BR-38 - FLEISCHMANN 716084
03/17/14, 10/09/15- This little 4-6-0 steamer closely
resembles the loco that came with the Minitrix Orient Express set that
I bought in the '80s, down to the olive and red livery. Even by today's
standards, that was an exquisite set with lighted cars, and the loco was
one of my best-running DC engines. While I didn't mind updating the cars
with flicker-free LED lighting, I was hesitant to touch the loco. It ran
under DC compatibility on my track, but made an annoying buzzing while
at rest-- something you wouldn't want to subject it to for any length
of time, since the sound is the result of the motor jittering from AC
power. I suspected that it would be a difficult DCC conversion and it
probably wouldn't run as well as a modern product.
My experiences with the Streamlined BR-01 and the Crocodile convinced me to look for locos with
a factory fitted decoder; presumably, that would ensure a good-running,
trouble-free loco out of the box (you'd think since these things are so
expensive!). The BR 38 didn't have sound, but it was from Era II and visually
was a near-perfect replacement for the OE loco. Best of all, it was available,
despite being released in 2011. Out of the box, it ran perfectly... at
Trouble in Paradise: I started to notice little quirks, like sometimes going in reverse for
a short distance on power-up. It would sometimes briefly speed up while
traveling at a slow speed. I attributed it to the load control feature,
but the documentation was sparse and didn't show how that feature could
be turned off. In fact, I didn't know who made the decoder or where to
look for additional details. Over time, the symptoms got worse so frustration
drove me to replace the decoder with an ESU Lokpilot. By this time, I'd
bitten the bullet and bought ESU's Lokprogrammer for other projects so
I knew I could fine-tune its performance.
Replacing the decoder initially solved the problems, although not without
some difficulties. This is a tender-driven loco, with the decoder hidden
in the loco's driver cab. The wires are neatly snaked through the hollow
tender/loco linkage (very nice!). In the tender, the motor is connected
via a small ciip-mounted circuit board with a resistor and coil, presumably
to supress RF noise emissions. I'd read that they weren't necessary for
DCC operation, so I cut the trace with an Exacto blade, right before the
coil. I thought I'd cut it completely, but maybe not. It ran well for
a while, but then it would inexplicably stop and not start up again. Sometimes
it would work only after re-situating the loco on the track, or cycling
the power, or taking the cover off the tender and giving the armature
a nudge. I didn't know what was going on. As usual, I started troubleshooting
at the wrong end, at the decoder in the cab. I thought that it might be
shorting since I'd had to solder the decoder in (there was no socket).
I spent quite a while going over the wiring, before ending up at the tender.
This time, I clipped one end of the coil to ensure that it was out of
circuit. The loco wouldn't run at all. From this I deduced that maybe
the circuit board routed power to the motor after the coil, not before
it from the large pads where the decoder wires were soldered? I'd never
removed the board to see if there was anything else underneath. The easiest
thing to do was to resolder the snipped coil and reflow solder over the
trace that I thought I'd cut. After this, the loco ran flawlessly for
a while, until the intermittency thing started happening again. Popping
the hood again, I wondered whether the circuit board that was clipped
onto the motor might be responsible? Surely, they wouldn't have designed
it for unreliability, would they? Maybe the circuitboard was for filtering,
required by law, and kludged onto the existing design so the board could
be easily removed by the consumer? Pure speculation. I soldered the clips
to the motor and it's worked flawlessly ever since (knock on wood).
[11/2015- ...until I had to desolder the clips to remove the motor to
give it an ultrasonic cleaning. Commutator grunge or the Digitrax Zephyr DCC controller needing to be reset may have been a factor
in the original symptoms. As I said, troubleshooting this stuff can be
Diagnosis, a Wild-Assed Guess: The original decoder was an older design,
hard-wired in. Replacing it with an ESU LokPilot was probably a good idea, even if it wasn't the problem.
Cutting the trace to the coil should have killed power to the motor, but I may not have completely cut the trace. At the time, I thought the clip-on
circuitboard was being jarred by the train's movement and didn't
reliably supply power to the motor. Therefore, I soldered the leads to
the motor (Having since seen many non-soldered motor connections in locomotives, I think they're probably not a common source of failure). I left the coil and resistor
since even if they weren't necessary for DCC, they didn't seem to hurt
performance (Nowadays for a wired installation, I like to grind off everything
except SMD resistors that power LEDs).
The main takeaway lesson is that troubleshooting and diagnosis benefit greatly from experience. There are so many different things that can cause odd symptoms, and you shouldn't rule out anything. Troubleshooting rule #1 applies: Start with the stuff that you can easily test (like resetting the DCC controller) before doing anything drastic. Cause and effect observations can be deceiving, so be skeptical until
you're sure of the correlation.
Electric trains are simple devices and wiring a decoder is actually
very straight-forward if you do it point-to-point directly to the parts.
Most circuit boards are used to help with production. Some may contain
electronics to add DC features; for DCC, special features are more logically
controlled by the decoder. However, using the manufacturer's circuit board
is often desirable since it's fitted to the space, may keep dust out,
or may have a light mounted exactly where it needs to be. If you want
to re-plumb using a fitted circuit board, examine the board completely
so you have an idea of what it's for, and where the wires and traces are
BR 38 Sound Conversion: This was my third sound installation
using an ESU Loksound decoder. By this point, I'd pretty much been "lok'd"
(yuk, yuk) into the ESU line of decoders since I'd bought their Lokprogrammer
that allowed me to program their decoders for motion and sound. It's a
great system, even for something as mundane as changing CVs since it's
done through a program on computer. Writing data to the decoder updates
the firmware as well, and you can download sound and decoder files for
specific locos, change sounds, and program how they operate. Very cool.
