Musings on sound decoders, A.K.A. "Silence is not always golden"

I am a somewhat regular participant on a certain prominent model railroad manufacturer’s forum. Said forum is notorious for having a few key players debate the finer points of a straw man argument for literally days on end until the average reader/viewer is nearly driven to breaking his or her computer keyboard across someone’s teeth just for stress relief. One of the more common arguments that resurfaces every couple of months, usually when some unwitting newbee to the forum asks a simple question, pertains to the proliferation of sound decoders.

One camp has its foot steadfastly in the “silence is golden” mantra. Members of this demographic tend, more often than not, to be older fellows who have been in the hobby for decades. They tend to vehemently decry the use of sound decoders in model locomotives, usually citing the perceived “poor” quality of sound. They maintain that the sound produced by the latest crop of sound decoders, both aftermarket and factory installed, sound nothing like the real thing, especially in the lower midrange and bass frequencies. Not even once do they consider the practicality of such a concept (or the fallacy of their argument); if it were even possible to wrangle 120 decibels of EMD 645 prime mover sound out of that tiny 2” speaker, what would it do for the dishes in your cupboard? Or for your relationship with your neighbors (and the local police force, for that matter)??? 

My response to the “silence is golden” camp is as follows: Don’t talk to me about a ‘lack of realism’ when your trains are all dead silent. Mediocre sound is better than no sound, I say, without apology. When was the last time you heard the prototype glide quietly by? Especially at a grade crossing.  I’ll up the ante and state that you CAN’T operate your trains in a prototypical manner without sound. (How’s that for throwing gasoline on the fire? Tee hee.)

The second camp is quite the opposite of the first, they love the idea of a sound decoder in a model locomotive and they can’t get enough of it. The lions’ share of their roster, or sometimes even their entire roster will be equipped with sound decoders, usually of one specific brand. These are the folks who usually start the Soundtraxx Tsunami versus Atlas/QSI portion of the ‘Legendary Recurring Sound Decoder Argument’….. “Tsunami is the greatest, the Atlas/QSI systems sound like washing machines” one of the more prominent members of this camp likes to say. Yet another member of this camp recently remarked that the QSI decoders sound like “a UFO from a 1950’s science fiction film” in comparison to the Soundtraxx decoders.

Now, I’ve spent a sizeable portion of time listening to the Soundtraxx Tsunami decoders. They certainly do SOUND great, and more realistic than the competing Atlas/QSI product. I’m perfectly willing to concede that. 

BUT, can you make them operate realistically? That is the 64,000 dollar question, and this is usually where the humming and hawing from the pro-Tsunami camp begins.  Cruise on over to Youtube, enter “Soundtraxx Tsunami” into the search box, and count how many videos you watch of a model locomotive making jackrabbit starts and stops. Last I checked, there were a lot more videos of that variety versus videos of Tsunami-equipped locomotives starting out smoothly and stopping in a slow and heavy manner, realistically simulating the behavior of a prototype locomotive.

Or, to state my argument to this camp another way, “What good is top-quality sound if I can’t OPERATE it realistically?” Sure, it can be done, I’ve seen at least one guy manage to program his Tsunami's for realistic momentum effects, but given the relatively rare occurrence with which this seems to actually be accomplished, makes me question just how complicated it really is.

Consider this. I can easily coax very realistic operation out of the Atlas/QSI-equipped Gold locomotives with 2 minutes worth of adjusting CV’s and using my Lenz 100 DCC system.  The first thing I do whenever I purchase a new Gold locomotive is to place it on the track, fire up my Lenz 100, and adjust CV’s 3, 4, 23, and 24 to some non-zero value depending on how much friction is in the locomotive mechanism. Some locos are maxed out at 255, whereas some (usually MP15DC’s, for those of you keeping score) are in the 75 to 80 range. I then program the roadnumber into the locomotive using the CV 17 and CV 18 workaround as detailed in the Atlas manual. Then I set up the Lenz handheld to recognize 128 speed steps with that particular locomotive. Using the large throttle buttons on the Lenz handheld steps up or decreases the speed in 16 step increments. Doing the math we find that 128 divided by 16 is 8, or eight throttle notches, plus an Idle (0 speed step) notch.  In other words, just like a prototype locomotive. Using the large throttle buttons in conjunction with the modified momentum settings, I get realistic drift and acceleration, along with a realistic speed curve.

In closing, my locomotive fleet is about 50 percent sound decoders, with all but one (an Athearn SD45-2) being of the Atlas/QSI variety. I standardized on these units not because I think the sound is the greatest, but because they represent a middle ground in acceptable sound and ease of programming/operation with my Lenz 100 system.

Sometimes it pays to consider an argument from another angle...

Modeling roadbed and superelevation

I like to model the different heights of roadbed found on the prototype. Full-scale railroads build sidings and spurs lower than the mainline to prevent errant rolling stock from creeping out onto the main in front of an approaching train. This is how the mainline got the nickname “High Iron” back in the day. Sidings are often 10 to 12 inches lower than the main, whereas spurs can be several feet lower, depending on their location.

I use Midwest Products cork roadbed to model this practice. I prefer the Midwest over other brands because it is more pliable, has finer grain, and thus is easier to shape.  My mainlines use two layers of cork; a layer of HO cork on top of a layer of O gauge cork. The two layers are the same thickness, but the O gauge layer is wider; this gives a realistic “two-step” appearance to the mainline roadbed cross-section. I use white glue to secure the roadbed to the foam top layer of the layout.

