Adding Details To Downtown Building Castings

March 19, 2020September 1, 2021 barkmannLeave a comment

The Featured Image above shows the results of our efforts so far. Below is what we started with… Three very rough castings obviously covering concrete blocks.

I’m designing the downtown market display case view blocks to conceal the edges of the concrete block on the first floor. I don’t really have a specific example of a building to model in mind, just a vague recollection from my childhood of what buildings looked like.

The building itself is based on some of the HO kits I assembled ages ago. These kits all start with the same narrow faced three story building, but add different details, like a funeral parlor or a five and dime. The bottom floor is always different, but the upper floors have similar window frames and clear glazing. Even brick patterned blanks are supplied to close off openings if desired.

I’ve already designed the window frames, experimenting with different glazing. With nothing to conceal the concrete block behind, it was obvious something was needed. First I cut some opaque black ABS to fit behind the frames. It definitely blocks the view! But unfortunately, it also means lighting is out of the question!

In the Featured Image, the building on the right shows the different between nothing behind and black ABS. The building in the middle show what the translucent green looks like compared to the ABS. None of the building have the display case view block installed yet. See the end of the post for a view of all everything populated.

I’m also trying out the “see through” dark green filament I bought, as a stand in for the soda glass of the day, or perhaps tinted glass on automobiles. Hopefully. It’s a little dark for the window glazing, but looks better than the clear filament, which is anything but clear. Translucent perhaps, but it just looks too white and opaque.

When I had just the black ABS cutouts behind the frames with no glazing it looked better than the translucent “clear”. And now the new translucent green looks better than the ABS view block. I need to experiment with back lighting and see if it still looks as good.

Even my “venetian blind” modification doesn’t look like anything but opaque white! It works only slightly better with the translucent green. Perhaps once it’s backlit it will look like blinds covering the inside of the windows. We shall see…

Iterations and Alterations

Just as the window frames underwent a number of modifications, so did the design, fitment, and assembly of the display cases. The original casting is meant to have two display cases behind large, transparent windows, one on either side of the entrance which is placed in the middle between the two.

The entrance represents a metal framed door with a transom above, typical of the era. For variety, the transom can be fitted with a more modern air conditioner. For now, I’ll concentrate on getting something there first. Details can be added later.

The entrance will be assembled from the three pieces previously described, due to limitations of 3D printing. Overhangs are a no-no. They can be overcome using supports, but it’s more work to clean up the traces left behind by them. The tradeoff is gluing the pieces together, but how to glue them together? We need something to glue together first…

Printing Storefronts

The beauty of the design’s symmetry is both display windows are identical, allowing both sides to remain a single, interchangeable assembly. Together with the door assembly, they can fill the void of the street level opening meant for them.

I recently purchased an assortment of “metal” filaments for the 3D printer. I’m hoping the silver looks enough like an aluminum frame to pass as metal. Even just the standard silver (gray) filament will look enough like an oxidized galvanized finish, so it’s not really a problem if it doesn’t. It still looks better than nothing!

I already printed window frames using the plain silver gray filament, so I print a set using the silver metal filament first. The showcase frames have a place for colored insets above both the door and windows, as well as the bottom kick panel of the door.

Rather than print the insets in the gray color, I decide to use red to make it “pop”. I refrained from printing more window frames in red for fear it would be overpowering against the brick red of the casting. The gray will work well enough for now.

But the real test will be the copper color! I’m thinking about using it for the cream colored brick building, together with brown inset panels, to really make it “pop”! I can’t wait to try it, but wait I’ll have to.

Assembling The Storefronts

My first prints of everything need revision! The frames don’t fit together the way I want them to without a redesign. Even then, the next version still needed some judicious filing, but that left the fit too loose. Even the insets don’t fit!

I quickly learned I need to leave about 0.010″ clearance between pieces if I want an exact fit, and by exact I mean almost an interference fit. I reprinted the insets and this time they almost “snapped” in place. I say almost because they are just loose enough to find their way back out on their own.

So not only do the frame pieces need to be joined together, the insets will need something to retain them in place as well.

Plastic Welding or Plastic Cement?

Even though we’re talking about adding details to cement blocks, the plastic cement we’re talking about is like styrene model cement, the kind that chemically dissolves the plastic together to create one solid piece once it flashes off.

