Monday, February 22, 2010

George Washington's birthday

Lots of little things have happened since the last post on progress. I made a trip to Caboose Hobbies last Tuesday, on the way to Denver airport, then the kids and I made another trip on Saturday. As the momentum builds on the layout I now have 3 kids with definite interest (the other could not care less, but that's cool, too).

So let's see what has happened in the past week:

  1. I built a short test section of scenicked track, using leftover blue foam and a bit of leftover Peco track. In the process I learned a few more things that don't work plus some that do, and now I have a test section that looks decent. I will cover the process I used in a later post after I've completed the ballast for first section of track on the main layout.

  2. Started the track scenery work on one section of the main layout. In the process I realized I needed to finalize and document my scenery plans for that area of the layout. You see, you can put ballast on the track, but properly done it will "spill over" onto the track side ground. This means you probably should have the track side ground scenery in place, or at least know what it will be so that you know you won't need to do this. If you start thinking about that you then realize that you should also know the exact dimensions of the right of way, which means you probably need to understand the whole scenic plan for that area. So, I've been making drawings and doing research on prototype rights of way and also of the dimensions of things like streets, blocks, buildings, etc. in similar midwestern cities (Google Earth has a nice tool for getting exact linear dimensions.)

  3. Lots of decoder work with Daniel. The big effort has involved his Bachmann 4-4-0 from a train set he got 4.5 years ago, a tiny locomotive with a design not updated since 1979. No one in their right mind would add a decoder to such a locomotive, but we have a cunning plan -- and actually it's working really well. We've taken progress pictures and will describe the process in its own post when we finish. Daniel found a way to isolate the motor from the track pickups without having to take the tender (where the motor resides) apart. We plan to house the decoder in the first box car behind the tender. At this point all is working except we need a permanent solution for the box car to get power from the track. We asked for advice on this at Caboose Hobbiesm but they told us to call on Wednesday when their N scale expert is back.

  4. Daniel also installed, with my help, a TL1 decoder in a tender of a Proto 2000 2-8-8-2. I installed a DZ123 in this back in 2005 but the rear light in the tender receives direct power from the track, with a small diode/capacitor light board to make sure the light only was on when the DC power was in one direction. On DCC the light was on all the time. The rear light now is controlled by the DCC F4 function, and Daniel did almost all the work himself.

  5. Athearn support has been just great regarding decoder problem with the HO scale F59PHI locomotive that I mentioned last week. They sent us a new board which arrived on Saturday, but unfortunately the symptoms persist with the new board. Sent them a follow-up email this morning.

  6. Daniel successfully installed 4 other decoders in various Kato and Atlas Santa Fe locomotives. We can now run the Super Chief train that we got in 2006, tested once on DC, and hadn't run since! Unfortunately, the motors for the F7Bs both are now failing to run intermittently. We've taken them apart and verified the problem is the motor itself. Kato doesn't have a support line like Athearn, Atlas and Digitrax so I'm stuck right now. I'll post a question to an on-line group.

  7. We started trying to figure out how to add a decoder to an Athearn 2-8-0. Couldn't find any instructions on line, except a couple comments that it can be done. I bought a TCS M1 decoder at Caboose Hobbies because this is thought to be the absolutely tiniest decoder available, and it does fit under the plastic coal load in the tender if you scrape out some of the excess plastic from the casting process. It looks like wiring up the motor won't be too hard, but the front light appears to be independently powered from the engine's pickup wheels, so I'm not yet sure how we'll wire that from the decoder in the tender.

  8. Daniel solved a clicking problem with a Kato F7A locomotive. Found some debris that got caught in the gears above one wheel. Good to get this experience, as we are both gaining confidence in our abilities to address N scale locomotive problems. We are certainly not experts yet, but we're not novices anymore either.

  9. Picked up some scenery books at Caboose Hobbies, including one on backdrops that my oldest daughter, Paige, picked out because she is interested in painting them.

  10. I also wrote some posts earlier last week while traveling, including a third one on the detailed design (for the freight yard) that is nearly ready for publication.
For the coming week the focus continues to be scenery. The goal is to complete at least the track and track side scenery for the SAMR line -- about 21 feet, double-tracked, with 9 switches. Almost certain not to be completed in one week, but want to have portions that are complete.

    Wednesday, February 17, 2010

    Evolution of the iNdoor design, part 2

    At the same time that I was using CAD to draft the detailed layout design I also maintained a written description of the purpose of each section of the layout. I archived these descriptions, by date, so they are now useful in reconstructing how the layout evolved.

    During the winter and spring of 2006 I worked on the layout design almost every day. My last post showed the layout design as of February 27. By March 3 I had updated a portion of the lower deck design as shown here:


    Compared to the diagram in the last post, one difference that is immediately noticable is that there are now text labels for parts of the layout. I would expand on this practice over time with more labels and detailed descriptions within the CAD drawing.