Installing a sound decoder and speaker in the tiny BR 38 would be a
very difficult project. Maybe not impossible for a talented tinkerer,
but there's an easier way to bring sound to an n-scale train: The sound
car. The concept is simple: Use another car to gain more space for the
installation of the stuff that would be difficult to fit in the locomotive.
This loosens up the size restrictions for components, and depending on
the car chosen, may make the job quick and simple. Of course, it does
mean that the "loco unit" has an additional car, and if you put the speaker
in the car, the sound won't be coming from where it should be coming from.
Only the most discerning would probably care since it's close enough and
not really noticeable at a distance. (By the same reasoning, sound shouldn't
be coming from tender, either.)
After some experimentation, I'd decided on my "formula": The decoder,
speaker, and dummy motor (more on this later) are installed in the car.
Instead of using a "keep alive capacitor", I connect track power between
the loco and car via a two-conductor plug. I felt that this was a better
solution since it provides more points of track power contact for both
the loco and the sound decoder, and the pickup points are spread out over
a longer area of track. As a side benefit, this also lets you coat a power-conducting
wheel with "Bullfrog Snot" to improve traction (if you need to) without
significantly affecting the power pickup (more power-conducting wheels
means that you can probably sacrifice one or two). I found this to be
a great solution on another tender-driven loco with dummy driver wheels
that turned sporadically.
(Note that this solves the problem of cars with unreliable track power pick up from the trucks. If the loco drives fine and the sound car doesn't glitch, it's not necessary.)
The downside is that you need a small plug and socket between the tender
and car if you want to work on them separately-- for this, closely-coupled
cars is not a desirable thing. Even the smallest plug is pretty
big at n-scale, and may cause derailment problems at sharp curves if it
binds between the two cars. However, if there's enough daylight between
the cars and the connector is oriented with this in mind, this problem
can be eliminated. (A side benefit is that you don't need to worry about
the car decoupling since it's tethered by the wires.) However, there's
no getting around the fact that it's ugly. This would likely be a deal-breaker
for most since Rapido couplers are so reviled. (Most of the European stuff
comes with Rapidos; they work well enough so they've never bothered me.)
Of course, you could wire the tender and car together without plugs, like
most tenders connect to engines. Power-conducting couplers are used on
Tomix Shinkansen sets and are used mainly for special situations where
on-the-fly decoupling isn't needed.
For steam locos, the Loksound decoder needs a load to trigger the steam
chuffs. I had some crummy/dead Bachmann locos from the '80s, so I used
their motors after cutting off the shafts. Through the LokProgrammer,
the decoder can be set so that the voltage to the motor is never high
enough to spin the motor, but yet trigger chuffs corresponding to speed
changes and triggering brake sounds.
(Prior to this conversion, I'd tried a high ohm resistor to serve as
the load but it didn't work, which is why I used motors for my first or
four sound decoders. I later learned that a ~200 ohm 1/2-watt resistor
could be used as the load. Much easier to fit in the space.)
This loco was going to be used with the Orient Express cars so the baggage
car seemed like a good choice. It had few windows and wouldn't look odd
if it weren't lighted. It also had the simplest construction, with only
the floor plate (with power-conducting trucks) and the body, with no detailed
interior. Besides, there were two of them in the set so I didn't mind
defiling one in case things got ugly.
This was an easy and quick installation and didn't require much defiling.
Despite it being a fairly narrow car, installing the dummy motor required
scraping off only a hair's worth of the clear plastic used for the windows.
The decoder easily fit with plenty of room, and if I'd been up for it,
I could have easily added LED lighting.
The speaker included with the Loksound was too wide to fit, even removed
from its baffle. I would have needed to remove a milimeter from each side
to make it fit, even without the window plastic in place.
Sugarcube Speakers: I'd read good things about Sugarcube speakers,
so I bought one when I ordered the decoder. The dimensions seemed too
good to be true, or at least too good to produce halfway decent sound.
When it arrived, I didn't have a clue how to use it. The speaker looked
like a metal rectangle, and I deduced that the side with the printing
was the back. The two clip things on the sides were probably for the speaker
wires. The black baffle ("resonator") didn't appear to be designed
to fit on the backside like a conventional speaker. After some more reading,
I learned that it needed to be glued to the front, with attention paid
to making sure that it was sealed completely, preferably with a flexible
compound. I used Pliobond, which is a rubbery shoe glue/contact cement
that adheres well to metal.
I tested it before mounting it and was surprised at how loud it was!
What surprised me even more was how bassy/un-trebly it was, especially
compared to the two other installations I'd done using conventional speakers.
I expected shrill and trebly, especially since I tested it outside the
body. I'd configured the decoder with a sound set I hadn't used before,
but on computer the sample I'd heard didn't sound much different from
Definitely alien technology. I was completely at a loss about mounting
it since it didn't seem to be very directional. I'd pulled window panes
out of the top of the car to let the sound out, but that didn't seem to
matter. The sound appears to be radiating from the entire car and not
louder or crisper from the front or rear where the openings face. Weird.
(Weird, but in a good way!)