Peco Track

I’ve seen a number of recommendations in favor of using Peco Code 55. Even one of the modelers of North American prototype railroads at my local hobby store says that Peco is what he recommends and that it is one of their best-selling items (and they certainly have a broad inventory of it). So I’m giving this serious consideration for use on the new flextrack-based layout. This page is to collect information and links to pages that I’ll want later as I start construction, assuming I go this route.

Peco is a British company, so their track is designed for 1:148 scale British N. That’s still standard 9mm N-gauge, but it means that the ties will be slightly larger and farther apart than track designed to 1:160 scale American and European N. I’m modeling Japanese N, which uses 1:150 for commuter trains and 1:160 for Shinkansen, so I’m not going to obsess over this aspect. I could use two different rail suppliers, of course, but that seems like more work than I want to take on (e.g., dealing with two different sets of standards when designing and laying track and planning switches).


Peco’s track and switches don’t conform to NMRA standards, but instead appear to be based on the European MOROP NEM standards, with their own non-standard quirks. MOROP and NMRA are interoperable for the most part, but it does mean that some things like flangeway dimensions are different. In general NEM allows for larger gaps, which are more tolerant of badly-designed wheels and wheelsets, but have a less prototypical appearance. In N-scale, most of that’s pretty hard to see from a normal viewing distance. Note that despite the deeper flangeways, Peco code 55 track will still have problems with “pizza cutter” wheels (badly oversize flanges) hitting the ties. I don’t think that’s likely to be an issue for me, but it’s something I need to check for any use of code 55 track from anyone.

Peco uses a curved turnout (and frog) with a constant radius, unlike most other switches that use frogs with a fixed angle (i.e., the frog and diverging track are normally straight and at a specific angle). Further the angle of the frog is 10°, equivalent to a #6 switch, regardless of the radius of the turnout (this is only true of their code 55 track, see this site for more info). For comparison, note that Kato’s #6 is roughly equivalent to 718mm, or 28” (about halfway between two of Peco’s sizes) and provides a 15° divergence.

Flex Track

By all accounts the way that Peco builds their Code 55 track is more forgiving of wide-flanged wheels than other manufacturer’s Code 55, so I’ll probably end up using it. Peco Code 55 track is really code 80 rail partially embedded in the ties so that it sticks up 0.055” above the tie, but still has the structural strength of code 80. This does, however, mean that you can’t simply connect it to someone else’s code 55 rail or switches with a rail joiner, as you’d end up with a vertical offset in the top of the rail that could derail trains. It also has low-profile spikes as well as flangeways at switches that are both wider and deeper than called for in NMRA standards.


Peco has a large range of track switches (aka “turnouts”), albeit with a quirky design of their own. I’m going to focus on their “electrofrog” code 55 switches, since I’m likely to use code 55 and only electrofrog switches are available for code 55 apparently. There are a number of differences in the details of switch design for other rail sizes.

In electrofrog turnouts, the frogs are live and power-routed (i.e., they get power from the point rails touching the side rails. That’s suboptimal for DCC, as you really want to have the frog fed separately with some kind of external reversing system. That can be done, but involves some cutting and soldering (see below).

In Code 55, Peco has three choices for normal switches: Small (229mm or 9” radius), Medium (457 mm, 18”), and Broad (914mm, 36”). That’s a bit problematic, since to ensure proper operation of the longest cars, I’d want to use broad-radius, which may be a bit large. That’s probably what I’ll do for Shinkansen, but I expect that medium will be a more typical choice for commuter lines.

They also have double-slip switches, a double-crossover, and curved switches. Generally I’m not likely to use these, but the double-crossover is found on some of the lines I plan to model.

Double Crossover

The double crossover is available in code 55 only with “electrofrog”. It uses a medium-radius turnout (457mm, 18”) and is 271mm (10 11/16”) in overall length. According to Wiring for DCC the switch does not require any cutting but an auto-reversing circuit needs to be connected to the two frogs on the outer side (the inner side is properly polarized and the guard rails are dead). The need for such a reversing circuit adds to the cost, but does make this a much simpler conversion than most Peco switches.

Wiring the Switches

One of the common recommendations for these switches is to improve their long-term reliability by cutting rails to isolate the points from the frog, and wiring each point rail separately to match its adjacent rail. This is a fair bit of cutting and soldering, and I’ll need to consider it more before deciding what to do. A good description of the “cut and solder” approach can be found on Allan Gartner’s Wiring for DCC Peco Switches Page. I’ve generally found Allan’s advice on power to be sound, so the fact that he’s in this camp makes me give that argument more weight. One thing I’ve learned over the years is that advice like this that makes more work up front is often founded on very real experience of problems anyone keeping at the hobby for the long term will encounter. I’ve had my share of ignoring such advice and getting bit in the past.

