Quote from: Doneldon on April 03, 2011, 07:35:06 PM

FOR ANYONE WHO'S READING THIS: Why are we so careful to call track switches "turnouts" and electrical controls for turnouts "switches" when we call switching layouts "switching layouts" but not "turnouting layouts?" Or am I just picking a fight?
                                                                                                                          -- D

A fair question and one that needs to be answered from time to time to remind new people about old terminology.

Talk to the railroaders who buy and lay full size track work.  They will tell you that a switch is part of a turnout.  A whole turnout consists of straight and curved stock rails (the outside rails that go right through), a frog (where the inside rails meet and cross,) guard rails (to prevent derailing at the frog,) point rails (that move side to side to direct traffic to one route or the other,) straight and curved closure rails (that connect between the point rails and the nearest end of the frog,) and extension rails (that connect to the other end of the frog.  The switch consists of the movable point rails, the switch rods (that join the point rails together to keep them in gauge and to make them move at the same time,) a mechanism of some sort for moving the points, and a throw rod to connect the mechanism to the point rails.

Now talk to train crew and ask how they send a train one way or another.  They will tell you they "throw the switch."  They do NOT throw the turnout!  It takes a large crane to lift a turnout, let alone move it, but a man can move the point rails using a switch stand, the levers in a switch tower, or an electric motor driven switch operator.  The latter can be activated to throw the switch from hundreds or even thousands of miles away.

Bottom line, the track appliance that comes in a neat package is properly called a turnout.  The part of it that you see move is properly called a switch.  When you are building your layout, you install turnouts.  When you run your layout, you throw switches.

Jim

There seems to be some confusion about what "28/128 Speed Step Control" really means and how this relates to using a Bachmann E-Z Command with 128 speed step decoders.  I hope to shed a little light on the situation.

Early decoders had only 14 speed steps.  The speed information was stored in half a register with headlight control stored in the other half of the same register.  This gave pretty big changes in speed when you changed your throttle from one step to the next.  Using more, smaller steps would smooth out the changes.  But this required an additional register in the decoder to store the longer speed commands and a switch to tell the decoder which register to look in and how to interpret what it found.  The first improvement was to double the number of speed steps to 28.  But having a whole register dedicated to speed commands allowed an even greater number of speed steps, such as 128 steps, to be stored.  The switch that we use to tell the decoder which type of speed steps it was dealing with is bit 1 in CV29.  When it is 0, the decoder uses 14 speed steps.  When it is one, the decoder uses 28 OR 128 speed steps.  So those are the two possible modes - 14 Speed Step Control and 28/128 Speed Step Control.

When Bachmann tells us that the E-Z Command can work with 28/128 Speed Step Control, it does NOT mean it can produce 128 speed steps.  It might make more sense if this mode were called "Extended Speed Step Control" or "Advanced Speed Step Control" because there would not have been any implication of the exact number of speed steps that it used.  Unfortunately, the term 28/128 Speed Step Control had already been adopted by others and Bachmann just followed suit.  Just like with that confusing term "DCC Ready."

Recently, there has been some controversy about just how many speed steps E-Z Command really does produce.  It was suggested that perhaps it produced 56 speed steps, the standard 28 steps plus 28 more intermediate steps generated by rapidly switching back and forth between two speeds.  I checked that this evening by connecting an E-Z Command to a decoder set to 128 speed steps.  With the decoder's output connected to a locomotive's motor and to a voltmeter, it was easy to check how many steps were generated when advancing the throttle from 0 to full throttle.  There were 25 steps.  Why not 28? probably because the last three got lost when the motor pulses became a continuous dc level at little early, at step 25 instead of step 28.  Unless the E-Z Command design has been recently modified, I suspect all E-Z Commands produce about 28 steps in 28/128 Speed Step Control Mode.

In another thread, I suggested that an E-Z Command could still be used to produce smooth, 128 speed step control by setting an acceleration rate in a 128 speed step decoder.  When the speed knob was turned, it would send a desired speed to the decoder and the decoder would accelerated to that speed at the 128 speed step rate.  Suppose for example you gave a locomotive 1/8 more throttle.  At 25 speed steps for the whole throttle, the locomotive's speed would increase in 25/8 = 3 big steps.  But with the decoder accelerating at the 128 speed step rate, the locomotive's speed would increase the same amount in 128/8 = 16 small steps.  The result would be a smoother acceleration.  After this evening's bench testing, it turned out I was partly right.  Using a Digitrax DZ123 decoder set for 128 speed steps, I measured the number of steps it took to accelerate from zero to full throttle.  Surprisingly, it took not 128 steps but 220 steps.  I have a strong suspicion this decoder was working internally at a 256 speed step rate but losing 10% or so of the steps at the top end as explained previously.  Even just using 220 steps, that 1/8 increase in throttle would increase the speed in 220/8 = 28 tiny steps.  In other words, very, very smooth acceleration, the sort of smoothness you would expect from a 256 speed step decoder.