Branchlines and sidings get one layer of HO scale roadbed. The most important part of this whole method lies in ramping down the top layer of the mainline into the single layer used on the siding. To do this, I lay the top layer of cork 14 inches past where the end of the turnout is located, into the siding. Once the glue has dried overnight, I use a Surform rasp (available at Lowe’s or Home Depot) to feather a downgrade into the siding. I work slowly and carefully, using a 12” steel ruler to ensure that I’m not introducing any vertical curves into the incline. The first four inches out of the turnout is dead level, while the remaining 10 inches ramps down to the base layer of cork. Given that the roadbed is 3/16” thick…Quick! Do the math!…falling 3/16” of an inch over 10 inches results in a nominal 2 percent grade, which is perfectly acceptable.  

The same general procedure is used for spurs, with the caveat that since N scale roadbed is used on spurs, the mainline roadbed has to ramp down slightly steeper since the N scale roadbed is only 1/8” thick.

“Hold on a minute” you say. “N scale roadbed isn’t wide enough to give a sloped cross-section past the ends of the ties in HO scale…”

It does when you backfill with ballast, shaping a new shoulder as you go. And the N scale roadbed is just the right thickness to imply a slight elevation above the surrounding ground.  

When I have finished shaping all the ramps, I like to paint the roadbed with light gray latex interior house paint. This accomplishes two things; (1) it seals the roadbed and makes it nearly indestructible to any moisture introduced during the ballasting process and (2) it helps disguise any areas with thin ballast coverage.   

In closing, a quick word about superelevation. I use a 30” minimum radius on mainlines and I like to superelevate. To do this, I superglue a strip of .030” x .030” styrene under the outside rail from easement to easement. I don’t bother fashioning transition ramps for the superelevation, I just let the track float for a short distance over the easement. It is fully supported once the track is ballasted.

Troubleshooting Micro Engineering turnouts

I really like the look of Micro Engineering’s turnouts. They have a smaller, more to-scale appearance around the guardrails and frog area; with prototypical tie size and spacing, including the headblocks. They also have a lead (the distance from points to frog) and overall length set up for 2”-on-center yard track spacing out of the box with no trimming or modification necessary. This is more than can be said for most other brands of HO scale turnouts. 

A while back, a post on a certain prominent model railroad manufacturer’s online forum (I won’t name names, but it rhymes with ‘Schmatlas’…) asked about these turnouts. The poster said he had “heard from a couple people down at the club that they are junk” or some such, I am closely paraphrasing here.  I was quick to come to Micro Engineering’s defense, essentially stating that anybody who thinks these turnouts are “junk” apparently doesn’t know how to install track properly. But maybe, just maybe, I shouldn’t have spoke so soon.

 That is not to say they are indeed junk.

Lately I’ve been having more derailments than I am accustomed to. A lot more derailments than I experienced back when I was using Walthers Code 83 turnouts, I should say. These derailments are almost always of the “splitting a switch” variety, where one truck on a given freight car decides it wants to be a rebel and not follow the same path through the switch as the others. And I’ve noticed that it more often than not seems to happen with a specific make and type of freight car (Athearn PS-2600’s and Athearn ACF 2970’s, which are both short-wheelbase cars, for those of you interested). I’ve also noticed that some cars and locomotives will make a faint clicking noise as they pass through the points of these turnouts. Recently I decided I’d had enough and broke out the NMRA  gauge to investigate.  I quickly found the issue… as it turns out, the Micro Engineering turnouts do not meet NMRA specs in the point clearance department. In fact, they don’t meet the standard by a sizeable margin. What this basically means is, if you have an under-gauged wheelset, that is to say, a wheelset whose wheels are spaced too closely together, one wheel will catch and try to follow the point that is opposite of the route selected (more often than not it occurs when the turnout is thrown for the diverging route, since the diverging point rail is carrying the brunt of the wheel-guiding force on the flanges), forcing the opposite wheel up and over its respective railhead. Wheelsets that are just slightly tighter than perfect gauge will click because they are hitting the end of the point rail momentarily before following their correct path. Checking the wheel gauge on my Athearn 2600 and 2970 hoppers confirms this, they are almost entirely tight-gauged. I would go so far as to say that one can safely assume that, given the choice between tight-gauged wheelsets and wide-gauged wheelsets, it is better to have them slightly wider than perfect gauge, if you are using these turnouts.

I know just what to do to fix this issue. It involves removing the plastic throwbar and soldering a new one in place made from N scale PC board tie material (available from Fastracks), correcting the point spacing to meet NMRA standards as I go. Anybody who has tried handlaying their own turnouts knows how to do this and how straightforward it is if you have the necessary tools. The caveat is, I can’t do this modification on my turnouts without pulling them up (out of fully-ballasted roadbed, mind you) and taking them to the workbench, which is more effort and risk than I’m willing to endure.  So I guess that means I’d better invest more time and energy into making sure all my wheelsets are in gauge.  Sigh…

In closing, I can’t help but wonder if the ‘guys down at the club’ who think these turnouts are junk, in all actuality have a bunch of tight-gauged wheelsets on their rolling stock.