle wires.
Here is a shot showing the aperture through which the actuating thread will pass, on its way to the coupler.
The thread passes through two guides along the way.
This is the crude but effective jig used to make each of the muscle wires. The one in the foreground is used to measure the length; while the one at the back is for the final testing.
The thermal length (between the ferrules), of the muscle wire is quite important. I found that 26.5 cm works best for me.
Brass ferrules 1.57 mm x .355 mm x 8 mm in length, are used as talaurits to secure each end of the wire. The wire is passed through the ferrule, through a solder tag (which has an appropriately colour coded wire attached), and then back through the ferrule. It is then crimped half way with multi-grips. I then run a drop of CA (super glue), into the ferrule to cement the wire in place.
I drop the completed end over the left hand side nail and then pull the wire around the right hand end nail to the "cut" mark and cut it with scissors. Even though the wire is only as thick as a hair, it's so tough that eventually blunts the scissor blades.
Then the other end is assembled and crimped.
Then the spring is attached and the harness is stretched on to the testing side. It's at this point that I add the polyester "unbreakable" thread.
The rig is attached to the DCC system. The power is OFF and there is just a small amount of extension of the spring, to prevent tangling.
Now the power is ON and the extension of the spring can be seen, as well as the change of angle of the thread.
Here is a box car that I prepared earlier 
by fitting the decoder, the two resistors and the two relays. The decoder and its stay alive capacitor are tucked in between them. The pick up wires from the trucks are soldered to the two buses.
Now the two harnesses are fitted and the threads are passed through the apertures and guides below. The blue (+) wires are pulled taut and a static anchor bar is cut and bent to fit each harness. The green (Aux 1), and purple (Aux 2), connect to the relays and the blue (+) wires connect to the resistors, which connect to the (+) track bus on one side (marked with a red dot).
The normally open (NO), relay terminals connect to the (-) track bus on the other side.
The decoder wires green/blue and purple/blue are connected to the primary coils on the ends of the relays nearest to the decoder.
A fine cut off wheel in a Dremel has been used to cut a small groove in the side of the coupler jaw. The coupler is held on full throw by over tightening its mounting screw. The thread is held in a helping hand so that it isn't loose, nor is it applying any tension to the spring.
A tiny drop of CA glue is applied to the thread where it passes over the groove.
When the glue has set up, the excess thread is snipped off with decal scissors and its end smoothed with a file. The coupler mounting screw is released so that the coupler returns to its correct position with plenty of free movement.
I have pinned the rear of each coupler box to stop them swinging.
Then it's tested on the test track. The decoder is given a new address which corresponds to the number on the A end of the car. The A end corresponds to F 1 on the DCC controller and the B end (with the brake wheel), corresponds to F 2.
Now to test it on the layout . . .
!