A counter-argument is that all this really does is protect against shorts at the frog in a derailment, assuming you’ve wired a feeder to the frog and gapped the running rails beyond the frog, both things you should do with any electrified frog. I’ve found a description of that approach using a Tortoise switch motor (which I was considering anyway) on this site. I don’t entirely agree with the author, who assumes adjacent track will power the outside rails and I’d wire feeders to them, but otherwise he makes a good case for a simpler approach. I do find this tempting. I’m going to be doing DCC, so I’ll have a circuit breaker to deal with shorts. And I plan to have small enough electrical blocks that a short won’t affect multiple trains. Thus I may not need the full “cut and solder” solution.

Another point to make is that since the Peco track has larger gaps between rails at the frog than called for by the NMRA standards, there’s less chance of a short from an out-of-spec wheelset. This is also true at the points, where the point rail pulled away will have the opposite polarity from the side rail, so any contact with a metal wheel would cause a short. These larger gaps mostly limit the risk of shorts to actual derailed wheels. That’s certainly a problem, but looked at one way it’s better to detect these by them shorting a switch and stopping the train, than to keep on rolling until they snag on something and pull the whole car off the track. This, of course, assumes you have a reliable circuit-breaker that isn’t going to let a short weld the wheels to the rails.

Peco switches have a strong spring on the points, which holds the point rails very firmly to either side, ensuring a strong contact with the side rail. This still shouldn’t be assumed to provide electrical contact (over the long term, dirt will get between them even if paint doesn’t). This spring is important if you’re using the standard Peco switch “motor” (actually a relay) or some other electromagnet-based way of throwing the switch, or if you’re using a manual ground throw (e.g., Caboose Industries or similar).

However, if you’re using the turnouts with a continuously-on slow-motion motor, like a Circuitron Tortoise or with a fixed-position motor (a servo), then you don’t want the spring fighting the motor when you throw the turnout. On some of the turnouts the spring is accessible from below and can simply be pulled out. On others, it takes more work (the site linked above discussing Tortoise use has a description). Once you do this, of course, you’re counting on the Tortoise or other motor to keep the switch in the thrown position. But that’s what it’s designed to do.

One problem I have seen with Tortoise’s in the past, is that sometimes the electrical switch feeding the frog doesn’t break contact with the old rail before the moving points make contact with the new rail, causing a short. This, however, was a problem on old HO Walthers/Shinohara pre-DCC switches that had “wipers” under the point rails, so the points made electrical contact with the side rails long before they made physical contact (and long after they broke physical contact). Removing the wipers solved that problem for me. So unless the Peco switches do something similar, I’m probably safe (I’ll need to buy some and see).

Or I could just invest in a specialized DCC reversing circuit for the frog (e.g., the Tam Valley Depot Frog Juicer). The only problem there is that there’s a limit to how many of these you can have per booster, although since I’m planning on using smaller boosters and relatively few switches, that’s probably not going to be an issue for me.

One question you might ask is “why use tortoises at all?”. Certainly Peco’s relays have a lot of fans, and a reputation for being reliable. And they are inexpensive: under US$10 compared to around US$15 for a Tortoise (assuming you buy those in bulk). The extra $6 will add up over dozens of switches.

I have several reasons: first, I’ve used both kinds of switch controls in the past (although not Peco’s specifically), and I prefer the silent slow-motion throw of the tortoise to the quick and quite audible “snap” of a relay. Second the normal mount of the Peco requires either a large rectangular hole in the subroadbed, or a very visible housing to one side, while a Tortoise can be mounted on the underside of the subroadbed, well below the rail, with a drilled hole. Drilling is easier than cutting. Finally, I still have about a dozen Tortoise’s from my old layout, so reusing them will save me some money.

An alternative to the Tortoise is DCC Concepts Cobalt. For DC use, these cost about twice the Tortoise at US$24 (in the U.S.; they may be cheaper in their native Australia). But for DCC there’s a low-cost decoder that slots into the housing, making the overall cost much more attractive compared to adding an external accessory decoder. And bought in bulk, the cost is much closer to that of the Tortoise. Also, where the Tortoise has a circuit-board for wires, which requires either soldering directly to it or use of a hard-to-find socket with soldered connections, the Cobalt has a simple friction-fit wire connection (push a button, stick the wire in, release the button). In terms of convenience, these seem likely to offset any added cost.

A third alternative is use of a servo-motor. This requires some additional control circuitry, so it’s not necessarily cheaper than the Tortoise, but it will be more compact, which may matter in places like a yard.