Bottom line, these tests indicate that an easy way to get super smooth acceleration using your E-Z Command is to use a 128 speed step decoder with some acceleration programmed in.

Jim   

Had a great time this afternoon teaching a group of kids at a summer day camp how to turn their train sets into model railroads.  Of the 14 elementary school aged children in the group, 4 boys and 4 girls had train sets at home.  All of them, even the half dozen without train sets, were very enthusiastic.

We built our layout on a 4' x 6' sheet of foam which I had previously painted dandelion green.  The track went down quickly using E-Z Track, followed by a station and half a dozen false front buildings to make a frontier town.  Next we added trees, dozens of trees, poked into the foam, and people, lots of people in various poses but all dressed in 1890's clothing.

Finally, we got to the stars of the show - the trains.  We ran an 0n30 freight behind a 2-6-0 and had a passenger train on display as well.  Everybody ran the train, including a couple of kids who were at first reluctant but then clamoring to run it once they saw how easy it was.

All of this was done in an hour, but with 15 pairs of hands, nobody was rushed.  And if I may say so, the final product looked pretty good.  My favorite comment came from one young fellow who told me he could hardly wait to get home and get his train set out.  I suspect he was not the only one who suddenly saw his train set in a whole new light.

I am already looking forward to a repeat performance next month.

Jim

P.S.  The black based steel E-Z Track performed faultlessly, without a single stall.  This was after being stored for more than two years out in my garage.  It had been cleaned and lightly oiled before I used it the last time, but today it got no cleaning at all.   

This is a reponse to a comment by Robertj668 on another thread.  As it was getting off topic, I thought I should start a new thread.

I have been into atmosphere sounds for a lot longer than onboard sounds.  I started off using Innovative Train Technology's modules.  These are very easy to work with and can be used with a separate timer or a push button to produce the sounds intermittently.  Lately they have been producing some visual circuits as well (arc welder, etc.)  See their website at:
Innovative Train Technology

My next step was using CD players.  Many of these can be put on repeat and CDs play for so long that nobody realizes that the program repeats.  The CDs are cheap to make and can include lots of dead air time so that the sounds seem intermittent.  I have used both portable CD players and ones rescued from junk computers.  Many of the latter have two extra push buttons which allow them to play without needing a computer.  A whole bank of them can be powered by a single power supply, also rescued from a junk computer.  A pair of amplified computer speakers can serve two of these players.  Mostly I used mono recordings as the sound recorder built into Windows is mono.

The latest step has been using mp3 players.  These are available on sale for less than $10 for a 2 GB player which will play more than 24 hours without repeating and will do it on a cheap AAA battery.  Again amplified computer speakers are a good choice.  I stepped up a level by using "movie edit pro 14" to edit the sound effects and convert them to mp3 format.  This program is like Windows Move Maker on steroids and was on sale for $40 just before they introduced version 15.  For audio, it has 32 tracks, each with its own dynamic volume control and each with its own fader to send it to the left or right stereo channel, or anywhere in between.  Last spring I did a sound track for the farm on our group's large scale portable layout.  The left channel is household sounds and is fed to a speaker in the farm house.  The right channel is barn and barnyard sounds, everything from cows mooing  to the hired hand building a shed and the farmer trying unsuccessfully to start his old truck.  These sounds are fed to a speaker in the barn.  Between the barn and the house is a creek and a patch of trees.  The continuous babble of the creek is in stereo so that the sound appears to come from a point between the barn and the house.  Birds in the trees are intermittent and again in stereo  to put them in the woods.   The advantage of the multi track system is that you can manipulate the sounds separately and yet they can overlap in the final product.  Changes become very simple - too many cows?  Kill a few cow sounds in the cow track.  It is also easy to move sound bites in time (just drag and drop them) so that the horse give a horse laugh at exactly the right time after the farmer tries to start his truck and finally runs the battery dead.

