First train!

I ran my first train on the new layout last night :)

If you come from another layout, it's about as unexciting as you can get, but it's the first time I've had a layout working well enough to run a train any distance at all since July 2002, so I'm pretty excited.

The train was simply this Life-Like / Mantua 2-6-6-2T in Canadian Forest Products colours, along with 5 assorted cars from the cheap table at recent train shows.  This locomotive is one of two DC locos I have, since I don't yet have my DCC system (I'm planning on getting an NCE PowerCab next month). In fact, I don't even have a proper DC power pack; I used an old industrial beast of a power supply (0-20VDC, 20A capacity) but it worked well enough to run the train. Simply swap the wires to reverse :)

I'm please to report no derailments or other track issues throughout the entire 48-foot (each way) run; forwards uphill and backwards downhill. No feeders run yet, just a pair of alligator clips at the end of the first rails and worked fine through 16 sections of flextrack (soldered joints on curves, which is 95% of the main line). This is the longest run between turnouts (which won't be soldered to adjacent track), excepting the helix.

Construction: laying track

Finally I get to lay some track!  The distilled wisdom of the Internet has told me that nailing is bad, gluing is better, and gluing with caulking is the best. Transparent caulk is what I see recommended most often (so it's not unsightly if it peeks through the ballast), but I noticed some nice gray caulk at Rona (again, same DAP Alex+ brand) for the same price as the clear/white/almond/brown colours. I don't know what they put in it, but I found that the 300mL tubes of white-becomes-translucent and gray caulking are significantly different in weight. Indeed, the scale reveals: clear = 350g (12.3oz) whereas gray = 469g (16.5oz). I assume the differing ingredients make the gray caulking that much nicer to spread. For laying track I run a bead up each cork half-strip (approximately under where each rail will go) and then with a cheap putty knife spread it out into a smooth thin layer (1mm [~1/32"] is all you need). Just like icing a cake. Once the caulking is spread out thin and smooth, just plop the track on top if it and press down gently. The caulk is sufficiently sticky to keep the track in place with minimal clamping.

Backing up a bit; before getting to the sticking-it-down part, I first lay out 3-4 sections of flextrack at a time along the roadbed, pinned and/or clamped in place temporarily, then cut the rail ends as needed to fit the curves. Once satisfied with the alignment, I solder the rail ends together, but I don't solder in the curve -- I clamp the the end of the first track a few inches back from the end and let it straighten naturally, then attach the next section of track tangent to the end. This way both rail joints will be soldered perfectly straight and can then be bent into the curve. Attempting to solder in a curve is a great way to get kinked joints.

My soldering technique is far from perfect (as is my equipment), but what I do is clamp some alligator clips to the rail on either side of the joint as a heatsink, then apply the iron to rail+joiner on the outside of the rail and melt a small bit of rosin-core solder into the space between the rail joiner and the base of the rail. As soon as that melts, I take off the iron and blow it cool (to avoid melting ties). Then repeat on the other half of the joint. Once both rails are tinned, a third application of the iron blends the solder across the joint (if it wasn't already) and smooths out the appearance.  I only solder on the outside of the joint to avoid any potential bumps and lumps where the wheels will roll.

At the end of the first day of tracklaying I have 33 feet of mainline down, from the yard, around the base of Marshall Hill and almost leading into the curve on the outside of the second level of the helix.

Atlas flextrack (SuperFlex 83) revisions

Apparently Atlas HO scale flextrack (specifically their code 83 Super Flex-83) has revisions. I happened to have two pieces side by side and noticed minor differences between what appears to be revisions "A" and "C".

On the underside of the track the most noticeable difference is the webbing between the ties. On "A" it is as wide as the base of the rail; on "C" it is about half as thick.  The writing has also been reformatted a bit, but the only notable change is the revision letter. The other noticeable difference is the tie thickness has changed: on "A" each tie is 2.35mm (0.0925" = 8" scale), on "C" it is 2.20mm (0.0866" = 7.5" scale). A minor difference to be sure, but it can definitely be seen when they're side by side.

The other, odder difference is on the top side. The spike detail has changed. Where it gets weird is that on one side of the track the spike detail got finer. "A" spikes are 0.75mm wide (0.0295" = 2.6" scale), but on "C" one side has finer 0.6mm (0.0236" = 2.0" scale) spikes, but the other side has clunky 0.95mm (0.0374" = 3.25" scale) spikes!

For comparison, the spikes on the classic Atlas code 100 flextrack are a massive 1.75mm (0.0689" = 6" scale) on 3.4mm (0.1339" = 11.7" scale) ties).