    You may note that the E-W main line was labeled "BNSF" and the N-S main line was "CN (Canadian National). The designation of the N-S line has varied over time as the layout concept has evolved. I like the CN railroad, and the layout will definitely feature it, but my current thinking is that the N-S line is owned by the SAMR (municipal railway), with trackage rights for every railroad that services the metro area. Similar to Chicago's Belt Railway.

    Looking at the above diagram in more detail you will see that the trackage on the east (left) and south (top) walls has been completely revised. I felt that a scene of the BNSF crossing the main river was a must, but I couldn't figure out how to make it work. My first approach was to somehow try to fit the river in the upper left corner, on the grounds that this would leave more space for the other railroad scenes. Eventually I dismissed that approach, both because the resulting river was too narrow to represent the major, navigable, river that the city history required, and also because the 36" minimum radius meant that the bridge track would have to be at least partially curved, something that would have been avoided in real life.

    So, I decided to try locating the river along the left wall. This meant sacrificing some potential industry there, but the resulting track was simpler and cleaner and thus aesthetically more appealing, at least to me. Moreover, the idea that the river would be at least 6' wide (960' in scale) meant that the bridge could be a signature scene on the layout -- if done well it would be something that people noticed first and admired. Then, I considered that immediately above the bridge scene, on the upper deck, was going to be the most active part of the freight yard. One tenet of two-deck design is that you want to balance the areas of intense operator activities, such as freight yards, so that no two such areas are located directly above/below each other. Thus, a river bridge would be an excellent counter-balance to the freight yard above.

    All in all I was -- and am -- very happy with the river and bridge concept. For the south wall I decided to include a short (5') commuter station, and the downtown intermodal yard. Again, I'm happy with the location of these items, although the final trackage would be significantly changed from what is shown in the above diagram.

    The final note from this March 3 iteration was a comment I wrote to myself about whether to transpose the identity of the main lines. I was concerned that this ostensibly BNSF-focused layout had the BNSF on only one deck. As it turned out, this concern would soon be answered in a different fashion.

    My next step would be to address the freight yard design, which is a big enough topic to justify its own post.

      Tuesday, February 16, 2010

      Evolution of the iNdoor design, part 1

      Previously I talked about the high-level design and the decision to use CAD for the detailed design. After I'd been using the CAD tool for a while I began archiving the design files with the date in the file name for posterity. I'm glad I did this, as these old files help me understand how the design evolved.

      Here is a picture of the lower deck design on February 27, 2006, the earliest file save I have a record of:


      Grid lines are 1' apart. This is comparable to the high-level design sketch from that earlier post.
      At this point in the design process I was still getting used to the CAD tool. I hadn't started using text labels, colors, or providing any details except the outlines of the benchwork and the main track. I also hadn't done much work on staging yet.

      Most of the track design at this point was nothing like the final design, but one area that was very close to final was the Union Station on the north wall:


      I put a lot of thought into how many tracks Union Station should have. My concept of commuter train operations wasn't as well formulated then as it is now, per the layout concept description I recently posted, but I did know I wanted enough tracks to support heavy traffic and make operations interesting. This was a highly subjective decision -- the sort of thing about which reasonable modelers can disagree. Eventually I decided that 6 tracks would be too few but 8 just enough to give a sense of a busy Union Station.

      The next challenge was fitting everything in. Each station platform had to be 10' long, per the maximum passenger train length. Granted, the maximum length for the commuter trains is only 5', but as this Union Station has a long history I couldn't seen any rationale for shortening the platforms from what would have been built in years past. (Reducing the number of tracks from years past does make sense in order to reuse the land, but reducing existing platform length would have had little benefit as the tracks would still be there.) When you then add the need for switch ladders on both sides, plus 36" minimum radius curves at each room corner, PLUS the need to have room for double track main adjacent to the station tracks, it was a struggle to fit it all in.

      Fortunately, one question I'd toyed with during the concept phase was whether to have the station be stub (tracks that terminated in the station) or through (tracks that extend in both directions from the station). At that time I studied existing and past Union Stations, and noted that Kansas City had a mix of stub and through tracks, and thought that might be a neat idea. It was important to note that for the modern St Albans Union Station 3/4 of the traffic would depart in one direction -- towards the junction -- so it made sense to me that half of the 8 tracks would be stub ended, servicing only that direction. This would also make operations planning more interesting, as it introduces a new constraint regarding track assignments. Finally, from the design perspective, this meant that the switch ladder on the east side (left in the above diagram) would be significantly shorter, thus giving me enough space to fit everything in.

      I've mentioned before that I love the intracate trackwork leading to congested city Union Stations -- not just Chicago but other cities as well. On the east (left) side I used curved switches to save space. On the west (right) side 3 double slips were used to help create a double crossover as the main entrance to the station. This is but a tiny fraction of what is used in Chicago, but it does give the feeling of complicated, congested trackwork that I was seeking.

