A couple of months ago there was a thread running (which I can't find now!) that included a pdf file done by Paul Romsky in 2015. It details an old reverse unit design apparently dating back to the 1970's. The pdf is well done and includes a good rundown on the theory of operation, a BOM, and schematic information in several sections. I thought it might be fun to design a practical board and test build one just to see if it actually works. I added 5 test boards to a recent board order, and I had most of the components on hand. Lo and behold it got done, and it works exactly as advertised.
Here is the finished board:
And here is the circuit:
The board is 32.5 x 50mm, and the output end is narrower at 24.5mm. It is sized to replace a failed DCDR/DCDS board (NLA at Lionel) and re-use the OEM heat sink for the TIP31/32 transistors. So if you have a shelf queen with failed TMCC electronics and DC can motors, this board could be used to allow simple conventional fwd-neutral-reverse-neutral operation. The only connections required are track AC input and DC output to the can motor(s). It doesn't get much simpler than that. When using the OEM heatsink make sure to use insulating kits for the TIP's. The tabs are common to the collectors, and it would not likely be good to ground them.
The TIP31 and TIP32 output transistors are good for 3 amp load, but the front end used 1N4001 diodes limited to 1 amp. So I beefed up the AC rectifier to a KBP306G and the necessary pcb traces to be able to output the full 3 amps. The trace widths are designed for 2 oz copper to handle the 3 amps. So if you order boards from say OSHPark make sure to spec 2 oz. Gerber file included below. The BOM is in the pdf file for the rest of the 37 components. All small diodes are 1N4148 glass. All resistors are 1/4W. Assembly is pretty easy, all thru-hole. The test board took about an hour to build. I used Dupont pin headers for I/O for the test build, but any 0.1" pitch connector will work.
The Romsky pdf file is also attached for reference. To give appropriate credit the entire circuit is as per Romsky's original pdf, except my changes as noted.
Rod