This sound system was very effective at the train shows last spring.  It was not too loud - in fact, it was overwhelmed by the train sounds when the trains passed by (this was by careful choice, not lack of available volume.)  The intermittent nature of the sounds almost broke some necks as a canary started singing (or??)  just as someone was walking by.  And we knew the stereo was working properly when people were looking in the woods to see where the bird sounds were coming from.

I am thinking that I would like to do a web article on how this is done, both putting the hardware together and making up the sound tracks, perhaps even putting up some sound tracks to encourage people to try the hardware.  I don't know.  What do you think?

Jim   

On another thread, buzz asked about wooden grave markers for a old west cemetery.  Here are a couple, one for 0 scale, the other for H0 scale.  The first photo shows laser prints on bond paper glued to a Popsicle stick for 0 scale and to a wooden stir stick for H0.  The Canadian 1 cent is for size comparison.



At left in the second photos, the Preacher consoles Widow Brown in 0 scale while the H0 grave robber prepares to dig up sheep thief Kemp.  Jackson Brown's marker is 18" wide by 3' high with mostly 5 point lettering while Thomas Kemp's marker is 20" wide x 40" high with mostly 3 point letters.



Jim

Various people, on another thread, have already said what you can and cannot do with copyrighted material.  It all sounds very restrictive.  But we can learn to live within the rules.  Most websites have a contact email address somewhere on the site.  Sending an email to that address requesting permission to use images or text takes only a short time.  And usually, permission will be granted within a day or two.  In ten years of writing for the web, I don't remember a single refusal although I have had a few "failed to answer," which had to be treated as refusals.  I have never yet refused permission to any other authors to use any of my images or texts, although I prefer being asked for permission, if only as a way of keeping track of who to send updates to.  I think most of us who use the web to share information on our hobbies feel the same way, and are only to glad to say "go ahead" if you ask.

Here are some tips to make sure the answer is "yes" when you request the use of copyrighted materials:
-  Be polite.  You are, after all, asking for a favor, not demanding a right.
-  Be specific.  "I gotta use one of your pictures" just won't cut it.  If it is an image, give the exact url for that image, not just the url for the page that it is on.  If it is text, give the page url and quote the entire passage you wish to use (use cut an paste so there are no mistakes or misunderstandings.)  If it is an entire article, then give at least the url of the title page and the title exactly as it appears on that page.  "I want to quote your article on couplers" probably will get you no response at all, if the author has written several articles on couplers but "May I use your article titled 'Using E-Z Mate couplers on Older Rolling Stock'  starting on page http://www.cplrsrus.com/EZ-old-cars/index.html and ending on page http://www.cplrsrus.com/EZ-old-cars/page7.html.'; is much more likely to get you a positive answer because the author knows exactly what you want.
-  Say why and how you want to use the requested materials.  Again, be specific.  Are you using the material in your own website article?  Or for illustrating a posting on a public forum?  Or??? And do you want to hot link to an image or upload it to your own web server or to a public image storage site?  For what period of time do you wish to use the material?  A month?  A year?  Indefinitely?
-  Credit the material, usually in the form "photo by S.Smith, used by permission"

Done right, the answer is pretty sure to be "yes"

Jim

Take a look at one man's model railroad that doesn't take up any floor space and doesn't have the view blocked by the tracks like many shelf layouts mounted high on the walls.
http://news.bbc.co.uk/2/hi/uk_news/england/7694255.stm

Of course, if space is no problem, you could build a layout like this one:
http://news.bbc.co.uk/2/hi/europe/7962573.stm

The greatest model railroading experience for a lot of us is going to a train show - as participants.  This past weekend, our little model railroading group set up our indoor large scale layout and ran trains, trains and more trains.  Ours was only one of a dozen or more layouts at a local show.  For us, it was a great opportunity to run our "summer trains" in winter, as none of us has a 20 x 50 indoor space to set up the layout except at shows.  But the best part of the show was not running the trains or rubbing elbows with other participants or the chance to buy/sell new and used equipment.  The best part was the thousands of smiling faces of the people enjoying the show.  We talked and talked and talked some more with model railroaders, potential model railroaders, real world railroaders, and people who just liked to watch trains.  It was great to be able to give back some of the pleasure that model railroading has given us over the years.  We are already looking forward to next year.

A previous discussion on powering the E-Z Command from either its own transformer or from a different transformer via a dc power pack left open the question of what would happen if you used both transformers.  Analysis of the partial schematic of the E-Z Command that I drew up some time ago and of the schematic of the 44212 dc power pack that Joe Satnik drew up recently led me to believe that if both transformer were used, the E-Z Command would preferentially draw power from its own transformer.  The transformer connected to the 44212 would supply only a small amount of power to  the 44212 to allow it to send a dc signal to the E-Z Command.  Tests just performed confirmed exactly this.