I’m not sure what glue works on PLA. Acetone works for ABS, and supposedly for PLA too. A quick test says I don’t know what they’re smoking, but it has no effect. Model glue does a little better, but still leaves a lot to be desired. That’s not working. Next. I read good reviews about using a plastic welder, so I decided to buy one and try it.

The ones most recommended double as a child’s “toy”, but I’m not sure what parents would hand their grade schooler a hot melt glue gun these days. That’s basically what it is, except it uses plastic filament for 3D printing instead of glue sticks. And at 210°C (410°F) for PLA, it’s about twice as hot. May as well hand your kid a soldering iron at that point!

When it arrived, my first attempt at plastic welding two parts together worked, but with mixed results. It effectively joins the parts together, but the joint isn’t very strong. And like welding metal, it leaves a visible bead of material at the joint. Let’s just say it doesn’t look like a stack of dimes…

After assembling the first storefront, I am thoroughly discouraged by the results. Even though I was able to keep the ugly weld bead hidden from sight, it doesn’t hold the assembly together long enough to press it into the opening in the casting, quickly snapping apart!

My assumption that it physically heated both parts to welding temperature while adding filler material was not a good one. It basically melts only the filler material. There is no molten “puddle” of material like metal welding.

PLA Cement a.k.a. Dichloromethane

A bit more research finds that dichloromethane is the solvent I’m looking for… a.k.a. chloroform. Great. Now where do I get that? Even after all the “Does this rag smell like chloroform?” jokes, it’s surprising how many products use it as the main ingredient! Methylene chloride was widely used as a paint stripper, now both are banned banned, replaced by a far less effective ingredients.

Just be bopping down to the paint department at the local big box store sounded like a good idea at the time, but they no longer carry methylene chloride, let alone in those gallon tins. Guess I should call myself lucky to still be able to buy acetone by the gallon.

What’s amazing to me is there are still many suppliers online. I went with “WeldOn” products, no affiliation. It’s expensive for just a pint can, but they deliver to my door! There are a number of different “acrylic adhesives” in their lineup.

I tried their #16 first, which worked incredibly well on plexiglass, but is really thick. Too thick to seep into the crevices between the parts via capillary action. It works, but has to be brushed on, wasting a lot of it in the process, kind of like using a firehose to water the flowers.

I ordered their #3 and #4 variations next. They are definitely liquid, thin enough to flow into the tiniest of crevices. The #3 is quick acting and the #4 has a longer working time. Both flow easily into the joints. All three should be used only in a well ventilated area or asphyxiation could result.

Along those lines, this stuff needs to be tightly sealed so it won’t totally evaporate away. Both the #3 and #4 came sealed in heavy duty ziploc bags over the sealed metal lid of the metal can! The first time I used it, I found out the hard way how quickly this stuff evaporates!

Learning the Ropes

The small squeeze bottles with attached hypodermic needle that came with them have a twist-lock “cap” of sorts to protect the tip. More like to protect you from stabbing yourself with it, but even that wasn’t enough to seal it. I had a third of the squeeze bottle worth of #3 left in the bottom when I went to bed. It was all but gone by morning!

From now on, any leftover gets immediate placed back in the can and then sealed in the ziploc. I also found the needle to be prone to clogging with the dissolved plastic after a short amount of use. Squeezing it to blow out the clog either doesn’t work, or if it does, floods the work pieces with solvent.

I found a better way to dissolve the clog was just to hold the squeeze bottle upside down until the solvent itself dissolves the clog. It doesn’t take long and it’s fairly obvious when the clog is gone. Hard to ignore the solvent that starts streaming out again. After the first few uses and it taking forever just to get enough sucked up into the squeeze bottle through that small needle, I decided to find some larger versions.

I now have so many of them, I can’t give them away, but the thought was they would fill quickly and I could use them to quickly transfer the contents into the hypo bottle. They came with an attached silicon cap to cover the larger, blunt tip, but even they don’t seal well enough to leave any leftovers inside for more than a day or so. I’ve gotten in the habit of pulling out a little solvent at a time, using what I need, then returning what’s left to the can when finished.

