Model Train Power
The ordinary model railroad uses a variable-DC power supply, commonly called a “power pack”. This creates a DC voltage on the rails from zero Volts DC to some upper limit, and can reverse polarity of the voltage, which causes a train to run in the opposite direction. The upper limit of the voltage tends to vary depending on the scale of the model. N-scale locomotives nominally have a 12 V DC upper limit, although in reality most “N” power supplies provide a higher voltage, while HO has a limit of 16 Volts. However, that’s only approximate. Some power packs will put out considerably more, and dirty track can make the effective power at the train considerably less without affecting train operation.
Historical factoid: in the early days, model trains were often run using automobile batteries. This led to original motor designs based around 12 volt DC motors.
DCC takes the basic DC system and extends it by placing a “decoder” in each train to provide DC to the motor, and by putting an alternating current encoded with control information on the track itself. This allows more than one train to run at the same time, as well as allowing for constantly-lit train lighting (both interior lights on passenger trains, and headlights) and smoother control over low-speed operation. In any large or complex layout, DCC is definitely preferable. And even on a relatively small layout, it can be beneficial.
Even if you run DCC trains, you should own a DC power pack. First, this will allow you to test trains for proper operation, and to run them in when new (an hour spent looping at moderate speed on a test track will make most models run more smoothly by wearing off any rough edges and working in the factory-applied lubricants). And second, it can be used to measure the “stall current” of the motor, which is important to know when selecting DCC decoders, as this is the maximum current, and thus the decoder must be rated for that much (or more) to avoid damaging the decoder.
What should the output voltage of a power pack or DCC Command Station/Booster be? That’s not actually a simple question. If you read advertising material or the specifications on models, you may get the impression that all N-scale trains run on 12V and all HO trains on 16V. That may be approximate, in a very general manner, but it’s not really a true statement.
HO used 12V when it was first getting started, just after the second world war. You can find lots of references to 12V motors, transformers and batteries in old magazines. But by the 1970’s power supplies would put out considerably more voltage, often 18V (50% more). Some of this was probably a “more is better” attitude: more power meant poor motors could run faster, and dirty or poorly-connected track could lose more voltage (due to high resistance) and still provide sufficient power to the trains. Some of it was likely also due to a feature the small transformers used in DC power packs: they sag under load, so unloaded power needs to be higher than necessary. But more efficient trains draw less power, so a power supply designed to output 2 Amps at 16 volts is likely going to run a 200 mA motor (0.2 A or 10% of maximum) at a higher voltage.
N Scale has held more closely to 12V, at least on paper, but most N-scale trains are probably operated on generic power supplies putting out higher voltages.
And, of course, power at the power pack is not power at the motor. On DC, small feeders or use of the track to carry power long distances will result in voltage drops at the motor. Dirty track will also reduce power in the same way. Ohms law, V = I x R (volts equals current times resistance) says that if you keep current constant and increase resistance, voltage must drop. And both small wires and track have fairly high resistance (compared to large copper wires).
The NMRA long ago stopped specifying voltages except in a very general manner. NMRA Standard S-9, Electrical, last updated in 1984, simply says that “full throttle voltage at the railhead shall not be less than 12 volts direct current at maximum anticipated load.” There are some other comments there about the characteristics of the power, but that’s all they say about voltage. Nothing is mentioned about scale differences or allowed ranges. When the standards for Digital Command Control (DCC) were added, some additional comments were made. S-9.1, Electrical Standards for Digital Command Control (last updated in 2006), has a section on “Voltage Limits for Transmitting Power through the Rails”. Curiously, though, its main statement is that “the RMS value of the [DCC signal], measured at the track, shall not exceed by more than 2 volts the voltage specified in standard S-9 for the applicable scale.” Since S-9 doesn’t specify a maximum voltage for any scale, that’s essentially a meaningless comment.
The DCC standard (S-9.1) does, however, provide a set of graphs showing maximum, minimum and typical power at both the Command Station / Booster (they use “Power Station”, which is NMRA-speak for either of those) and for the decoders. From this we can see that all power stations have a range of 7 to 22 volts (which implies that DCC has a range of 5 to 20 volts, although don’t read too much into that). And that N-scale, for example, has a typical voltage of 12 volts (at the command station). Decoders are similarly limited, although here the range is 7 - 27 volts for “non-N” scale, and 7 - 24 volts for N-scale.
None of this really tells us what voltage levels “should” be, or except in very general terms what a safe range is. I’ve heard of small N-scale models (in Japan) that aren’t designed to run on more than 10 volts. Most modern N-scale has to work in an environment where power packs often put out 16 or more volts if turned all the way up. Whether they can actually operate for long at that voltage without overheating and damaging themselves is anyone’s guess.
