[Mark VerMurlen]

I really like the new yard. Playing with the simulator in Scarm I found it much more functional.

Functionality is what counts in a yard, so if you like the way the new one works, then I think that’s the way you should go.

Mark

 

[Mark VerMurlen]

Looking at the layout it has the two complex intersections. I've been looking at alternative options to the single and double slips I have selected.

View attachment 400682

The only way I've figured how to remove the double slip on the reverse loop intersection is through the use of a 3 way turnout. Not sure if this is a better option, it adds a S-bend to what would be a heavily used route.

The other intersection uses a single slip, and could be replaced with a cross and two additional curved switches.This doesn't effect traffic or throw anything out of alignment, it may be the better option.

What are your thoughts?

I don't have experience with single or double slips or even with a three way turnout, so I can't really give any opinions on reliability of any of these. Hopefully some others with experience with these track pieces can jump in with their opinions.

I've been looking at these 2 junctions for a couple of hours. I can't quite put my finger on what bothers be about them. You've labeled your red track as "mainline", but I don't really think that it is. Its really serving 2 different functions. Its your entry to and exit from your staging. I also think you'll be using it as a passing track more so than as mainline. So my question to you is do you agree that's how you'd use this track? If so, can you separate the 2 functions and thus simplify both junctions?

If you want to keep all the route options you currently have, I think the left hand reworked junction works fine. The right hand reworked junction has an odd bow on the outside track leading to the staging area. I think if you move the last turnout before the staging area farther away from the main junction, you can get rid of that weird bow and avoid the S curves that result from it.

Mark

 

[Thelic]

Mark, I agree the two junctions are a bit of an oddity, and not at all prototypical in this region. They do save a ton of space though.

I guess we should discuss operations.

In the prototype intermodals are somewhere in the 6 to 8 trains a day range, while coal and grain are a little less. Some of these spend a little time in the siding beside the yard before continuing their journey, but none of these trains are modified in town.

The prototype has several team tracks in town, they unload things like salt for highway maintenance (my business) and cement powder.

Town also has a bulk plant next to the yard, and a scrap metal yard with sidings on a branch line.

As mentioned previously, in the past town had used rail for moving chips up the branch line to a pulp mill and lumber from the sawmill to the port.


As for my operations...

I had envisioned on eventually running automated thru trains from staging on the red route, that's why I referred to it as "mainline".

The blue line would serve as local short term storage for these trains. The requirement then would be to be able to store an intermodal train without blocking the A/D track.

My Kato Maxi-I's are 22.25" long each, so three of them and two 5.5" Kato GE ES44AC's so the siding would need to hold about 78" or 72" if using a two Maxi-I's and a Maxi-IV. Coal and grain trains could easily be shortened to the nearest car under 72".

Local operations would consist of accepting a mixed cargo train, breaking it up and delivering cars to their respective industries while avoiding delaying the thru traffic, which the prototype values highly.

Basically, receive a train from staging into green A/D track, break train in half or thirds to avoid fouling main and sort into yard. The yard has room to comfortably fit twenty four 50' cars.

Then deliver cars as required to industries on the branch lines exchanging full for empty or vice versa. The industry off the main provides an interesting scheduling challenge. But this turnout exists in the prototype as well.

Finally build train from cars returned from industries, turn it around and dispatch back to the staging yard.


And of course, for maximum railfaning you can run out of the staging yard, around both reversing loops and back into the staging yard on the opposite side of the layout.



So yes, I had regarded the red line as the mainline as it connects to the outside world.

The blue functions as a long siding and turn around. The prototype uses a wye, but its far less continuous run friendly.


Is there a better way to run this?

 

[Thelic]

Took a little while, but I'm much happier with it than when we started.

I reworked the yard to have an always accessible run around/engine maintenance track. It has room to comfortably run around three 50' cars. The spur for the bulk plant is also off this route.

The main level is 6.5" above the staging yard, this should give access to the staging yard in excess of 5.5". This should give me enough room to install switch machines. it will be difficult to be sure though.

I'm still not sure what the best switch machine to run would be. I like the idea of having a DCC decoder on each switch, along with micro switches for frog power and LEDs. I was considering Tortises, but the digital Cobalts seem to be more feature packed. I see the Smails are also comparable to the digital Cobalts.

Any advice?


I guess my next step is to block this into detection sections...

 

[Thelic]

Alright, so I've diagrammed the track (excuse the MS paint skills, or lack there of). I've also been doing a lot of reading.