The Test of Time

So long story short, after a few false starts, we finally managed to get those three buildings looking more like a downtown business than a concrete block. They still need modifications to continue to serve us as downtown marketplace buildings though. The elements were not kind to the PLA plastic, as seen in the picture below.

Faded Glory — Time takes its toll on the PLA plastic exposed to the elements

I should have used some adhesive caulk to hold the windows in place too. Slowly but surely the dogs managed to knock out every single window. It got to the point that every time I went out to work on the Barkyard, I had to pick up a couple more windows, just laying in the street or on the ground. I started a collection of them in the garage so as not to lose them.

Next go round I’ll need to paint the plastic with UV resistant paint, maybe even just a clear top coat, but an opaque color will provide far more protection from the sun than clear coat ever could. At some point I really need to get some display cases put behind there though.

I finally removed the ground floor display cases and stashed them away too. Ann says it looks like a ghost town. So we’re thinking of making more castings, enough to build an entire ghost town downtown for Halloween! I’ll need to accelerate my efforts to add lighting and other “spooky” effects, but it should prove entertaining regardless.

Our First Casting

October 5, 2019August 23, 2021 barkmannLeave a comment

Why Casting?

The title should really be “Puppy Proofing”. They can certainly inflict some serious damage! One of the main themes of the Barkyard Railroad is sharing the space with the “pups”, so everything must be built much stronger than normal in order to withstand a direct impact from a hundred pound German Shepherd at full gait.

Another consideration is the longevity of any structure outdoors, under constant bombardment from the elements. Most plastics quickly becomes brittle after baking in the sun. Even treated lumber quickly rots in this environment.

Along those lines, I found an interesting approach to outdoor structures online, in the form of cast concrete. More accurately, cast concrete patch. I will include some links to this inspiration, as well as my source of pattern sheets for casting.

A brief introduction to the technique is warranted. Using extruded foam insulation as a sort of “backer board” for the pattern sheet(s), inserts for openings like doors and windows are also fashioned from the rigid foam, then “T” pinned in position over the pattern.

Pinning Together — Pinning it all together

There are many different types of pattern sheets available. Brick, concrete block, lap siding, and corrugated roofing are but a few examples of the many types of pattern sheets available. I started by acquiring an assortment of different pattern sheets from various sources.

Although a specific pattern sheet may be found online, its availability is certainly hit or miss, especially when searching for G scale items. But that seems to be a common theme for anything G scale. The next obstacle is finding suitably sized pattern sheets.

Patching a mold together across multiple pattern sheets is problematic at best, creating a “parting line” effect in the casting. To avoid this, I chose to start with a 15″ x 15″ square brick pattern sheet. The resulting casting will cover one side of a concrete block, roughly 8″ x 8″ x 16″ in size.

Why a concrete block? Because it is immediately more puppy proof than just a thin casting. This technique also calls for an embedded wire mesh frame to provide strength for the brittle casting. The issue becomes how to manage parts that require multiple castings to complete due to their shape or size, a pitched roof casting for example.

For a pitched roof, only one face or the other can be cast at once since the mold must be level. But the mesh must be formed at an angle across the roof peak, with the other half protruding from one side’s casting until it cures. The other side can then be cast, but this presents the unique challenge of fixturing, somehow propping the previous casting in place while the new one hardens.

There is no easy way to cast a four sided building using a single piece of wire mesh. The challenge of keeping the first wall standing and squarely aligned while casting the second wall is difficult enough, but consider trying to hold three walls in place as the fourth is cast.

That last wall would also require somehow connecting the ends of the mesh together, either from the start, or allowing enough to extend from the first wall to fasten to the other end of that last wall. Extending the thought of somehow connecting the ends of the mesh together, just folding up the edges to leave about half an inch of the mesh exposed should do the trick.

The Pour

With everything pinned together and the mesh cut to size and in place, it’s time to mix up the concrete patch and pour. Having never done this before, the consistency of the mix is in question. “Pancake batter” can range from runny to thick enough to require a ladle, so opting for somewhere in between seems the best bet. A dedicated mixer is recommended, so opting to hand mix presents its own challenges.

While the mix slacks, everything is sprayed with the mold release agent, WD40 in this case. Although the “batter” seems to be thoroughly mixed, the patch wants to settle out to the bottom of the mixing bucket. The closer to the bottom of the mixing bucket, the thicker the consistency becomes. Working around the inserts is difficult with the thicker mixture and even a small trowel isn’t helping much.