Has anyone else noticed this?  Any idea why they did that?  Of note, the finer spike detail is on the fixed rail and the coarser spike detail is on the sliding rail.

05-Dec-2010 update: I found some revision "B" track in my collection. It looks superficially similar to "A" track, I haven't bothered to pull out the calipers to see if any of the dimensions differ, but they look very close. I have attached a scan of both top and bottom for reference.

Construction: laying cork

My first order of track came in a couple days earlier than expected, so I now have an assortment of switches to play with to get final alignment of some unlaid segments of spline subroadbed, as well as final alignment of the cork roadbed. For track, I opted for Walthers/Shinohara code 83 DCC-enhanced switches combined with Atlas Superflex code 83 flextrack. One disadvantage of this combination is that while the rails are both 0.083" tall, the Atlas ties are 0.020" thicker than the Shinohara track (apparently this is so that Atlas code 100 and code 83 track come out to the same railhead height). What that means for me is I need to find a source of inexpensive 0.020" (0.5mm) material (possibly sheet styrene?) to use as a shim under all the turnouts.

A quick trip to the local Rona showed that there is plenty of inexpensive silicone caulking available. I picked up several tubes of DAP Alex+ clear caulking for $2/tube. Well, "clear" is yet to be confirmed; according to the label it applies white (this I can confirm) and cures to clear after 7-14 days (we'll see). If it does indeed go clear then I'll use it to attach the track to the cork. Alternately, it does come in several other colours (white, almond (very light brown), dark brown and gray. If the "clear" caulking isn't to my liking, I'll use the grey one for attaching track; the colour should be very close to ballast colour.

The wisdom the Internet had revealed to me showed that the standard strategy for gluing down cork with caulking is to lay down a bead on the subroadbed and spread it with a putty knife. This worked, but I found it a little messy. I tried applying a bead directly to the underside of the cork strip, not spreading it, and applying the cork to the subroadbed and pressing in place. For my particular subroadbed construction this seems to work very well, as there are some minor irregularities between the triples of spline and Styrofoam strips, and with the bead of caulking on the cork directly, pressed firmly in place, the caulk seems to fill in all the necessary gaps without excessive spillover. On the subject of subroadbed irregularities, despite my best efforts there is inevitably some variation in top level of the 6 hardboard splines and 3 Styrofoam strips, but a Stanley Surform Pocket Plane is a fabulous tool for quickly smoothing it all out.

So, by the end of the night I'd emptied a box of cork (25x 3-foot sections) which accounts for about a third of the mainline. Next on the hitlist (now that I have actual physical track turnouts to check alignment with) is finishing the last piece of mainline subroadbed, the passing siding on the top of Marshall Hill. After that I can continue laying cork for most of the mainline and then (hopefully) onto track sometime next week!

Construction: subroadbed continues

Work continues. I've got most of the main line spline subroaded laid out now. I can't join it to the helix until the helix is built to all 5 levels, and I can't build the helix until I get my order of track in. I also can't exactly position the subroadbed for the sidings and spurs until I get the track in, so I'm a little stuck until the end of the week. Nevertheless, I have experimented with attaching two sections of cork roadbed; one with white glue to plywood in the helix; the other with white latex caulk to the splines. The caulk is admirably tacky, it takes very little effort to keep it in place in the correct alignment.

I attached a rough printout of the planned 4-foot truss arch bridge (to be based on CPR's Stoney Creek Bridge). It gives some sense of the chasm to be spanned. I remember there was an excellent article on building exactly such a bridge in the April 1991 issue of Model Railroader Magazine. Unfortunately I sold my 50-year collection of MR about 10 years ago, now I'll need to find a copy of that issue somewhere (although I can always order from the publisher if the train shows don't yield results).

Design: reworking the yard

 

 Based on feedback from LKandO, and my own dissatisfaction with the original design (especially the proximity of the yard throat to the mountain at the bottom-right), I've completely reworked the arrangement of the yard. Compare to plan in previous blog post. Key changes:

• turntable moved from end of yard to entrance of yard.
• horseshoe curve no longer contain storage tracks, just two approach tracks (30", 27") and a switching lead (24")
• found room to put back in the car float (apron at least, float would be removable off-layout) I had originally conceived in a much-earlier design
• quite importantly, design is now based on actual measurements of the space with backdrop and benchwork in place (previously was just rough dimesions, not counting the 6"+ lost to backdrop etc).

Feedback on the new design is very welcome.

