I was a happy early adopter of MTL sectional track. I liked the idea of integrated roadbed and secure, snap-together connections. The more I used it, however, the less I liked it, and when the switch sets were released I became exasperated with their performance, even with some of my MTL locos and rolling stock.
I finally sold off the whole batch of track, mentally writing it off as a 'learning experience.' What didn't I like, besides the usual grumbling about the switches?
I thought that there were some problems with rolling stock stalling and derailing on what looked and felt to my finger like smooth, correctly-aligned track joints. I didn't spend a lot of time investigating these stalls, but they were perplexing.
I didn't like the fact that the joints between sections required the finicky and fragile metal joiners on the rails. I would have preferred a combined mechanical/electrical connection via the roadbed, perhaps with removable pins for the electrical connection which would have made building isolated sections quite easy. (I realize that any variation from the 'tried and true track joiner system would have made it hard to incorporate Marklin track or any non-MTL roadbed track into a layout.)
I didn't like the roadbed itself: I didn't like the color, the rather coarse texture, or the fact that, maddeningly, it seemed to have a problem staying flat. Curved sections especially seemed to have a tendency to twist.
Finally, I didn't like the lack of completer pieces: isolation tracks, short sizes, etc.
And getting back to the switches, I thought the designers made a strange choice to make them power-routing but not power-isolating. Also, although it is not 'prototypical behavior' I really missed the ability for a loco to 'barge through' a closed switch, making it easy to back a loco off a siding and then reverse it along the main line without actually 'switching' the switch. I'm just a lazy runner and I liked the rather passive nature of Marklin switches, power everywhere and points floppy enough to roll through backwards in the 'wrong' direction.
So, time has passed and for a variety of reasons I recently bought a new batch of MTL roadbed track on eBay, a batch which I intended to use for a switchless, mostly straight-line project, and only for that project. There was one left-hand switch in the batch, however, and when I bought the few pieces of additional track I needed for the project I also bought a right-hand switch. And when I snapped-together my back-and-forth layout I inserted the switches to see how they would perform.
Based on the minuscule sample of my two new switches, my unscientific locomotive tests turned out this way:
A Marklin CG-1 rumbles through all configurations and directions just fine, as does a Marklin F7. Furthermore, my MTL GP-35 and GP-9 chassis seem fine in both directions in both switch states.
My AZL SD-70, (an early, super slow Norfork Southern beauty,) had significant problems with the switches until I re-gauged its wheels. Now it negotiates the switches well.
My Marklin I.C.R.R. 4-6-0 has problems with the front pilot jumping around (sometimes causing derailment; often sparking as it bounces off 'live' tracks, especially on the right switch) but the passenger cars from that set seem fine.
A Milwaukee Road Mikado 2-8-2 (one of my better runners anyway) handles the divergent paths just fine, but a lowest speeds can get hung up on the straight path, lifting the drivers just enough that they lose traction but continue to get the electrical feed and so 'spinning its wheels.' Not bad, though, for a non-weighted Mikado chassis. An old Marklin railbus has the same issue, getting hung up at low speeds entering the straight path. For both engines the right switch is more problematic than the left.
A lowly DB89 tank loco, (a reasonably reliable runner) handles the MTL switches nicely, even at relatively low scale speeds, around 20mph. The little tanker does complain at the track joints, however. (See below.)
Most of my rolling stock, MTL or Marklin, breezes through the new switches, with the exception of a Chicago Railcar Observation car, which has brass trucks and metal wheels. I haven't checked the gauge of the metal wheels but I wouldn't doubt that they're a little off.
All of this seems consistent with a diagnosis of the switch 'frogs' being finicky about width of the wheelsets, from back of flange to back of flange. If the wheel sets are too narrow, or too wide for that matter, problems abound. Furthermore, steam locos, or any loco with a set of axles fixed in a rigid block, generally have problems with Marklin Z-scale switches, as well as MTL switches. Bottom line is that the new pair I own seem better (more tolerant) than the first switches I bought, but they're still more finicky than I would like.
I have figured out, I think, what might have contributed to what seemed to be random stalls/issues at track joints: there's a slight natural 'dip' in the track surface in the joints.
I was slow running my Burlington SW-1 along a section of new 220mm straights; the Burlington paint-job featuring a straight yellow horizontal stripe along the body work. I could clearly see the little engine bob at every joint. The effect is not as apparent with longer engines with more flexible trucks, but it is real. (Check out the video, especially in the forward pass as the second axle passes over the joint. Sorry about the size of the video but you need the full scale to see the effect, and even here it is difficult. If the video pauses on initial download, the second play should be from cache and should be smoother.)
If you can't see what I see; my digital micrometer reads 4.58mm from trackbed bottom to rail top in the center of a random 220mm piece, but 4.50mm at the joint end. Another random piece: 4.6mm in the center, 4.48mm at the joint end. Admittedly, these are random measurements only as accurate as my micrometer and my measuring technique, but still, I find a consistent variation from center to edge. Let's see... if it is .08mm then that's a scale 17.6mm, just over 3/4 of an inch. Maybe a railroad man can tell me if that would be acceptable in real life. It would be a very visible dip, I believe.
For now (July of 2009) I'll probably continue to use MTL track for my "bannister railway" experiments, and perhaps for other large, simple layouts with lots of straights and wide curves, but I'll continue to miss the missing pieces (and the roadmap of what pieces are in the pipeline) and I'll continue to have to slip little pieces of painted IBL cork roadbed under the Marklin pieces I need to use, and I'll continue to grumble with joints that aren't flat by design, and with trackbed that doesn't like to lie flat. And if I use switches with MTL roadbed track, I might end up with old, loose, floppy Marklin switches.
Back to FABbNet main page