My first shelf layout

Whew, I've been dreading this for days, but I finally got the next section of track put down. This included a pair of turnouts with wires to feed through the foam, and two curves that needed held in position with jigs, plus some straight pins to correct the alignment where things wanted to drift. I finished with some metal blocks to keep things flat. I'm also getting better at not being so sloppy with the caulk. Fortunately everything else after this is pretty simple in comparison, now I'm just waiting for the caulk to set up for an hour so I can verify the track is all positioned correctly.



I used some thinned raw umber on the basswood ties last night. It's rather too bad I'll be painting over it because that leaves a really nice color in the wood. On the first shelf I painted the ties, then laid down the track and had to guess where the lines where supposed to be. This time I figured I would just paint everything with the ties in place but I wanted to give the wood some kind of initial coat in case the pain didn't reach everywhere. Now I was able to see my lines and get all the track laid in the correct positions (really helpful for those curves!) so I think it went a lot smoother.

The next pair of turnouts go at the upper-left. The wood ties are already in place, I just need to solder the feed wires to the track and they'll be ready to put in place, then I can start prepping the flex tracks that lead to the last turnout on the mainline.

Hmm interesting. I was looking at the page for the 8'6 ties and I thought it said one package provided enough ties for 6' of track. Yeah it's all very confusing, you would think they would make a point of letting people know how many feet you get in your order, and point out that they also sell the bare rail. Ah well.

I've been using ME track for the flex and the turnouts. And I'm really loving the look of the code 70 track, especially in the close-up pictures. I actually got a heck of a deal off ebay awhile back, two packages of ME bare rail (33 rails per package) for $70 including shipping. The first package (single) of this stuff that I bought cost $66 with shipping! If I ever get around to building a big layout, I think I have enough rail on hand now to build most of the turnouts.

I got the new pair of turnouts glued down a couple hours ago. I also trimmed up the excess track around the bridge and confirmed that it maintained the correct position through the curve, so when I'm ready to lay in the bridge track everything will be in alignment. Now I'll just be prepping the flex track to install the next turnout, and that will take me to the far end of the shelf and once I finish the bridge it will complete the mainline.

It occurred to me that I haven't posted an update pic in the past week, so I figured it was time. I've been laying down the turnouts and now have the mainline plus a run-around finished. I also have the track prepped for the first siding. Hopefully I left enough gap in the rails this time. I got some jumper wires soldered to the track to bring power to either side of the bridge, but I've also been thinking about power routing this weekend... I have all these wires coming from the turnouts to supply power to the frog, and I realized I could make the power more reliable if I connected together the common rail from all of them, and connected each set of turnouts that are in a common block. It also provides a point to connect the power wires to without having to solder more wires to each track.



So the next step is getting the track around the bridge prepped, do the final alignments on the track that crosses the gap between shelves, and then I should be ready to get DCC hooked back up again so I can test how the locos run between shelves and across the bridge.

Someone who was on the forum said there was a video online on modelling water with toilet paper and Modpodge Gloss. I mixed water and white glue 50-50 soaked the tp in it and laid it down. I used a small paint brush to create a ripple effect and let it dry. I think I waited a day or due and then painted it and let it dry again. I then applied the Modpodge allowing each layer to dry before doing the next. I never had any problem with it being tacky or brittle. If you can locate video it will walk you right through it or you could try a search on this forum.

PS: Back then I remember that I made a comment about calling the water Schitz Creek. I did a forum search on that and found the thread. It's under New England Model Railroad and post #247 gives the Youtube link and a better description.

I've been going through ideas for the past week on how to make the bridge track removable but still conduct power, and I think I finally figured something out. I have just enough room to work with, so I put regular rail joiners on one side of the track so I can slide the bridge track in place. The other side needed a way to hold the rails without capturing them. What I ended up doing was taking a rail joiner and cutting the flange in the middle on one side so I could open up the top half of the joiner. I opened up the opposite side on a second rail joiner. The picture isn't too clear, but this left an open flap on the inside of both rails, while the rail joiners remain untouched on the outer side of the rails. So when I drop this other side of the bridge track in place, I simply pinch the rails together so they slip inside of the rail joiners. It's enough to keep the rails from moving and keep them fairly well aligned with the track that is already glued in place. Note you can also see the wires I added here to electrically join the track on either side of the bridge.



So that was a bit difficult to put together, but once I confirmed it was working as intended I dropped a small bit of solder on each of the four rail joiners to attach them to the glued track, and now I can pull out the bridge track at any time.



If you check the alignment of the rails in the next image, you can see that they don't make a nice smooth curve yet. This is mostly because of all the missing ties plus the last couple ties on each end coming loose from the caulk, so the track wants to return to a straight alignment.



Not to worry though... The next step to finish up the bridge is to extend the beams that the ties sit across out into the embankment. This will give me room to add four more heavy bridge ties on each side to fill the gap, then I can put in a small track nail along the outside of the curve to hold the rail where it should be. And of course once the track is glued to the bridge it will help retain the curve there too.

I need to re-check clearance on the bridge as well. It seems to only barely have room to move a flatcar through. I've already made the cross beams wider that the bridge blueprints I was modeling from, but it's pretty tight. I think the loco is going to be the widest thing going through, and it did have clearance when I had this sitting on the desk, so I probably just need to adjust the track position.

Anyway, I'm happy to finally have the track in place so I can finish up the bridge. At some point in the future those rail joiner flaps will be rolled back into the rail and soldered in place, but for now there's a lot of scenery that needs completed, including the river itself, but meanwhile I should be able to drive the locos over the bridge while the track gets tested.