Once poured, the mix needs to be vibrated into place to remove any trapped air bubbles, like real concrete. A handheld jigsaw was recommended, minus the blade of course, but a palm sander will do in a pinch. The idea is to vibrate all around the mold, and for long enough to allow the bubbles to float to the surface and burst. Now all that remains it to let it set up for 24 hours then remove it from the mold.

The first attempt at removing the bubbles proved to be an exercise in futility. While the plywood bench all around the insulation base was vibrated, many bubbles remained trapped. And while all those minute voids in the casting can be filled, it’s more trouble than it’s worth. The take away is next time, the mix must be thinner, and the mold itself will be vibrated directly.

Once the casting had cured a few more days. it was painted. And a nice brick red it is. The air bubbles still show, but aren’t discernable from a distance. The “Five Foot Rule” is in full effect on this first one, a learning experience for sure.

Another lesson learned is to wait until the casting is no longer “green”, that is, fully cured before fiddling with it… let alone flexing it or the exposed mesh along the edges. The missing edges where the casting fell apart because of that flexing are much more noticeable than those tiny voids in the brick face from the bubbles.

The Final Product

After painting, the casting was secured to the concrete block using thin set mortar. The missing edges were filled in as well, to look like a “repair” of sorts to crumbling bricks along the edge of the building. Now all it needs are windows for the upper floors and display case windows and an entrance for the ground floor.

All in all, our first building casting for the downtown marketplace is a success. Is it perfect? By no means, but it certainly looks more like a three story building than a bare concrete block does! Some “window dressing” and other details will make it an even more believable model.

Lighting is another consideration, and by extension, how to power the lighting. But all that will wait for another installment. We started working on the downtown marketplace back in June of 2019, so there is plenty more to come. Stay tuned!

If you have any questions, or you would like to see more detail, leave us a comment.

Building A Howe Truss Bridge

October 18, 2015August 23, 2021 barkmannLeave a comment

Building Our First Bridge

Our garden railroad is progressing, but we need a bridge. As with the trestle bents, this post is a foreshortened version of all the research and planning involved. I’m an incurable rivet counter when it comes to details and modelling, a hold over from my HO scale days, so it is difficult for me to accept anything less than prototypical.

The approach I settled on is, shall we say, less than prototypical. But it does allow for “quick” assembly, combined with a modular approach. And again, like the trestle bents, it requires a jig. Beyond that, it also requires additional brass hardware… Threaded rod, nuts, and washers for examples. I chose to use #2-56, but #1-72 would be closer to scale.

The modular approach is an adaptation of a commercially available system for building a Howe truss bridge. The modular concept to based on “opposed overlapping” end pieces, coupled together by multiple standard overlapping sections. A prototypical Howe truss bridge is composed of massive compression members, and comparatively “puny” tension members where iron rods assume the tensile forces.

Sounds like too much engineering mumbo jumbo? The quick and dirty version is wood doesn’t take kindly to stretching, and metal doesn’t like to be pushed, even in the form of a spring. The design of a Howe truss accounts for this and it’s appearance. Another feature of its appearance is the reason why my rivet counting background makes this difficult for me.

The prototype uses a metal casting to bring all those components together, a component I cannot replicate with the modular approach I adopted. Someday perhaps, but not now. The prototype does not use the massive 12×12 beams, both above and below, to contain the trusses and terminate the tension rods. But I’m getting ahead of myself. We don’t even have trusses yet!

Building A Truss Jig

After a number of false starts, I finally came up with a final design and a jig to hold thing together for assembly, which starts with placing upper and lower members. Then the first course of diagonal compression members are placed over those.

Another set of upper and lower members is placed over those, but these are a “cell” shorter than the previous. Then another course of diagonal members, this time in the tension direction. One last layer of upper and lower members, yet again a “cell” shorter.

Hopefully the pictures are worth thousands of words, but perhaps I should start with how I made the jig. I learned from my experience with the trestle bent jig and started with a chunk of ¾” plywood. After carefully calculating the placement for the upper and lower truss members, I proceed to cut ½” wide slots for both members, 3⁄8″ deep, using the router.