So the way I understand it is any length of track that could potentially hold a train should be its own block.

So short runs between switches are not necessary to block separately and can be combined with the switches that make up that intersection? Basically the "On Switch" block encompasses the entire intersection.

So then for stopping trains at signals accurately within a block using JMRI each block would be broken down into at least 3 sub-blocks (for travel each way) allowing the computer to sense where the train is in relation to the end of the block? (I haven't drawn these in yet)

Is it possible to accurately stop a 72" train in an 80" siding using block detection and JMRI, as well as do it slowly?

And then getting even further into it, using transponding you *should* be able to route all 3 sub-blocks through the same detector?

Unsure how to deal with the spurs. Do they typically get their own block, or are they just considered part of the parent block?


I'm going to need like 16 different colors of wire...

 

[Thelic]

So after sectioning I end up with 12 Blocks, each with 3 or 4 sub-blocks.

If I use BDL 168s for detection and RX4's for transponding I "only" need 3 BDL 168s and 3 RX4s.

Each RX4 handles the common wire for a given block, and each block is split into up to 4 sub-blocks. This should give enough information to determine what is in each block and where it is in relation to stopping points.

The junctions don't have any detection, I'm not sure whether to stretch the blocks into them or not. Guidance would be appreciated!

Please stop me if I am wrong!

 

[jackpresley]

I'm a big believer in simulation, which I assume SCARM does from one of your earlier posts. My eventual goal is similar to yours as far as automation is concerned. Simulations of my designs usually end up with a lot of surprises that I sure would not have wanted to have during the actual build.

Good luck with your design and build.

 

[Mark VerMurlen]

I was never able to figure out JMRI, so I'm using TrainController as my automation controller. So everything I'm going to say is based on TrainController which may or may not carry over into JMRI.

I chose to make my spurs separate blocks. Not because I plan to park full trains there, but because I want to move a locomotive and a car or two into and out of it automatically and maybe park there as well. If you put slower speed limits on those blocks, the trains will slow down before entering them which helps with realism.

I did not break all my blocks up into 3 (or more) sections. TrainController can profile the speed of locomotives so it can fairly accurately stop a train at a certain spot, plus or minus and inch or so. Most blocks don't need that amount of accuracy. I do have a couple of blocks that I'm considering putting additional sensors into because I'm running train lengths that just barely fit. If you break your blocks up into three sections like you have, you'll get absolute accuracy, but its also taking a fair amount of hardware to get that. Its up to you if you need that everywhere or not. I would certainly never break a block into 4 because it really doesn't serve any purpose that I'm aware of.

I chose to include my turnouts into my blocks. The going recommendation on the TrainController forum is that you shouldn't put the turnout in any block. I didn't know that at the time I wired my layout. I really haven't noticed any serious problems with what I've done, so I'm happy to leave it as is. Note that if you put a turnout into a block, it must go into the single block on the points side of the turnout and not into either of the 2 blocks on the other side. I would encourage you to read some of the threads about this on the TrainController forum. There are some issues with not detecting on the turnout. When using current sensing like the BDL168s, the electronics of the detector will have a slightly lower voltage than the raw bus feed. This can cause some issues with accurately knowing when a train enters a block because it can sometimes draw power from the higher voltage track even though its partway into the next block. If you always have a small sensing zone on each end of a block like you do currently, this won't matter to you. You can also add diodes to the turnout power feed to knock down the voltage slightly. I'm not an expert on this, so that's why I encourage you to read the TrainController forum on this topic.

I don't have any RX4s on my layout, so I don't ever detect which train is in a block. On my layout, I had to initially tell the software which train was in each occupied block, but I've only had to do that once. The software tracks the train as it moves from block to block because its thrown the turnouts so it knows where the train should be. I think the only place on your layout that you really need the RX4s is on the A/D track outside your yard. The advantage of putting it there is that the software can then read the train identity when it comes out of the yard and it can then track it from there without the need for any other identification sensors. If you didn't put it there, you'd have to manually identify the train for the software. The other place you might want to put the RX4 sensors is on your staging tracks. If you plan to manually swap out locomotives in staging (or maybe spurs too), then it would be advantageous to be able to automatically identify the locomotive in those blocks. I personally don't think there's any need for identity checking anywhere else.

