Train Power - When DC is not Direct Current

This is the second (or third if you count the previous diversion into history) in a continuing series of posts on powering model trains. The first installment covered some basics of DC motor design, and if you haven’t read it, it’s worth at least a skim before reading today’s post, since it provides a lot of background information (or read the DC Train Motors page, which has been updated considerably from the original material). Today we’ll get into how modern “DC” power packs aren’t really DC, and what aspects of DC motor behavior make that a good thing.

Most modern “DC” power packs don’t really put out DC in the sense of a fixed voltage for a given throttle position. The reason for this is that DC motors don’t really run very well at slow speeds. They work best at a fairly high rotational speed. In the early days hobbyists had to be content with models that went from “stop” to “rather quick” with no intermediate speed. This is often referred to as a “jack rabbit start”, and it’s hardly prototypical. Real locomotives weigh several hundred tons. Nothing that heavy accelerates quickly, and if it’s pulling a train that can weigh upwards of 10,000 tons it’s even slower to get up to speed.

Passenger trains (multiple unit type) tend to weigh less, and often have more and distributed power so they can accelerate faster. But there are limits to that, as passengers tend to object to being hurled backwards. And “weigh less” still means that a train will weigh 100 tons or more without passengers, and a full load of those can easily add 50 tons. That’s a lot to accelerate quickly.

I’ll get into the details below, but the short version is that modelers discovered that “pulsing”, or varying the power in a repeating manner, allowed a train to start more gently. This wasn’t without its problems, and various different approaches were tried over the years, although the earliest technique proved quite effective and continued in use for over twenty-five years. In fact many systems today use a technique known in the 1950s, but not considered “the best” at the time. It’s still a compromise, but a popular one. Today all but entry-level DC power packs produce pulsed power in some form, and it’s so commonplace that most don’t even mention it. There are exceptions, of course.


Occupancy Detection Yet Again

About fifteen months ago work on adding occupancy detection to Sumida Crossing stalled. That was in part because I’d planned to use the BDL168 detectors to also do transponding, and a few months earlier had abandoned that plan since I was unable to get that aspect to work reliably, even on a simple test track. The number of solder joints on the BDL was also a nuisance that caused me to put off further work.

Recently I’ve been rethinking my approach. The BDL168 is an amazingly cost-effective solution. Ignoring the transponding part, you get 16 detectors on a board that includes a LocoNet bus interface for US$120 (street price). That’s $7.50 per detector (if you can use all 16). That’s really hard to beat for a bus-connected detector. I’d originally planned to install one per table on the layout, and my cost would have worked out to around $10 to $15 per detector on average.

On the other hand, I’m thinking that I might want to move to either a OpenLCB/NMRAnet bus (if I want a feature-rich bus for the future) or a really dumb serial bus (like S88 or C/MRI). The latter is attractive since I can potentially interface to it with an Arduino, opening up some room for home-brew devices. Of course I could do that with NMRAnet, but today that requires a US$45 shield to add to the Arduino (or one with it built in), which kind of takes away from the appeal of using $10 Arduinos to do things like drive signal masts.

While thinking about this, I went off and started researching what was available commercially or as home-brew circuitry and software libraries for these busses and for doing occupancy detection with them, as the latter would be a good way to get my feet wet and solve my “don’t want to solder those #$@! BDL168s any more” problem.

But in the past week I’ve been sidetracked into looking at homebrew inductive-coil detection circuits.


Corrected PWM

Update (3/29/13): The “percent of maximum” numbers are all wrong. I’m never going to get this right. I need to go back and recalculate those, but that’s not happened yet. When it does, I’ll correct this post (and maybe make another one).


I really thought I was going to be done with PWM and back to working on decoders this week. But, as is now noted on the last post, I got it wrong. As I mentioned then: “I’m not an electrical engineer. I think I know what I’m talking about, and my conclusions appear to line up with observed reality. But I could have fallen down the rabbit hole and just not noticed. Take it all with due caution and a grain of salt.” Yep, down the rabbit hole I went.

So, as Bullwinkle used to say, “this time for sure!”. Well, sort of. I’ve fixed my problems, but my model isn’t an exact one. For the details of that, and what conclusions I can draw from what I have, read on.