Construction: Marshall Hill lower track relocation

Skipping ahead a bit to current events: the mainline roadbed has progressed, at least in partial form, from the helix around the loop at the other end of the room (around the future logging camp) and back around to the upper level loop around the top of Marshall Hill (center of room). I had planned to offset the two loops somewhat so that the lower track could be at least partially visible (not entirely in a tunnel), but with the upper roadbed temporarily in place it became obvious that the small offset I had allowed would not allow any reasonable hill slope (or even a vertical cliff) in most places. Fortunately I also realized that the aisle next to the helix was 39" wide, far larger than my standard 24", so I decided to cut down the aisle by extending the lower track loop outward. The lower track should now be able to stay exposed for 75% of its circle, only entering a tunnel where I actually wanted a tunnel in any case.

In the photo the old roadbed is still there (with a piece of flextrack pinned in place), and the new path of the lower track roughed in about a foot to the right.

Construction: Helix (part 1)

The track plan had always included a helix, but I hadn't really planned on putting it in place until later on in the construction process. The track runs from the main yard at one end of the room, eventually to the terminal at the other end, having climed nearly 30" via 3 lengths of the room and 4 loops, but the helix is simply to return from the upper terminal to the hidden staging area (and optionally back to the main yard and/or continuous running). However, I realized that the second loop out coincides with the location of the helix and that at the very least I needed the exact location of the helix determined so that I could lay roadbed for the mainline around it. But I quickly realized that the mainline at that point coincided nicely with the second level of the helix in location, curvature and slope, so I decided to simply double the width of the second-level helix roadbed pieces to have a 3/4-circle of double track.


The helix itself is 4.5 turns of 28" radius, rising 4" per turn (width of a 2x4 on edge for spacers, plus 5/8" plywood roadbed), for a slope of (4" / 2•28"•π) 2.7%, which is about average for the entire main line. The roadbed is 4"x24" strips of 5/8" plywood, cut at 22.5° angles at each end. The wider second level deck (not pictured here, yet) is the same idea, but extended to 7.5" width to accomodate a 30.5" radius track on the outside. Spacer blocks are cut from 2x4s, with 45° angles on both ends to make an isosceles trapezoid shape (about 5" on the long side, 1.5" on the short side). You can see how they're laid out in the photo, alternating inside and outside placement. This provides a relatively large support area for the levels above, while still providing plenty of clearance from the 28"-radius track. The only issue I'm mildly concerned about is alignment on the inside edge of the roadbed joints. The outside is nicely supported across both pieces, but the inside is floating, and any warping of the wood would throw it out of alignment. I think I'll simply glue on a splice plate of 1/8" hardboard under the joint to ensure vertical alignment. Even with cork, track, and splice that still leaves 3" of clearance.

Constrution: spline roadbed

I had experimented with spline roadbed construction on a previous layout that never as far as track laying (had to move before then) and liked how it worked, so I'm doing that again. My method is a little different from anyone else's I've seen.  1/4" hardboard would probably be the better choice, but I had a supply of 1/8" hardboard so that's what I'm using. I make a sandwich of hardboard and Styrofoam: two triple laminations of hardboard strips (one under each rail), with a strip of 1/2" Styrofoam in the middle for spacing, and then another strip of 1/2" Styrofoam down each outside to widen the roadbed (including as a scenery attachment point). It's all held together with yellow carpenter's glue, and I have no issues with adhesion, it's as solid as could be desired.

So how to cut all these strips? The hardboard I ripped into 7/8"-wide strips by first ripping two 4'x8' sheets of hardboard into four 2'x8' sections, then with the edges lined up, drill through all four strips and bolt them together so they can't move relative to each other. So I effectively had a 2'x8' sheet of 1/2"-thick (4-ply) hardboard. I then rigged up this guide frame & support with a 2x4 and some 1x4's on each end of my bench saw, carefully aligned with the rip fence, so I could then just set the stack of hardboard on one end and slide it through the saw without worrying much about alignment. It worked very well, all the strips came out nice and straight and uniform width.

But sawing hardboard produces dust. Lots and lots and lots of dust. A dust mask and eye protection are essential. This is what my saw looked like after ripping 4x24 8-foot strips, and that's with the shop vac hooked up to suck dust away. There was more than a grocery bag of dust just on the floor immediately under the saw, not counting all the dust floating around over everything. This is something that should ideally be done outside or somewhere well-ventilated which you don't care if you get dirty.

In the end I had a big pile of hardboard strips all set aside ready for installation. Now I needed to figure out how to cut 1/2" Styrofoam into uniform strips. I rigged up this simple contraption with 3 blades. I found that trying to cut the whole 1/2" at once was too much, but cutting 3/16" or so with each blade lets the sheet slide through smooth and easy. I just have this rig screwed to the wall and it's extremely easy to cut strips of Styrofoam from the 2'x8' sheet.