It’s likely that any layout will lose a volt or so in the track wiring, often 2 volts if you aren’t careful about using a heavy bus and putting feeders out to each section of rail. So the supply must be oversized relative to the nominal maximum track voltage of 12 volts “at the rail”. A value of 14 or 15 under normal load seems reasonable. With DCC the same limits probably apply, although you’ll lose an extra half-volt or more in the decoder, so 12 volts at the rail is really 11.5 or a bit less at the motor. In most cases the difference doesn’t matter.
A well-regulated supply would put out close to that maximum whether you were running one light engine (around 0.2 Amps), or a double-headed passenger train with sound and lighted cars (perhaps as much as 2 Amps). The reality is that the voltage in the two cases is likely to be significantly different.
I’ve changed my opinion on what makes a good voltage several times, and you may find a variety of numbers used in my pages (I’m not great at locating and updating old material). Today I think that designing the power for 12 - 14 volts at the rail is probably the right general idea for N-scale. This implies that the power pack or command station / booster shouldn’t put out much more than 15 volts. Larger scales have larger motors, and may have a wider safety range. And N-scale motors may be designed to handle much higher voltages safely. But there’s really no way to know, so I aim to be conservative.
In addition to a constant DC voltage, most DC Power packs also apply a “pulsed” voltage. This is an output signal in the form of pulses where the voltage goes from zero (or some low voltage) to a maximum level and back to zero. This has the effect of making peak voltage higher than average voltage. The pulses provide better low-speed operation because the higher peaks provide more torque to overcome friction (see this paper for more detail; this is also further discussed on the Power Pack Testing page). At one time power pack manufacturers made a point of pulsed power capabilities in their advertising. Today it’s so commonplace as to often not be mentioned.
Pulsed power is also most necessary for smooth low-speed operation, and it unnecessarily wastes power (producing heat) at higher speeds, so more sophisticated systems will reduce or eliminate the amount of pulsing at higher throttle settings. Basic power packs may not be pulsed at all, and simply provide a constant voltage, that only varies as the throttle is increased or decreased.
Surprisingly, Kato USA’s power pack, which I have heard described as producing DC power, actually produces a form of pulsed power because its outputting an unfiltered DC power produced from the rectified AC input. Technically this is called “ripple”, but it’s effectively doing the same thing: providing a higher peak voltage to overcome forces that keep a stationary train stationary (static friction) while providing a lower average voltage to keep a moving train moving at a slower speed. Ripple is an simple and effective way to have peaks higher than average, and it turns out that the rounded shape of these waves is better for motors than the spikes found in some transistor throttles.
Continuous Lighting (CL)
Although it’s not something I use, Tomix’s Continuous Lighting (CL) system is worth mentioning. This superimposes an AC signal on the track, in addition to any DC present to run the train. This allows lights to draw power even when the train isn’t moving, if they’re designed for it. One effect of this is that Tomix DC trains typically have a filter capacitor connected across the motor that needs to be removed when converting a train from DC to DCC.
CL is also reported to damage DCC decoders, so running a DCC train “on DC” using a Tomix power pack is probably a really bad idea.
Not all Tomix packs have CL. Specifically the low-end ones for trams and small train layouts that I use for my Urban Tram Layout do not.
Modern N-Scale trains are much more efficient than past designs, and LED-based lighting for passenger cars uses much less power than bulb-based systems did (although some trains are still produced to older designs using bulbs). When sizing DCC decoders, the "stall current" is what matters, and a 500 mA limit is typically more than sufficient (200 mA would work in many cases, but don’t count on that unless you’ve measured it).
And when sizing power supplies for a layout with multiple trains, efficiency is even greater as typical running current is often under 100 mA for freight or unlit passenger trains, and likely under 200 mA for LED-lit passenger trains. In short, on DC a power pack with even 500 mA is plenty for most modern N-scale, even if you run double- or triple-headed freights or large, lit, passenger trains with two locomotives. Sometimes a DC power pack has a “VA” rating, rather than a maximum current. Just divide VA by the output maximum volts to get current (e.g., a 14 Volt power pack with a 10 VA rating has a maximum current of 10/14 = 0.714 Amps, or 714 milliAmps.
What this means for DCC is that even a "3 Amp" system like Digitrax's Zephyr (which probably should be assumed limited to 2 Amps to allow headroom for the circuit breaker to trip) can support at least ten simultaneously operational trains, and possibly more than twenty.
DCC or DC?
A DCC system requires upgrading all trains with DCC decoders. Although it’s possible to run a DC train on a DCC system, this isn’t very good for the motor, and doesn’t work all that well in practice anyway. It’s also possible to run a train with a DCC decoder on a DC layout, but in my experience it doesn’t run as well as it did before conversion. You really have to pick one method, and use it for everything for best results. And if you can afford to do it, DCC will prove well worth the investment in the added capabilities (e.g., non-linear speed tables, BEMF, speed-matching paired locomotives, etc) that it provides.
However, if you find that daunting, a good modern DC power pack designed for N-scale use will do a fairly good job of running trains smoothly. It’s important to avoid high-voltage ones (more than 15 Volts is bad) unless you plan to be very careful about how high you turn up the throttle.