Rather than use a bunch of different colored wire, I chose to use tie wraps with label heads to write the block number on. I put a few of these tie wrap labels on the wire periodically so I can identify them. All my rail power is black/red wire, my switch machines are white/yellow, my auxiliary DC power for lighting is orange/gray. Keep a drawing and spreadsheet with your block assignments, BDL168 assignments, turnout motor assignments, signal assignments, etc.

Hope this helps.

Mark

 

[Thelic]

Thanks for the words of encouragement Jack.

Mark, as always you're a wealth of information.

If the software is really that good at monitoring trains that makes things much easier. Encouraging to know you've only have to ID them the one time.

Some of my blocks are very tight for a 72" train to fit, so I think I'll stick with the small blocks at each end.

The reason for breaking a block into 3 is to get those small end blocks as you've surmised. However breaking the block more times gives you the ability to have chains chase each other within the block, I figured the long decent through around the rear of the layout and down the 1-1/2 turn helix might be nice to break up since its over 4x the maximum train length, though maybe that's just overkill.


I will have a look over that forum.

 

[Mark VerMurlen]

If the software is really that good at monitoring trains that makes things much easier. Encouraging to know you've only have to ID them the one time.

I've maybe stretched the truth slightly, but the software really is quite robust. The software is designed to work with a wide variety of sensors, like momentary contact sensors, so it has to be robust in tracking trains. If the train "picks the point" and actually goes down the wrong branch of the turnout, the software will be confused because it will detect occupancy in one block when the train is expected in the other. I've had this happen rarely. When it does, chances are the locomotive and some cars are likely to derail anyway which means you have to stop the software. In the software, I may then have to manually drag the train from one block to the other. But it is true the train is only configured once and I may have to just drag its position from one block to another. I also don't tend to manually reconfigure my trains a lot, so I've not missed automatic train identification.

The reason for breaking a block into 3 is to get those small end blocks as you've surmised. However breaking the block more times gives you the ability to have chains chase each other within the block, I figured the long decent through around the rear of the layout and down the 1-1/2 turn helix might be nice to break up since its over 4x the maximum train length, though maybe that's just overkill.

In this case, I would break up the helix and lead lines into 2 or 3 equal size zones and make them logically independent blocks. I'm not sure about JMRI, but TrainController defaults to only allow one train in a logical block at one time. You can override this, but I think it should be used cautiously in special circumstances. So if the helix & feed is 2 or 3 blocks, the software will automatically chain trains going around the helix without any special handling.

Mark

 

[Thelic]

I've been looking at the brochure and browsing some videos on train controller. Seems like for the price it could save a massive headache.

Don't worry, I didn't take that ID once thing literally, but I get the idea. No need to constantly ID the train every block. Just when using the 0-5-0 switcher.

Mark what do you use as far as hardware? I was thinking the simplest route would be Digitrax everything. The BDL168 seems like a fantastic device, and the SE8C also seems ideal. I just really dislike their throttles!

What method would you recommend for turnout control? Switch machines with individual decoders or centralized with DS64s?

That said, insulfrog or elecrtofrog? Bare in mind I'll likely only be running 6 axle diesels, maybe one 4 axle as a passenger train at some point. It would be a Kato EMD F40PH, with all wheel pick up.

 

[Mark VerMurlen]

I have the Gold version of TrainController, which is actually very expensive. I was originally very opposed to spending that much money on the software, but in the long run, its been worth it since TrainController can do so much and is relatively easy to setup and use.

I have all Digitrax hardware, except for the Loconet to USB adapter which I bought from RR-CirKits based on reviews I've read about it. I figured that since this was my first DCC layout and I wanted full train automation, I wanted to go with a single manufacturer since that would give me the highest degree of compatibility. There's also a fair number of examples around the web that use it. I also like the Loconet communications mechanism and its ability to be extended in many ways, such as using Arduino or other micro controllers. Its worked out well for me, but I know that there are others on this forum that promote other systems and hardware suppliers.

My DCC Controller is the DCS100. I use PM42s for power distribution and short circuit management. I have BDL168s for occupancy detection. I use SE8Cs for turnout motor control and for signaling. (I'm still working on building signals and don't have any of my signaling working yet.) I chose to use Tortoise turnout motors since my past experience with solenoid motors has been problematic. The SE8C can directly run Tortoise motors for turnouts without the need for other hardware. Be aware that Digitrax has released some newer hardware that integrates short circuit, detection, and train ID into one new board. I don't have any experience with it, but it sounds good.