More About PWM

Update 3/16/13: Ok, my graphs (and some conclusions drawn from them) had a serious flaw. I’d modeled both the growth and decay curves incorrectly, and this caused current to drop to zero in a lot of situations where it wouldn’t have. Don provided an interesting simulator model, which led me to do some more reading and correct my model. It’s still not complete, but rather than re-do this post, I’m adding a new one that is the correction. I’ll leave this one here for history’s sake, but please see the new one.

PWM Motor Control

DC motors are controlled by varying the voltage and polarity of the DC power connected to them. In a simple DC power-pack a rheostat is used to provide a voltage to the track that varies from zero volts to the power pack’s maximum, which is often around 16 Volts. A simple switch is used to swap the positive and negative outputs to change the polarity (and thus the direction the motor turns). This tends to waste a lot of power as heat, but since that’s happening inside the power pack (and “a lot” isn’t really all that much at these voltages) that’s acceptable.

DCC decoders need to take a constant-voltage AC input from the rails, and control a DC motor somehow. Even if they could use a rheostat, wasting power as heat inside a plastic model is more problematic. The technique normally used instead is called Pulse-Width Modulation, and it’s a fairly simple and commonplace, and efficient, method of controlling DC motors from a digital controller. The same technique is used in many other applications. Read More...

Decoder Wars II - Lightboards

Comparing decoders for cab cars is actually relatively simple. These don’t need to do very much, so it’s really about checking basic functionality. I’ve laid out the full testing details on my Decoder Comparison Testing page, and here I’m going to summarize the findings for the capabilities of interest to me.

Decoder Wars I

Edit: see the comments for some additional notes; also, I’ve edited the text to correct some errors, but those edits are marked.

A long time ago, in a distant land, titans met to do battle...no, wait, I mean recently, on my kitchen table, I started testing DCC motor decoders for N-scale EMUs. And as with most wars, after it started I began wondering why on earth I’d thought it was a good idea. Still, I have reasons for this, and the result is important: the winning candidate will go in my (so far) 27 trains that don’t support Kato’s plug-in decoders, and some have two motor cars, so it’s closer to 30 decoders.

Configuring The EM13 Part I

Kato’s EM13 DCC decoder (29-351) is a specialized decoder used for the motor car in an EMU/DMU model or other “DCC Friendly” models made by Kato. DCC Friendly (the English term is used even in Japanese, rendered as “DCCフレンドリー” or “DCC furendorī”) isn’t the same as “DCC Ready”, and it is a phrase used by others to simply mean that a model is relatively easy to convert to DCC. But when Kato uses it, the phrase means “will accept Kato FR11, FL12 and/or EM13 decoders”. And the models that do so are primarily Kato’s N-scale Japanese prototype models: commuter trains, limited express trains and Shinkansen (bullet trains). They also use it in some steam locomotive models, including the American-prototype GS-4.

Imperial Train

Japan’s Emperor is the head of the world’s oldest continuing hereditary monarchy, reputed to have been established in 660 BC. He is also the last monarch in the world reigning under the title of Emperor. Japan’s current constitution stripped the position of emperor of political authority and, although it still has formal duties, the holder of that title is largely relegated to a ceremonial role. However the Emperor, both the office and the individual, is still highly regarded in Japanese society, and his duties include diplomatic ones such as “receiving foreign ambassadors and ministers”.

When the Emperor or his immediate family travel, they do so like any head of state, with a great deal of security, press coverage, and attention. Mostly such travel today is by car or plane, but given the predominant role of trains in Japanese transportation, this mode is sometimes used as well. If the Emperor travels by Shinkansen, a reserved car will be used. But for travel on the narrow-gauge network, there is a special Imperial Train. Since 2007, the E655 shown above has been used.

Other Lightboards

Up until now I’ve been concentrating on my Kato models as far as DCC conversion goes, and at the same time I’ve only really paid attention to the interior lighting of those cars. Now that I’m working up to large-scale DCC conversion of my non-Kato stock, which is mostly MicroAce and Greenmax at this point, I need to think about lighting the interior of those cars.