 Then it's just a matter of sticking it all together. Pretty simple, really, as long as you have clamps. Lots and lots of clamps.  I found these beautiful ones (oversized clothes pins) at the dollar store (10-pack for a dollar, can't beat that) and they have about a 2" opening (enough for almost all of it, but can't quite clamp with the last layer of Styrofoam on the outside. The inner notch just below the hinge is perfect for starting out, it provides tremendous clamping power for up to 3/8", which coincidentally is just the size of each set of hardboard splines, then the end of the jaws can hold the hardboard-foam sandwich together just nicely. I do also have some heavier-duty spring clamps and C-clamps, especially for the ends of each spline on curves where a lot of clamping pressure is needed to keep it from springing out of alignment.

Track plan

I guess a copy of the track plan would be useful. I spent the month of July reading track planning books (such as John Armstrong's Track Planning for Realistic Operation), and doodling rough plans on paper. I spent August refining the design in XtrkCAD (which is now freeware). This is what I came up with:

I can cheat a little bit by wrapping the triple-duty loop (helix, hidden main line, visible main line) on the left side into the space under the stairs, which gives another two feet of room length and makes things fit much nicer.  What isn't well conveyed by this track plan image is the elevation changes. This is a mountain railroad with a perpetual grade of 2-3% along the 100+ feet of mainline. The visual crowding on the plan shouldn't be the same in 3D, since apparently-close tracks are quite vertically separated. For example, the two loops around the centre peninsula are about 24" apart in elevation.

I did omit from the above plan a logging line that goes from the terminal at the left side, along the bottom wall (above the staging area) to the logging camp above the access hole in the lower-right corner. This line will be unconnected to the main track of the layout (potentially narrow-gauge) and feature sharper curves and steeper grades. Main line is 30" radius at 2-3% grades; logging is around 18" radius and 5% grades.

Construction: 2010 September

I'm two months into construction on my new layout, but methinks I should start a blog to post some progress photos and tips, since I found these useful on other blogs.

I'll run through this and catch up on two months of photos with some comments:

[07-Sep-2010] This is what I started with. Completely empty room (roughly 19'x14'), bare concrete floor and walls (top half of walls overlaid with about 2" of pink Fiberglas insulation wrapped in plastic), open studs for the ceiling, unfortunately no studs on the walls, and a single dim lightbulb in the middle. Unfortunately there's two plastic pipes in the way down the left side, but I can work around that.

[16-Sep-2010] I replaced the single lightbulb with five 4-foot fluorescent fixtures and it's looking much better. Studs are laid out with a horizontal pressure-treated lower piece on the floor, then the vertical stud laid parallel to the wall offset by the thickness of a 2x4 to leave room for the insulation behind the top half.  Bottom is screwed to the lower member, top is screwed to the floor joists above. Nothing is secured to either the wall or the floor. In this photo I'm starting to put up the backdrop, which is 4x8 sheets of 1/8" hardboard (aka Masonite) glued with LePage No More Nails construction adhesive.

[13-Oct-2010] This scene looks a lot happier; the backdrop is all up and painted. The right half of the room will be mountainous and only need half the backdrop height as the track will come up above that level. The section on the left will be a yard at the lowest elevation of the layout, so it needs the full 4-foot backdrop height. Backdrop is painted with a couple coats of white primer, then blue #059-63 from RONA for the sky colour. My artistic skill for drawing and painting is frightful (I can't draw a convincing stick man), so I was terrified of ruining the nice smooth blue sky, but I went ahead and used The Clouds cutout templates from New London Industries and some Krylon flat white spray paint, and after a few attempts at practicing, they actually come out quite well. The key is to keep the template a little away from the backdrop, maybe 0.5" or so. If it's right snug against the backdrop you'll have an unconvincing sharp outline of the cloud, but the small separation gives a nice feathering effect. Don't worry about a lower half for the cloud at all, the overspray is highly desirable for producing a realistic fading effect. Start at the top of the backdrop and work your way down, overlapping the lower cloud on the overspray from the previous cloud. I found you can do about 10-12 linear feet (2-foot high backdrop) of fairly dense clouds per can of spraypaint. Spraypaint, however, is highly undesirable in an enclosed room. Despite my best efforts at ventilating the area, and wearing a dustmask (prevents breathing paint particles, but not fumes) and trying to breathe as little as possible, I could only work for 20 minutes or so before taking a break to recover.  Much ventilation and a respirator would be recommended.