I fully understand your dislike of the Digitrax throttles. I started with a Zephyr DCS51, which I liked, but I don't like their other throttles. I'm currently using digital throttles on my touch screen computer or my iPhone/iPad. I'm not totally happy with the digital throttles because they lack the mechanical feel that I want. I'm still trying to figure out what my next step might be regarding throttles.

Mark

 

[Mark VerMurlen]

One other thing, I think that TrainController can actually operate 2 control systems simultaneously. So I think it might be possible to have something like an NCE throttle and command station with Digitrax Loconet based detection, switch machine control, and signaling.

Rudy and Fredrick on this forum are much more knowledgable about TrainConroller and would know better about if this is possible and how to do this.

Mark

 

[Thelic]

I've been reading the manual, looks like for transponding you need at least Silver. It looks like it will be way easier than JMRI to implement.

I'm going to take another look at my layout, I think I could use a BXP88 (8 detection & transponding sections) to detect 4 staging yard tracks, and the A/D track. Not sure where the rest of the sections would be used best. I may just leave them open.

The remainder would simply be BDL168 detection sections without transponding.

How does the software handle complex crossovers? Is it smart enough to allow multiple routes through my complex intersections at once?


So SCARM tells me I have 31 switches. One of which is a double slip (2 motors), another is a single slip (2 motors), and a third is a three way (2 motors). So that brings me to 34 switch machines. Yikes!

That said there are a number of crossovers you could drive with a single motor, not sure if this is easier, or just a headache in the long run. Opinions?

If I remove the yard for construction at a later date I'm still at 25, which is still four SE8Cs so I may as well bite the bullet and go with the five I'll end up requiring to avoid silly wiring later. Ground throws with electrical switches are also only a few dollars cheaper if you need to power frogs, so might as well just do it with a Tortoise.

 

[Thelic]

Woohoo! Trains have arrived!

Surprised to learn that the Maxi-I car is not in fact 22-1/4" long as claimed in the brochure, but 20-1/4" long. This opens up some serious space on the layout! I can simply go with 3 Maxi-I cars instead of 2 and a Maxi-IV and still come in slightly over 72" with a doubleheader.

All the cars are surprisingly heavy and run fantastic with the metal wheels, a single engine pulls the trains effortlessly.

These Kato engines are a dream, I've never seen something so small so smooth. The couplers...not so great...

Don't know if it how I'm pushing the cars together or what, but they seem to resist going together, once together they stay, but I'm having to set each car down into the other.

Does adding the magnetic decoupler help? I haven't added it yet.

 

[Mark VerMurlen]

I've been reading the manual, looks like for transponding you need at least Silver. It looks like it will be way easier than JMRI to implement.

I'm going to take another look at my layout, I think I could use a BXP88 (8 detection & transponding sections) to detect 4 staging yard tracks, and the A/D track. Not sure where the rest of the sections would be used best. I may just leave them open.

The remainder would simply be BDL168 detection sections without transponding.

How does the software handle complex crossovers? Is it smart enough to allow multiple routes through my complex intersections at once?


So SCARM tells me I have 31 switches. One of which is a double slip (2 motors), another is a single slip (2 motors), and a third is a three way (2 motors). So that brings me to 34 switch machines. Yikes!

That said there are a number of crossovers you could drive with a single motor, not sure if this is easier, or just a headache in the long run. Opinions?

If I remove the yard for construction at a later date I'm still at 25, which is still four SE8Cs so I may as well bite the bullet and go with the five I'll end up requiring to avoid silly wiring later. Ground throws with electrical switches are also only a few dollars cheaper if you need to power frogs, so might as well just do it with a Tortoise.

TrainController does support single and double slips. I'm not sure about 3 ways, but that's easy to mimic with 2 turnouts. So I'm sure it can handle your complex track junctions. As long as you have parallel independent routes through your complex track junctions, TrainController should be able to pass simultaneous trains through those junctions.

I was looking more at your track block plan. I think you need to look again at your A/D track. You've got a turnout in the middle and at the end that go into your yard. Since the rule of thumb is to not make a turnout part of any block, I think you need to break your A/D track into two separate blocks on each side of the center turnout. I'm not an expert on this, so I would encourage you to ask questions about the best way to split your layout into blocks on the TrainController forum. If the end turnout on the A/D track isn't in the A/D block, then a train can actually pass from the mainline into/out of the yard without ever hitting the A/D block. I think that means you'll want another transponder on that blue mainline block then.