I used a 30 ohm, 25 watt precision power resistor as a load and used a clamp on ammeter to confirm which transformer was supplying the current consumed by the load.  When the E-Z Command was used alone, its transformer supplied the power.  When the E-Z Command was used without its transformer but was connected to the 44212 and its transformer, the 44212's transformer supplied the power.  And when both transformers were used and the 44212 was connected to the E-Z Command, only the transformer connected to the E-Z Command supplied power to the load.

The significance of these findings is that if you want to use your 44212 power pack to power turnouts, building lights, and other accessories without robbing power from your tracks, but you still want that extra control and/or extra DCC address, you can do so by using the two transformers.  Which of your two transformers powers the 44212 and which powers your E-Z Command does not matter - they are identical.  The E-Z Command manual does not say you can do this.  More significantly, it does not warn you not to.  Perhaps they don't tell you to do this because there is one idiosyncrasy that you need to live with.  Once you have connected your E-Z Command to your tracks and to your 44212, you are left with the question of which to power up first - your 44212 or your E-Z Command.  It turns out that it does make a difference.  If you power up the 44212 first, it will automatically take over control of analogue locomotives and free up address 10 for DCC locomotives.  But if you power up the E-Z Command first, then it will not relinquish address 10 to the 44212.  This means address 10 will still control analogue locomotives and not be available as an extra DCC address.  And it means the 44212 will just be along for the ride - it will not control anything.  I cannot guarantee that your setup will cycle exactly the way mine did but it makes sense that something in the power-up sequence has to decide how to use address 10, and most likely that something is the microprocessor or PIC controller that controls just about everything in your E-Z Command.

While doing the above tests, I was also able to confirm that both methods (address 10 or using a 44212 power pack) used zero stretching to power analogue dc locomotives. 

Rich recently posted photos of DCC waveforms but was unable to capture the zero stretching that makes it possible to run dc locomotives on DCC.  Nor was I until recently given a storage oscilloscope which is capable of capturing the signal and displaying a single, non-repetitive view of it.  I have put together a simple web page on the Saskatoon Railroad Modellers website which can be reached by clicking on the link below:

http://members.shaw.ca/sask.rail/dcc/DCC-waveforms/DCC_waveforms.html

Just finished a trip on the Rocky Mountaineer and had to buy the model.  The Locomotive appears to be made by Athearn.  Anybody know who makes the coaches? 

Attempting to use a Digitrax DZ125 decoder to operate 0n30 passenger car lights gave strange results.  The lights intermittently turned off and would turn on again only when I shut down the E-Z Command operating them.

I connect two strings, each consisting of four 3.3 volt, .050 amp lamps in series, to the motor output of the DZ125.  My idea was to have a light dimmer to allow adjusting the lighting to suit either my E-Z Command or my Digitrax Zephyr under various ambient lighting conditions as can be found at various train shows.

With the two strings of lamps in parallel, the total current was only .1 amps, or 10% of the maximum allowed current for the motor output of a DZ125 decoder.  But this was causing the decoder to heat up and shut down from thermal overload.  Shut down occurred for both the forward and reverse directions of the decoder and for a variety of lamp brightnesses.  Significantly, shut down did not occur when I replaced the E-Z Command with a Digitrax Zephyr.

Using a high quality moving coil voltmeter and correcting for square wave versus sine wave output, I measured the output of the E-Z Command as 19 volts peak (= 19 volts r.m.s.)  I later confirmed this with an oscilloscope.  While this seemed a tad high, it is less than the 22 volts peak maximum voltage specified by NMRA in their Standard S9.1C.  Standard S9.1C also calls for the maximum voltage to be within 2 volts of the r.m.s. maximum voltage for the scale as specified in Standard S9 but unfortunately the latter specifies only a minimum voltage of 12 volts for all scales (not too helpful here.)  So 19 volts seems to be acceptable under NMRA Standards as presently defined.

Turning to the decoder, Digitrax specifies their DZ125 as a 20 volt decoder.  I suspect this is the voltage at which the four transistors in the H-bridge motor output circuit start to conduct current even though they are nominally turned off.  Why Digitrax would use transistors with this low a rating I do not know but it could be because they are smaller physically or it could be that high voltage equivalents are hard to get because of the war.  In any event, it appears that the 19 volts output of the E-Z Command is close enough to the transistor breakdown voltage that the inevitable transients cause enough extraneous conduction to over heat the decoder. 