And it turns out, they’re pretty much identical. Both MicroAce and Tomix make their own interior lighting kits, and they’re very similar. A third-party company, F&MOKEI also makes lightboards, and claims they work with both manufacturers’ cars. And Greenmax notes compatibility of its cars with Tomix lights (Tomix is far larger a company than MicroAce, so it makes sense they only mention one).

DCC Voltage and Cab Lights

’m turning my attention to the cab car decoder install now, and a recent discussion with Don along with a question from a reader had me thinking about potential problems with DCC conversion of N-scale EMU cars with cab lighting. And the one that really worried me was overvoltage from high DCC track voltages, and its harmful (fatal) effect on LEDs. DCC decoders essentially pass track voltage (minus a small bit) through to their function outputs.


I went off on a tangent this past week. It all started when I asked myself: what gauge wire should I use on my decoders? And the root of that question was thinking that the wire I’d used on my first install, two weeks ago, was too thick. The answer turned out to be less obvious than I thought it would be, and consumed (and is still consuming; I’m not done) a lot of time.

November 2012 Status

November, as you may have noticed from recent posts, went largely to laying the groundwork for installing wire-in DCC decoders, and a bit of testing of same. After a few delays, most of what I was waiting for finally arrived, although a few things are still backordered. In particular, the six-pin NEM651-compatible plugs and sockets mentioned in the comments last time have arrived. For the curious, the parts list has been added to my page on DCC Decoders. Read More...

Wired Decoders II

Although the rest of my decoder order still hadn’t arrived, I decided to start work on my first wire-in motor decoder, to let me get some experience with the installation process, and to do some more experimentation. I used the Micro Ace Sobu E231 as previously planned, and to start with, a DZ125 decoder from Digitrax.

DCC Speed Tables

My plans to start installing decoders were somewhat upset when a large quantity of the ones I’d ordered turned out to be out-of-stock, and a box arrived containing only a couple of decoders and some wire. I actually have a number of decoders on hand, though not enough to do a full EMU the way I want, or all of the models I wanted to experiment with. So, while I could have made a start, instead I decided to spend some time working out my standard configuration settings for the Digitrax decoders. I’m going to have a number of these even if I don’t use the DZ125 wire-in decoders, since my Kato “DCC Friendly” models use Kato’s EM13, which is essentially a Digitrax FX3-Series decoder. And I have a few models with lightboard replacement Digitrax decoders.

Decoder Programming Prep

As noted last time, I’m going to (finally) install DCC decoders in some of my commuter (and other) trains that aren’t Kato “DCC Friendly” designs, meaning I have to use wire-in decoders. And since these are EMUs where the motor car is in the middle of the train, that means installing three decoders per train, a Motor Decoder and two “function-only” decoders for the cab cars.

But to start with, I need to set up my workspace since it’s been a while since the last decoder install, and the various elements had all moved off the bench to other uses. And the bench had filled up with important stuff (meaning junk I couldn’t stuff somewhere else), so I needed a better workspace. Besides, I’m going to want to sit down for this work, and the workbench is really better for standing work. Read More...

Wired DCC Decoders I

This is the first of what I expect will be several posts about wire-in DCC decoders. Up to this point I’ve either been using the Digitrax-made Kato decoders that snap into Kato trains, or lightboard-replacement decoders for locomotives. But I have a large number of trains that don’t take either of those, many of them the commuter trains I’ll want to run on my new commuter line (once I finish the DCC wiring for that). Read More...

NMRAnet - Why You Should Care

In August, the NMRA adopted standard S-9.7.1, NMRAnet Physical Layer, and a short article about it appears in the November issue of the member's magazine. What is this, and why should you care about it?

Well, if you care about Digital Command Control (DCC) for controlling a model railroad, it's an important addition to model railroading that will enhance that. And if you don't care about DCC, it's compatible with other control systems, and you may still want to use it. Read More...

In Search of the Perfect Post

One of the hardest lessons in model-railroading, at least for me, has been that “good enough” really is good enough. I spent fifteen years on my HO layout doing very little, in large part because what I did do fell short of what I’d set out to do, and I’d get frustrated and go do something else for six months. With Sumida Crossing, I started with the premise that I wasn’t trying to do a picture-perfect layout of the kind featured in magazines. Neither my skills nor my available time were up to that. Read More...