Do you know if you'll make all your cars detectable or not? If you make them all detectable (by adding resistors between their metal wheel sets) then you'll definitely want your spurs to be independent blocks so they don't make the mainline track appear to be occupied when a car is sitting in the spur.

You've got a complex track plan which requires a lot of hardware, both switch machines, controllers, and detectors. So its not a surprise that you need a lot of hardware to make this all functional. Because of this complexity, I again urge you to go to the experts on the TrainController forum. I'm expecting you'll get more engagement there than you've been getting here. I'd just hate to see mistakes made because only the two of us are looking at this and not a wider group of experts.

Mark

 

[Thelic]

So I've been "testing" my models, seeing what looks good, what operates well. Trying to figure out exactly what I want to accomplish.

Also picked up a second pair of Kato Es44ac's bringing the fleet to 4! Found some of the older ones available online so I've got road numbers #2801, #2825, #2814, and #2822. That gives me enough power to double head both sets of cars.

Some clarifications. The Kato couplers on the engines and Maxi cars are great, its the ones on the coalporters that don't work for coupling. They really need to be dropped into each other (as the directions show), not really good for switching operations. They will not go together by pushing the cars together, they have a shock absorber built into them, which works great when pushing a long string of cars, just not for switching.

I've been watching a lot of layout videos and doing a lot of driving for work, so I've gotten a fair share of railfanning in. That said, I'd like to be closer to the prototype.

So some things I'm going to change:

1) Minimize switches, extra cost, extra maintenance, same as the prototype. Eliminate all "fancy" switches, moved to ALL #8 frogs.

2) Longer visible mainline run, I'd like to make use of some signals eventually. Having so much hidden doesn't really work for the visual appeal.

3) Double ended yard, the local yard is double ended with one side having a long yard lead. I'd also like the yard tracks to be able to handle a full length train, 76". This also mean I can do more staging on scene, and eliminates the need for hidden staging.

4) Passing siding, double track mains do not exist in my region. I'd like the challenge of using a passing siding rather than just run them around on their own tracks.

5) Hidden reverse loops, these look rather unprototypical for my region, but I like the idea of sending a train off layout then having it return flipped around. I'll probably double track the loops so that I can have a train staged in each loop while not blocking thru traffic.

So I've come up with the following...

The red portions with tunnels will lead to lower level reverse loops.

Let me know what you think, I'll probably adjust some of the spurs, but that's the location they will come off the main.

 

[Thelic]

So it all fits, 2-3/4" clearance between top of rails at all crossing points, lots of room for switch machines. Should be able to build it from 3/4" plywood or spline.

Ruling Grade: 3%
Minimum Visible Radius: 17"
Minimum Hidden Radius: 14"
Mainline Run: ~400"
Maximum Train Length: 80"


I think I may have room to pull off a small container facility in one of the lobes. The bottom lobe is actually situated very nicely for switching from the siding. You can break up a full length train using the siding as a drill track.

The left lobe can be switched using the siuding and bottom lobes single track as a drill track. This was one of those "happy accidents".

Don't really like the two spurs into the middle of the layout, may just join them into a 5th track in the yard. This would allow me to cut the benchwork further into the corner which will help with reach.

Every iteration I get happier with the plan. Just need to finish the tile in the basement to get the final go ahead from the CFO.

 

[DonR]

Thelic

I agree, lots of interesting train running plus
switching challenges.

However, your latest additions have created at least
3 reverse loops, if I follow your colors correctly.
Not a problem if DCC. A Digitrax PM 42 has capability
of controlling up to 4 isolated reverse loops.

Don

 

[Mark VerMurlen]

Your latest plan looks good. I would consider using the 2 track spur in the middle for an engine house. The other middle spur could be something like a refueling and sanding siding. Since those spurs are along side your yard tracks, I would pick yard services to put on them.

While Don thinks there are 3 reverse loops, I only see 2 (the balloon shaped returns under the main layout). You do have fairly long leads to get to your reverse loops, I would consider trying to shorten those up so that it doesn't take so much time for a train to disappear and return. You can always extend the time by slowing the train or stopping it, but its hard to shorten the time when you have a lot of track to traverse. I realize you need enough length so you can drop below the other tracks at a decent grade, so maybe you do need all that length. Just something to try to optimize if you're getting down to the nitty gritty details.

Mark