I have not as yet tested additional DZ125 decoders with the E-Z Command but plan to do so.   In the meantime, I would recommend caution if you are planning to use DZ125 decoders with the E-Z Command.  Either pretest the decoders before installation or reduce the output of your E-Z Command by connecting two or three 1N5300 diodes in series in one direction and paralleling this with the same number of 1N5300s in series in the other direction.  Then put the whole mess in series with one track lead of your E-Z Command.

Simply put, the more pickup wheels, the lower the odds of losing contact with the rails.  Just for fun, lets play with some numbers.

Suppose the electrical pickup from any given wheel fails only 10% of the time.  Failure could be from dirty track, dirty wheel, dirty wiper, or any other reason, but the other 90% of the time each wheel picks up just fine.

Now suppose that we built a locomotive with only two pickup wheels, one on each side.  Then the chances of both wheels picking up power at any given instant in time is 90% x 90% = 81%.  That in turn means the chances of failure to pick up power at any instant in time is
100% - 81% = 19%.

If we double the number of pickup wheels to 4 (two on each side) then for each side, the chance of failure is 10% x 10% = 1%.  So the chances that both sides will pick up is 99% x 99% = 98%.  The chances of no pickup is 2%.

Continuing in this vein, we can develop a table for number of pickup axles versus failure to pick up.  First lets do it for 90% pickup reliability per wheel.

number of axles     failure, % of time

     1                             19%
     2                               2%
     3                               .2%
     4                               .02%
     5                                .002%
     6                                .0002%

A 2-8-0 with tender pickup is an example of a locomotive with six axles of pickup, four on the locomotive and 4 half-axles of pickup on the tender (they only pick up at one end.)

Ninety percent reliability per wheel is probably about right for clean wheels on clean track with clean, well adjusted wheel wipers.  But suppose both wheels and track are moderately dirty.  Then 50% reliability might be nearer the case.  Then the chances of failure for different numbers of axles would be something like this:

number of axles    failure, % of time

     1                                75%
     2                                43%
     3                                23%
     4                                12%
     5                                 6%
     6                                 3%

Now something amazing comes out of this.  Dropping the pickup reliability per wheel from 90% to 50% increases the failure rate by 15,000 times!  That is, the chance of failure to pick up power at any moment in time rises from .0002% to 3%.

The other day, I pulled an old coach off the shelf and ran it.  The interior lighting had only one pick up on each side and the lights flickered terribly.  I would guestimate they were on for only 10% of the time.  I wondered how much better they would be if I had pickup from 3 wheels on each side or maybe even six wheels on each side and calculated the following table for 32% reliable pickup from each wheel.

number of wheels       failure, % of time
on each side

     1                                 90%
     2                                 71%
     3                                 32%
     6                                 .2%

From this I concluded I would need pickups on each and every wheel, not just each and every axle with the usual one pickup wheel per axle arrangement if I wanted excellent pickup.  Alternately, I would have to add some anti-flicker circuitry and/or better pickups.

I've been playing around with some hand laid track, even though the track plan for my proposed 4 x 8 portable is still being developed.  I thought I would start with a stub switch.



The turnout has an all metal frog and is built from Atlas Code 100 Nickel Silver rail on cedar ties.  It is powered by an old Tortoise machine that had previously been repaired after it wore out in H0 service and had then been modified for use in G-scale.

I used cedar for the ties because I like the way it weathers with ferric acetate.



Because this type of turnout relies on "bending the iron," I used some Atlas H0 ties to serve as tie bars between the moving rails.  I may yet replace them with styrene rods with brass wire glued into the ends and soldered to the rails - it all depends on how they look once the turnout is ballasted.  That will have to wait until I finish building a "ballast grinder" to crunch up kitty litter to the right size.  (Note to Joe Satnik - no danger of cats visiting my house - my dogs won't allow it.)

I did a lot of hemming and hawing before cutting the ties.  In the end, I decided to cut them 5/32" wide by 1/4" high by 1-1/2" long.  In scale, that works out to 7-1/2" wide by 6' long, which may be a tad short.  The 1/4" height has no real scale meaning as they will be buried in ballast.  But that height serves two purposes - it means the spikes do not have to be driven into the Baltic Birch plywood under the ties and it makes the track a perfect match for Bachmann E-Z track.  Hey, you didn't think I was going to hand lay ALL the track, did you?  Where it is hidden in tunnels and a helix, E-Z track will be my E-Z way out.  The last photo shows a transition.