More Wires

After saying I was going to finish up the DCC electronics several times, I’ve finally made a start. There are two parts to this. The first is finishing up the remaining DCC power protection and occupancy detector panels. I did the first three tables last year, starting around January. The third was the first I’d done as a removable panel, which I wired up back in November. At the same time I’d begun the work of setting up a third accessory power bus, adding a switch and meter to the main panel. However at that point I’d stalled, with my attention off the layout over the holidays, and when I returned it was to work on buildings.

Bus Wiring and November 2011 Status

DCC is often said to simplify a model railroad because it requires “only two wires”. While that’s true to an extent, most real model railroads will require quite a bit more. Or maybe I just like to over-complicate things.

DCC Power II

The DCC Power work continues, but it’s still not done, as I have to do the power panels for the urban scene and the unsceniced return loops, but with the panel I built and tested back in September/October finally installed under the tables of the River Crossing scene, I’m at the halfway mark (having done the Riverside Station scene back in March).

To recap, the board contains a PM42 DCC circuit breaker that provides four separate circuit breakers and a BDL168 occupancy detector that can provide up to sixteen occupancy detectors and eight transponding sensors (see my pages on occupancy detection and the BDL168 for more details). This board provides occupancy detection for six electrical blocks, two each on the commuter tracks and one on each subway track where it loops below the expressway. Although I’m also wiring up the transponding sensors, as mentioned previously I’m having doubts about them due to problems in testing and the anticipated low power draw of my trains being borderline to register on them.

Fun with JMRI II and September 2011 Status

I’ve been playing around with JMRI some more, and trying to debug my transponding problem with the first of the electronics boards. This is really baffling. I checked the wiring, and it was fed through the RX sensor properly. I replaced BOTH the PM42 and the BDL168 circuit boards (I’ve got a stack of them waiting for more electronics boards once I get this one working) and I tried using other blocks. And I had more transponding sensor failures. On both sets of RX sensors. One defective set I might accept, but two?

So I tried a variety of things, and noticed that the non-functional detectors would, every once in a while, work. In fact, I discovered that with the train motionless, one of them would periodically cycle from detection to non-detection, emitting a LocoNet message reporting the change in status each time. I tried moving the wires. I pulled a fresh RX1 set out of a bag, and set it up atop a trash can (see above) with every wire fed through it fully separated from every other wire in mid-air (about the middle of this I was holding things in both hands and wishing I had a third arm). And that failed too, reliably as it were.

Fun with JMRI

Some weeks I don’t get much done. Well, that’s not quite true, I did a lot of work on the website this week, but the railroading time suffered. I did find some time to play with the test track I’d set up last week though, configuring JMRI to report block status based on the BDL168 outputs.

If you’re not familiar with JMRI it’s a free software package that allows a computer to interact with any of several DCC command stations that support computer interfaces, including Digitrax. It’s also available for Windows, Linux and Macs. And while not the most Mac-like program, it works, even on the ten-year-old iMac I’m using for the layout controller.

So what I have is a test track that’s an oval divided into two electrical blocks, each connected to an output of the BDL168. I named them, creatively enough, “One” and “Two”. And I used the Layout Panel editor to draw a schematic of the track. When displayed, red shows occupied (it’s configurable) and green shows clear. Even on my old computer, the change is nearly instantaneous once the locomotive crosses the insulated rail joiners from one block to the other.

DCC Power I

After assembling the first of my DCC protection and Occupancy Detection boards, I decided I wanted to test it and get some experience with using it. So I set up my loop of test track with insulated rail joiners separating it into two halves, and connected the feeder to outputs 1 and 2 of the BDL168, which correspond to RX4 A, detectors A & B. All of these are powered by PM42 section 1. For DCC, I used my Zephyr, and for the DC supply I used the 2 Amp, 12 Volt supply I plan to use for these systems (it’s the black box just above the Zephyr in the photo above). Powering it up, nothing went “Zap!”, which I count as a success.

Miscellany and March 2011 Status

March was another of those “not much obvious happened” months. I did manage to get the layout back together, with two of the circuit-breaker/block-occupancy-detector systems wired up. And I installed some lighting in the Subway Station as a test. But I still don’t have the track back together and operational (I’m waiting on some more DCC electronics on order). In the meantime, I’ve amused myself with several things and some work on the website as I plan my next moves.

