Skip to main content

A few years back Dale Holzmann made available his circuit for CD operation of Atlas switches. These are his words, edited slightly:

“Atlas switch machines as factory wired can be burned out if the switch is held down too long. These switches can be thrown with a DPDT toggle switch and two 2200uf capacitors. The method is burn out proof since only the energy stored in the capacitor goes to the machine coil (thus “capacitor discharge”). The switch indicates turnout position. For indicator lights a 3PDT toggle can be used. A 4700 uf capacitor can be used to throw a pair of machines in tandem.”

OGR member Tom Frye and I have collaborated to design and test this handy little PCB, which takes Dale's basic circuit and puts all the needed components on one compact 75mm x 24mm PCB. The board can be located out of sight under the control panel and only the 3PDT toggle switch and two indicating (red and green) leds stick above the panel. It needs only 12-14VAC (or DC) power source, and has only 3 wires running to each switch for operation. Each board can operate one or two switches (see User Notes) and any number of them can be mounted side by side on 1" spacing as required. Use of offshore parts such as the 3PDT switch and the caps will keep the cost per board under $5.00.

Here are some Pics of the completed prototype board, and one operating a Lionel 072 test switch.


Here is the board layout and the circuit:

Board LayoutCircuit Snip

For ease of wiring under the panel the board is designed for 3.5mm pitch screw connectors such as the KF350 series available on ebay, like below. But it can also be hard wired and soldered if desired.

KF350 Connectors

Here is a 3D board view

Side Top 3D View

Tom uses Atlas switches and has tested a breadboard version of the circuit with them. All of my testing has been with the Lionel 072, and if it works well with that, it will likely work with any brand of twin coil/plunger type switch machine!

I have not priced these boards at OSHPark, but a fair estimate would be about $15 for three. Obviously for anyone intending to build several, other sources might make more sense. The components are all readily available from several ebay sources.

A pdf version of the User Notes is attached, as is a BOM, and the gerber files for anyone interested in building their own.



Last edited by Rod Stewart
Original Post

Replies sorted oldest to newest

Rod, you're really going whole hog on this PCB stuff!  Good looking project, and something that anyone with Atlas switch machines should probably consider!

Just a note, this type of CD circuit isn't required for the Lionel Oxx machines as the coils only get power until the points move, then they are disconnected.  However, for Atlas switch machines, this is highly recommended!

Rod, you're really going whole hog on this PCB stuff!  Good looking project, and something that anyone with Atlas switch machines should probably consider!

Just a note, this type of CD circuit isn't required for the Lionel Oxx machines as the coils only get power until the points move, then they are disconnected.  However, for Atlas switch machines, this is highly recommended!

Yes John I am quite enjoying all this PCB stuff; it's been a blast. A big thanks to you for bringing me up the learning curve with DipTrace back 2-3 years ago. Who says you can't teach an old dog new tricks? It is good that the Lionels have coil disconnect switches built in that prevent coil burnout, as K-Lines, and most others do I believe. (I have been playing with and repairing 022/072 switches now for about 65 years; so I am kinda getting to know them fairly  well haha! I recently put a post in the Lionel Tips & Tricks forum about doing an on-the-layout repair of a failed 022 switch machine. It sure beat the heck out of tearing up all kinds of track and switches to get to the machine! )

But I don't own any Atlas switches so my testing of this board was all done with Lionels. I figured if it would work on these old girls it should work on most anything. And it throws them very well, clean and crisp from my testing. And for those who don't necessarily care for the nostalgia of old Lionel switch controllers on their panel, this board offers an alternative.

I have sent a few blank boards off to my good friend Tom Frye, who operates Atlas switches on his layout. He is going to build up a couple and give them the acid test so to speak. We should hear his results fairly soon.