Those upper and lower members are only ¼” thick, but the jig must be recessed enough to allow for the diagonal members to be placed over them, which accounts for the 3⁄8″ depth. Again using the router, I cut slots for the first course of diagonal members, this time only 1⁄8″ deep. I’ve never really had much success with using the router, even with a guide, so I’m a bit nervous.

Those with a keen eye will notice that I missed by half an inch on the lower slot and had to cut another next to it in the correct position. I must have miscalculated, but thankfully was able to correct for it without having to start all over again.

Building A Truss

By building two of the end assemblies, one can be flipped over and placed on top the first such that the diagonals overlap and fill in where the other is missing them. A truss constructed in this manner will be rather short and somewhat useless, but it demonstrates the approach. Another picture is in order. The horizontal members are roughly nine inches apart. Where the diagonals meet them are about four inches apart. The minimum truss length is about two feet.

A truss of arbitrary length can be constructed by first starting with a pair of end sections. Any number of intermediate sections can then be inserted between the two end sections, using the same offset technique for placement of the diagonal members, except the length of all the horizontal members of the intermediate section are all identical. The intermediate sections don’t necessarily all have to be the same length

This is the modular approach I borrowed in hopes of building the bridge trusses a section at a time, then later assembling them all together into one long truss, four feet long. The jig allows me to build intermediate sections up to two feet in length. Combined with the two end sections yields the desired four foot span.

Each and every joint requires cutting a piece of the threaded rod, and that requires chasing the threads, mangled by cutting. The rods come in three foot lengths. The joints need a chunk less than an inch long. Before cutting, nuts are threaded on the rod to chase the threads after cutting on their way back off, but this proves difficult and ineffective.

While using the nuts is good enough for proof of concept, the rivet counter in me prevails. A good set of threading dies make the task much easier. The correct sized machine bolt would be perfect, but the only items available in these small sizes are machine screws.

We’ll stick with the more prototypical nuts, although hex nuts aren’t strictly prototypical either. Square nuts and square headed bolts would have been used on the prototype. Also not strictly prototypical is the way the horizontal beams are joined, but short of stamping or casting joining plates that the prototype employs, this will be close enough.

The most frustrating part of assembling a truss is having to partially disassemble the modules just to join them together and reassemble. It’s hard enough for my fat fingers to fumble with threaded these tiny nuts on the rods once, but having to do it twice or more is trying my patience.

Perseverance prevails and I finally have two fully assembled trusses. Now to join them together into the final bridge span.

Building A Bridge

The two trusses are held together by pairs of scale 12x12s and tension rods. The scale 12x12s are cut to extend beyond the sides of the trusses. Holes for the threaded rods are drilled in the ends. Threaded rods extend through the pair of 12x12s, above and below the trusses, secured by nuts and washers on the tension rods.

Sounds simple enough, but no amount of hands seems enough to hold everything together while tightening the nuts on the rods. A different approach is called for. By preassembling the pairs and rods, sliding them over the ends of the trusses, then tightening to hold everything snug works much better.

Assembly continues by just snugging the rods at first, then tightening them until they can be “plucked”, like tuning guitar strings. More like bass strings as thick as the rods are, but they still make their own music of sorts.

As the rods are tightened, the structure becomes more rigid, but wants to twist the tighter they get. It appears the trusses are deforming under the force. Attempting to twist things back in shape causes the trusses to deform further.

We have a bridge! Good enough for now, but there shortcomings that will need work.

Items To Address

The first issue encountered is accurate and repeatable placement of the holes for assembling the truss members. A few modifications to the jig will allow the use of my dremel drill press and a means to accurately index the holes. It should help speed up assembly, but assembly will remain clumsy and difficult using the modular approach.

Another disadvantage is joining the horizontal members “together” where the diagonal members join. A less than rigid truss assembly produced. Even though the joints are staggered, the forces applied by the diagonal members tend to force the horizontal members apart, much weaker in comparison to a single member.

Slop inherent in the modular truss approach could be overcome using a number of different approaches.

Making Trestle Bents

August 3, 2015August 18, 2021 barkmannLeave a comment

Our First Post

This will be our first blog post ever and we’re excited! Just as eager to get our garden railroad off the ground and running, we’ll need a reason for it to be off the ground. And that reason would be a bridge. A bridge over something. And to get from the ground up to that bridge we’ll need an approach trestle.