The ties sure look a long ways apart for a mainliner's eyes, but I laid them at 2640 to the scale mile, which I suspect is about right.

While waiting for some 0n30 rolling stock to arrive, I thought I would see if I could still scratch build in 0-scale.  This is only the second building I've done in 0-scale since 1955 but it was a whole lot of fun.  The board-on-board siding is 1/4" wide stir sticks over a Baltic Birch plywood frame.  I don't know what type of wood these stir stick are made of but it is a delight to work with and it reacts well to iron acetate aging treatment.  The muntins had me puzzled for a while as I usually use styrene solvent welded to polycarbonate glazing.  The solution was narrow strips of birch veneer, the iron-on kind with factory applied hot glue on the back.  It stuck well enough to the window glass and I am pretty sure it is going to stay there.  While this building is only 16 ' square (4" x 4") I think several of these plus some half tents and perhaps one or two more substantial buildings would be a good start on an early town.

The following error or errors occurred while posting this message:Your session timed out while posting. Please try to re-submit your message.

Maybe I am typing too slow or spending too much time worrying about correcting my spelling and grammar, or making sure my facts are correct, and generally putting all my ducks in a row, but I seem to see the above message pretty often these days.  I know, thanks to Gene, that I can hit preview again and then post, but I am still left with the nagging suspicion that I am taking up too much of the Board's time.  Is anyone else troubled with this?

I don't often visit the ON30 section of Bach Man's Board because I don't (or rather didn't) have anything in ON30.  But after stumbling in by mistake and then compounding my error by staying to look at the photos of the 4-4-0, I now own one.  Wooden cab, oil headlight, small drivers and all.  The wife says I don't need another railroad but I can feel an 1890's period piece coming on.  All I can say to her is "The Devil made me do it!" 

Signals can add visual interest to the layout as well as providing information to the operators.  The information provided can be anything from indicating turnout positions to a full blown signal system mimicking the signals on your favourite railroad.



The link immediately below will take you to the article on building the signal masts and heads like the ones seen above.

http://members.shaw.ca/the.trainman/LV_Workshop/signals/plastic/index.html

while this next link will take you to a preview of other signal articles presently being written.

http://members.shaw.ca/the.trainman/LV_Workshop/signals/index.html

Does anyone remember what colours oil drums were?  I am talking about the steel 45 gallon drums (55 gallon drums in the US.)  I remember Imperial Oil painted theirs red with the middle third painted white.  But I cannot remember the colour of the ends.  I think maybe the contents were stencilled in red on a white end (or was it stencilled in white on a red end??)  I also seem to remember green drums with a white centre stripe, and think they might have been Texaco.  But I am hoping someone has a better memory or can direct me to a website that lists the colours of steel barrels.

After Ten Wheeler's problem with a noisy 4-4-0 http://www.bachmanntrains.com/home-usa/board/index.php/topic,1234.0.html
I thought I would investigate the noise in my Bachmann DCC-on-board GP-40.  As received, it growled at all throttle settings above a dead stop.  I did not find the growling all that disturbing as it was quieter than the side winders and other coffee grinders of a couple of decades ago.  But some of my cohorts were nervous about this locomotive which, to them, sounded on the point of self destruction.

Removing the shell revealed a cast metal frame with the motor sandwiched between upper and lower sections of frame.  The motor was slightly loose in its mountings because the upper section of the frame was slightly warped.  But it was not so loose that it could rotate and touch the plastic shell.  It appeared, however, that it could rotate enough to trap the decoder motor wires between the motor and the plastic shell, possibly transmitting motor noise to the shell via the trapped wires.

One solution might have been to unsolder the wiring, remove the decoder, remove the top section of the frame, and then try to straighten the latter without breaking it.  But being lazy, I tried something easier first.  I relocated the motor wires so that they ran between the bottom of the motor and the bottom section of the frame in the area between the brush holders and the farther end of the motor.  The plastic end of the motor is narrower than its metal case, so there is more room to run the wires vertically up past the motor to the decoder printed circuit board.  When the body was snapped back on, the locomotive was quiet, to the point that my (slightly deaf) ears can no longer hear it run.  I hope my model railroading buddies will find it the same.

This is a classic case of a noise (in this case, motor noise originating from Bachmann's non-silent decoder) being coupled to and amplified by the plastic shell of the locomotive.  If you can break the coupling (in this case by moving a couple of wires) then the sound is no longer amplified by the shell.