Just ran across this post.  Great job Rod!  I also adopted Dale H.'s capacitance throw philosophy and use it to throw my Atlas switches.  Since I use Arduino control for the trigger source I needed to devise an interface arrangement that worked for my layout.  Our capacitor boards are eerily similar.  The other board is a distribution board that interfaces between the Arduino and the cheap relay boards with opto-isolators available on eBay.  I use the 4 pin header bridge to connect the distribution board to the relay board and 20awg bare copper to form a buss between the relay board and the capacitor board with a 2.2 ohm resistor to the common pole on the relays (prevents sticking contacts).  Glad you've gotten things to work for you.



Images (4)
  • mceclip0
  • Distribution Board
  • Header Bridge
  • Assembly

I'm curious if the electrical gurus have an opinion on the need for R1 in the design: a 33 ohm 1/4 watt resistor to limit the inrush current when the capacitors are recharged.

How important is having such an inrush limiter?  Is it for the sake of the circuit itself, or is it important to limit the load on the transformer?

A big spike would very likely be more than the DB107 bridge rectifier limit, but would only be a split second so not likely of consequence. What I can tell  you from playing with the circuit quite a bit, is the recharge time for even 4700uF caps is well less than 1 second, so there is no big time penalty by having it. The 33R size was Dale's suggested size in his original circuit sketch, and I simply chose to stay with that. There have been no issues with either the bridge or the 1/4 watt resistor. The instantaneous current flow is just over 1/2 amp with 18VDC at the bridge output, and the 33R resistor. So it is well within the DB107's 1 Amp capability.

It is worth noting that Dale originally suggested a large 5 or 8 amp bridge feeding as many CD circuits as needed. This would make the wiring simpler if you were hand wiring a bunch of these. In designing the board (because building each using a board is quite simple) I chose to use a small bridge on each so that commonly available AC layout power could be used for any number of them. Once the cap is recharged, there is no current flow except for the tiny amount to light the led. So the total power draw on a transformer is of no real concern IMO.


Last edited by Rod Stewart

I don't have an inrush limiter on supplies with smaller capacitors, but when you're adding 5,000uf to 10,000uf of capacitance, it makes a bit more sense to knock down the big charging spike.  For instance, for my YLB - RailSounds Battery Replacement, I have to charge the 1.5F supercap from a 5V supply.  For that application, I added a 10 ohm resistor to limit the inrush current to a reasonable 1/2A or less.

UPDATE: Tom Frye kindly loaned me a spare Atlas switch machine of his, and I got some daylight today to test the CD switch board with it. Went quite well. I used 3300 uF caps in the board, and set 14VAC supply.

I found that less than 14 VAC the action was a bit sluggish and not consistent. My test 1044 transformer will only put out 15 VAC, and that was marginally better, but 14 VAC works satisfactorily.

The action with the Atlas switch I noted is quite different than a PW Lionel. Lionel switch coils are about 6.5 ohms, but the Atlas are about 29 ohms.

So the inrush throw current will be much lower for an Atlas. And instead of a snappy throw it is more of a slow steady movement from end to end, because the cap stays connected throughout. (As we said earlier there is no built in switch that disconnects the power part way through the stroke like most other switches use.) No matter; it gets the jobs done just fine.

I also noticed that since the cap is fully discharged after each throw, it does not get fully recharged if you make several throws quickly, and after 2 or 3 quick throws the subsequent throws are not fully completed. Makes perfect sense when you think about it.

I noted the 33R resistor maybe gets a little warm after several quick throws, but nothing of any concern in my opinion.

My conclusion is that minimum 14VAC works best with Atlas switches (as it does with Lionels), and for the minor cost difference I think I would use 3300 uF caps with Atlas switches (instead of 2200 suggested by Dale); and 4700 uF for two in tandem.

I can appreciate that for Atlas fans who are used to Atlas pushbutton switches, there may not be any attraction in switching to mini toggles. But for many of us who are used to “lever” type switch actuators, the mini toggles may seem more familiar and easier to use. And with this CD board there is no worry about coil burnout.


Add Reply


OGR Publishing, Inc., 1310 Eastside Centre Ct, Suite 6, Mountain Home, AR 72653
800-980-OGRR (6477)

Link copied to your clipboard.