First, I’ve taken more photographs of the Overhead Transit Station (photo above) and the associated platforms I’m using on the Urban Station scene, and updated my pages for it and for the Unitrack platforms. The photos were also added to the Stations photo album. Once I get the Riverside Station track operational, I’m going to be turning my attention to the Urban Station for a time (and as noted last time I’ve added a page about the Urban Station itself). I have some new track (Kato’s new V15 20-874 set and 20-875 single-track concrete-tie track) on order for that, about which more after it arrives. I really like the combination of the Overhead Station (and expansion for a second platform), the new platforms, and the V15 set; this makes for a really nice modern-looking station.

I’ve also done some more testing of DC power packs, checking out the behavior of pulsed power on motor temperatures (no effect that I could measure) and examining yet another power pack. The notes on both have been added to the DC Power Pack page. Photos were added to the Electronics photo album.

And I built another of Don’s LOLBoosters, and ran some tests for him. Not much to say on that, but I added a couple of photos and some text to my page on it.

How I learned to stop worrying and love the BDL168

Well, perhaps that’s a bit strong, but I’m coming to terms with its design flaws and the poor state of the documentation surrounding it and the PM42. I now have the first set of PM42 circuit breakers and BDL168 occupancy detectors (with two sets of RX4 “transponding receivers”) installed, and have done a bit more reading over the weekend. I’ve discovered a few things and come to a few conclusions as a result of that work, and that’s making me feel that I have a handle on this now. But I had to work through a number of issues to get to this point.

February 2011 Status - Occupancy Detection Revisited

Work has progressed slowly this month, partly from distractions, and partly because I’ve been reluctant to finish up the block occupancy detector wiring. I finally realized that the reason for this was that I wasn’t happy with my hybrid approach to occupancy detection and transponding.

To recap, my Subway and Commuter loop tracks were to be divided into blocks, with Digitrax BDL168 occupancy detectors and PM42 circuit breakers (circuit breakers are typically one per track per table, whereas there may be two, three or even four detectable track sections on a single track on one table, and more in a couple of cases). The PM42 provides for four circuit breakers, which is a nice fit for the four tracks, and the BDL168 is divided into four independent quadrants (so each can be wired to a separate circuit breaker), each with four block detectors. I’d originally planned one PM42/BDL168 per scene, meaning that wires would have to cross a table boundary in the Urban and Riverside Station scenes.

And that was a problem, for several reasons. First, running wires between tables violates my “keep all wires except bus wires local” design goal (it makes the layout harder to disassemble), second while the BDL168 can support 16 occupancy detectors, in some places I needed more than four on one track, which broke the association of the PM42 circuit breaker element to a single track, meaning a short would shut down a second line. And finally, I wanted to do Transponding, and the BDL168 only supports 8 transponding sensors (using a pair of RX4 sets), meaning some blocks would be able to report which train was in them, and some would only be able to report that some train was present, but not which. None of these were fatal flaws, but they were eating at me. And I finally realized that I only needed two more sets (seven instead of five) to fix these problems.

Building A DCC Booster

It’s been longer than usual between updates. That’s partly because other things kept me away from working on the layout or this website (the layout is still apart awaiting final assembly of the backdrops), but mostly it’s because I’ve been spending evenings the last couple of weeks building and testing a DCC booster kit. I didn’t actually need another booster, but Don offered me one and it seemed like a good way to gain some soldering experience (which it was, although I had a bit of a scare when it didn’t work).

DC Motors, Poles, and Skewed Windings, Oh My!

So, what is a “5 pole” motor, and why should I care if my models have one? And do they? That question occurred to me a few weeks back, and I’ve been doing a bit of reading since then. What I discovered was a mixture of fact and Internet lore of uncertain origin. I also discovered that some models do have “5 pole” motors, while others have “3 pole” ones. But oddly, at least in the case of Kato (and possibly also Micro Ace), the “3 pole” motors are the better ones. Or at least the newer ones, although I think they’re better too.

Kato Lightboard Flickering II

I seem to spend an inordinate amount of time worrying about light. First it was the layout lighting, then it was the flickering interior lights, which annoyed me when I detailed my first commuter EMU, and which I wrote about nearly two months ago when I did the original planning for a flicker-prevention circuit. Now I’m back, having built several prototypes and refined my design. I think I have a winning solution.