Many folks have many different ways of creating a trestle. Even the different railroads had different ways of doing it, right down to the trestle bents. What’s a trestle bent? Well, there’s plenty of information online, but simplistically it’s the individual “tower” structure that is repeated and lashed together with the rest to create an assembly called a trestle.

The key word here is repeated. Most folks agree that a jig will help immensely when repeatedly constructing all those trestle bents needed to assemble a trestle. I’d be happy to make just one bent to start with, but know that I’ll need many more than that, so I make my first cut at making a jig.

Having just replaced the fence between us and our new neighbor here in Mount Dora, I have plenty of well weathered “scrap” wood to rip into the raw materials I’ll need. I start by ripping the scale 6×12 and 12×12 members I’ll need to cobble together that jig.

Our First Jig

Starting with a chunk of ½” plywood, the main “T” portion is aligned, then the angled members drawn in at 5° and 10°, respectively. The straps are drawn in every 10″, a scale 20′ at “half doll house scale”. What’s half doll house scale? Well, if “doll house scale” is one inch represents one foot, or 1:12, then half doll house scale is a ½” represents a foot, or 1:24.

Some would argue that G scale is 3⁄8″ represents a foot, 1:32. USA Trains uses 1:29. Since the Bachmann Big Haulers we have are supposedly 3′ narrow gauge, that would be 1:20.3, sometimes called F scale. I use 1:24 because it makes the math easy and it’s closer to F scale.

Add to that our goal of sharing the space with the “pups” – everything must be built much stronger than normal in order to withstand a direct impact from a hundred pound German Shepherd at full gait. A 12×12 post that is a ½” thick is stronger than one 3⁄8″, so there you have it.

Back to that chunk of ½” plywood… We’ll call a jig for now. Short pieces of the ½” thick scale 12x12s are attached with self tapping screws on either side of the five posts to act as guides. More are placed to help keep the top beam in position. The template helps hold the pieces in alignment during gluing and assembly, but it has shortcomings that must be worked around.

Laying Out Jig
Trestle Bent Jig

Cutting those posts to exact length proves to be another challenge. No matter how carefully cut, their lengths are just enough different to force hand selection of each piece. So much for assembly line efficiency!

The scale 6×12 pier sills that delineate each new section of the bent and associated scale 3×10 cross bracing must be attached from the top side and bent then removed from the jig to attach to the other side since no provision was made for them in the jig. Brass brads are used to simulate what would be nuts and bolts on the prototype.

The brass will weather to a dull brown to simulate rusted hardware, yet survive for years without further corrosion as iron or steel would.

Actually, the cross bracing is added last, hand cut to fit and tacked into place using some HO scale track nails! Those will rust away to nothing eventually, but they were only meant to hold the sills in place until the glue cured, and will provide a realistic rusted patina over time.

Our First Bent

This is all brand new, so I’m learning as I go. Ann isn’t very happy with me assembling these in the living room, but there is nowhere else that I can. The garage was really just the old carriage house, two strips of concrete over a dirt floor, with no lights or electricity whatsoever. We’ve added a wood floor, a side door in place of the old window, electricity and lights.

But it’s still a work in progress. I’m fairly limited without an easily accessible workshop. Thankfully most operations are limited to drilling pilot holes and tapping in brads after gluing the pieces together. I need a few quick clamps here and there to hold things together, but manage to get the first bent assembled.

After trying to remove the bent and reinsert it face down, it’s obvious that the jig isn’t even symmetric! It may have drawbacks, but at least I can assemble bents for testing and assessing other operations, such as assembling multiple bents into a trestle.

This post is a woefully foreshortened version of all the research and planning required. Some railroads used poles in place of posts. Some used four posts rather than five. Other configurations exist, such as those for two mainlines, with more vertical posts than one. This design is a compromise between the differences and meant for a single track.

It may not be symmetric, and it may not be pretty, but it’s a start. Next up is determining how to stain and seal the bents while giving the impression of creosoted posts and beams. Stay tuned for further developments.

Barkyard Blog

Our Barkyard Railroad Adventures

Category: Scale Models