Wow... 7 months. I knew it had been awhile since I had time to work on this, but really? Well I'm happy to report that some small amount of progress has been made finally. I got the last two turnouts prepped yesterday and attached them this morning. I also got one of the sidings attached, and just about have all three tracks for the yard ready to lay down. It's going to be a busy weekend but maybe I can get those finished up next week. Then that only leaves the elevated hopper siding, but that will involve building some trestles so it will wait for a bit.

The next step will be working on some block sensors. I have a few different types of electronics to test out, but I want to figure out what will work before I start wiring up this half of the shelf. I also need to do some experimenting with IR sensors. Between those two devices I should have pretty good feedback about where the trains are at.

I also realized the bridge is still too narrow, and after looking at Latestarter's design I think I can update mine to be more realistic.

So... back to making baby steps, but at least something is happening.

And a month later I finally got the yard laid down. I actually prepped the tracks around the time of my last post, then they've just been sitting there waiting. The spur in the back corner is the only remaining piece of track left but it will probably be awhile before I get around to building the trestle for it. I'll probably go ahead and lay down the flex track on the ramp leading up to the trestles though, then I can do all the wiring and some initial painting.



You may think that the three yard tracks look awfully close together, and you would be right. They are spaced 1-5/8" center-to-center, leaving only 3/16" between cars. Certainly no room for fingers, but I don't have to worry about a long reach or even very long trains as the spurs will only hold about three or four 40' cars. I decided I wanted to make a little more room away from the edge and from the mainline before I built the turnouts, and I think it's going to look really nice when the yard is filled with hoppers.

I have some other things to finish up around the house over the next three weeks, but eventually I will start experimenting with current sensors and try to decide what I'll use for block occupancy detectors. After that I'll work on the IR detectors (I have a 38kHz variety and want to see how well they work to bounce the beam off the bottom of the cars). Then I'll finally be ready to wire up the servos on this half of the shelf, and get the DCC controllers set up again.

Lemonhawk said:

Look at the Digitrax BXPA1 and BXP88 for track occupancy and transponding operations.

Click to expand...

I'm kinda lost with all the loconet stuff, can you fill me in? Do the transponder functions with with my existing DCC decoders like the tsunamis I have in my more expensive locos? And how does that report the loco locations back to my computer?

For the occupancy portions, I see it detects cars fitted with wheel resistors, but what's the resolution? Will it give a fairly accurate count of wheel sets in the block once I tell it approximately how many microamps each resistor draws? For comparison, I figure each 10k resistor will draw around 1.5mA and the sensors I'm working with have 15-bit accuracy, so if I have a maximum of 5A per block, the sensors will detect down to 0.15mA. Of course there's going to be huge fluctuations once the loco enters the block, but I can keep a count of the cars entering the block before the loco arrives or after it leaves.

There's also the question of cost. I couldn't find how many blocks the BXPA1 works with, but the BXP88 is rated for 8 blocks. Even if the BXPA1 is also set up for 8 blocks, it seems a bit pricey at $50? The sensors I think I'll be using are $2 each and will detect two isolated blocks each. Paired with two $5 ESP8266's that comes to about $26 to handle all 15 blocks on the shelf. The ESPs can also manage the IR LED sensors in the same area and send all the info over wifi. So that's kind of my baseline for the minimum of what I expect any system to do.

The BXPA1 is an autoreverser and only detects that one block. BPX's are also circuit breakers by the way. They work by just detecting when current is being drawn, the BXP's can't detect individual cars that are in the block. If your using JMRI the BXP's report to it, if your rolling your own then they have an output that will drive an LED or an input to an Arduino. Some occupancy detectors use current coils where the power lead to the block is wrapped around a coil a few times and the circuit senses when there is current flow, some use back to back diodes and note if there is any voltage across the diodes and hence a current draw. There are also IC's the detect current and convert it to a voltage that a A to D can read. Lots of ways to do this, but counting cars in the block would not be easy, even if you knew what each cars resistor was, because of all the other places where resistance will be present (dirty wheels, track wire length ...). I only bring up the BPX's because it['s a known system that works with JMRI and DCC.

Ah yes, I forgot to account for dirty wheels and such. Well at least the system I'm planning to set up has 10x the resolution needed. I figured each ESP8266 can simply count each step in current draw, regardless of the exact value, to keep track of how many wheel sets appear to be in a block. This should work up to the point where the loco enters the block, but after that I think the fluctuations from the motor will swamp out the ability to detect individual wheel sets.

What I'm hoping to do is add an IR detector at about 3" from the end of each block, which tells the loco it is time to stop what it's doing before it goes off the track. I might also add block detection to the frog within each turnout. Between those various sensors, I'll have enough information to write up a script to have a single loco navigate through all combinations of turnout positions and attempt to map the entire layout. Once I have a table of which blocks neighbor each other, and which turnout positions lead to which other blocks, I should be able to use that to determine that if a train is in this bloc, moving in this direction, the next expected block should be this one... and thus keep a running record of the exact location of each train. (I'll have to expand this a bit to include trains that occupy more than one block at a time.)

This information will also be useful in generating routes -- if I want a train to get from point A to point B, the code can return what path should be taken. More importantly, if any of those blocks are already occupied then I could generate an alternative route (if it exists). The real fun begins when I try to also add collision-avoidance. I'll start by adding signalling for the surrounding blocks, and priorities for each train. But then I have to figure out solutions, such as when two trains approach each other and are stopped -- do I resolve this by having one train back up, or is there an available siding for one train to move to? And that will lead to code which attempts to plan ahead, keeping a timetable of when it thinks each block will be occupied, avoiding a situation where two trains could ever collide.

In my setup, DCC will probably be the only standardized system I'm using (running the locos and turnouts), everything else will be using my own circuitry and code. I'll probably have to include a raspberry pi for the heavy code (I'm thinking of using MQTT to keep a record of all the data), but hopefully I can have the ESP chips handling a lot of the work interpreting the data being received.