A Matter of Time

Railroads have always been concerned with time. Early ones used timetables alone to keep trains on the same track from colliding. This didn’t work very well, particularly given the accuracy of mid-1800‘s pocket watches and the lack of synchronized time sources, and many lives were lost. Signaling systems and other protection methods were gradually developed. But timetables continued to be important for all trains in scheduling the use of tracks even with other systems used for protection, and timetables were required for passenger trains, as trains from different places needed to coordinate their arrival and departure at interconnection points, so passengers could move smoothly from one to the next as part of a longer journey. Railroads gradually developed standards for time-keeping (and they’re responsible for the adoption of the standard time zones used in the U.S.), and also influenced the development of clocks and watches to provide accurate and synchronized time sources.

Japanese passenger trains today are famed for their obsessive adherence to schedules, with deviations measured in seconds, not minutes. So it only seems reasonable that a model railroad of a Japanese passenger line should operate to timetable (well, eventually,I need a yard/staging tracks before that will be much fun).

Freight Trains, Electronics and October 2010 Status

Not much got done on the layout itself in October, mostly I’ve been running trains (as documented in an earlier post with a video) and doing a bit of electrical work (mostly the previously noted update to the power panel). I’ve spent a good bit of time on a couple of other things though.

Adding a DCC Meter, And Related Topics

Although I wrote about the power panel previously, there were a couple of things left undone. I was missing one of the ammeters for the lighting power (the store I get them from had run out), and I’d misplaced one of my LNRP LocoNet Repeater panels, so I’d substituted a UP5 panel I took off the old HO layout. Also, I’d realized that my portable RRampMeter DCC volt/amp meter was useful, but wiring it into the system when I wanted to check something was a real nuisance. Over the summer I bought or found the missing bits, and finally decided to order a new RRampMeter (the version without the nice case) to mount as a permanent part of the circuitry (I’m keeping my other meter--the one with the case--for workbench and other testing uses). So it’s a good time to summarize what the panel does, and talk about the new meter a bit.

Car Lighting Power Protection and September 2010 Status

There isn’t much status to give for this month, as the only major layout event was finally completing the Rapid/Shinkansen loop and running DCC trains (a major milestone, but one I’ve already covered). But this month also marks the one-year anniversary of when I first started keeping this record (I didn’t actually get it online until November, but I was making offline entries and writing down design information from September 2009), and it’s worth a short look back on the year.

Detailing a Kato Commuter EMU

I haven’t done much work on the layout recently. Aside from having fun running trains, I’ve been working on finishing up my first DCC train. I started added DCC decoders (motor and cab) to this train last October.

Kato DCC Decoders and a Decoder Tester

Now that I have some real working track, I’m even more motivated to get some of my trains converted to DCC (so far there’s just one, the Jōban E231 described on my Adding DCC and Lights page). So, back I go to the hell that is Kato “DCC Friendly” decoder installation. Really, the marketing person who thought that phrase up was truly evil. There’s nothing friendly about this process. DCC doesn’t have to be hard, but with Kato it all too often is.

Tunnel Roofs, Castings and Track

Just a brief update: For the last month or so I’ve been working on the Subway roof that carries the Rapid/Shinkansen tracks along the front of the Riverside Station scene, above the subway tracks at the front of the table. That’s now been “finished” (WS inclines glued where needed, plaster and roadbed applied over that, all painted) and it’s supporting the tracks quite nicely. The decision to use 2mm sheet styrene turned out to be a good one. The track here is a bit hard to visualize (and some of it is missing in the above photo), see the track plan for a diagram.

Myths The Internet Told Me and August 2010 Status

The Internet is a wonderful invention. Without it, I wouldn’t be able to order trains from halfway around the globe and have them delivered in under a week, nor would I have any idea of the difference between an E233-1000 and an E233-2000 (both are commuter trains, but the 2000 is a narrow-bodied variant with a end door for emergency exits; this model runs through onto the Chiyoda subway and nobody makes a model of it yet, but I want one). I hear that some people have more serious uses for the Internet, as well.