      This next picture focuses on the northwest corner of the lower deck (the lower right of the diagram). Although all of these tracks were redrawn at some point the basic schematic still remains in the implementation today:


      The main junction between the two double-tracked lines is shown at the top. One main line is shown going diagonally from the lower left towards upper right (this is the N-S main). The other (the E-W main) may be difficult to identify because there are so many tracks in the picture. The three tracks to the far right in the picture, the ones going straight along the right wall, are the interchange tracks between the two main lines. If you follow these downward you'll see they merge into one, using curved switches in the bottom right corner, and that one track then connects them to the double main line. Although these tracks appear to be right next to the main line tracks, what the diagram doesn't show is that there is a widening elevation difference as the main line gains elevation and the interchange tracks slope downwards.

      The elevation difference was because I decided early in the design process that the junction would be a bridge of one line over another, not a level crossing. This was another one of those decisions that required a lot of thought. A level crossing is certainly more interesting operationally. However, given the traffic density for these lines it seems to me that a bridge crossing would have been a requirement -- especially a century earlier when the Union Station would have hosted 40+ tracks and automatic traffic control wasn't even a concept. On the plus side, this gave me a good reason for introducing gentle grades (1% maximum) to add visual variety.

      At the same time, I wanted the scene to contain evidence of there having been even MORE traffic in past years. Although both main lines are now double-tracked, the layout will have signs that more tracks were used on the main lines in years past. One example of this is the track which goes from Union Station to the E-W line. You'll see that it connects with the main, but also that two other tracks extend along the main line towards the junction. The purpose of this track is to go from Union Station to the commuter train storage and maintenance yard (on the layout peninsula). There is a dual track for a short distance which serves as a runaround track. These dual tracks are positioned as if they are on the same roadbeds that in years past hosted the 3rd and 4th track of the E-W main line -- back when traffic density justified that many main line tracks.

      Perhaps the most interesting item in this diagram is that trains leaving Union Station on this side of the layout have 3 choices of direction -- one toward the E-W main line, and two toward either direction of the N-S main line. The radius for these tracks is tighter than the main line -- one of them actually is set at the 24" minimum radius for slow speed yards. All in all they succeed in providing the sense of an interesting, complex, network of track at the entrance to the Union Station.

      There were a number of things that needed adjustment in this diagram, and these would be taken care of in later iterations. The biggest issue was the "S-curves". The back-to-back switches are just not realistic. There needs to be at least one car length between switches to avoid the S-curve problem.

      Otherwise, referring back to the first diagram at the beginning of this post, the rest of the track shown for the lower deck would not survive future revisions. On the left side of the diagram you'll see a single track spur from the mainline. The idea here was that this spur would meander through city streets then ultimately end up on a landing along the river, going under the main line, and that the bridge over the river would be in the upper left corner. As I reworked this later I would toss out the idea of having remaining industry along the river as being both inconsistent with my vision for the city, and also a sign of trying to do too much in the space. This last point is purely subjective, but it was the conclusion I came to.

      The last point I'll make regarding the lower deck is my decision to include a commuter train yard on the peninsula. Although the track for this yard would be revised, I kept the yard in that location. The question about whether a downtown yard made sense was another one of those subjective decisions. One option would have been to argue that there would be commuter train yards at the suburban ends of each line, where real estate is cheaper and where most trains would originate. This would also allow me to put some sort of interesting industry on the lower deck peninsula instead of a commuter yard. However, I eventually decided that consolidating all maintenance operations in one central location would be a money saver, and thus a likely result of the 1970s consolidation of all commuter railroads in the city, and that the real estate in question would have been available from other railroad uses that the SAMR (St Albans Municipal Railway) would have owned. In addition, the commuter yard operations are turning out to be at least as interesting as those of a freight industry, and have the benefit of being unusual for a model railroad.

      So that's where the lower deck design was on February 27, 2006. At that time I'd also done some work on the upper deck, but it was not as far along:


      This does roughly follow the high-level design concept for the upper level. The peninsula is used for a branch line, terminating in the railroad museum. The railroad museum concept would later evolve so that it now matches what is described in the high-level design post. There is a commuter station on the right side of this diagram -- that would remain in that location, but only after extensive modification.

      There is a first attempt at a freight yard on the left wall, with part of the yard spilling over to the bottom wall. This attempt was entirely unsatisfactory and resulted in a long study of frieght yard design which I'll cover in a separate post.

      As a final note in this post, you'll see there is no evidence of staging in either of these diagrams. In fact, there were staging tracks in this version of the design, but they were in separate CAD layers so are not shown. At this point they were very rudimentary. Much of the work over the next several months would be struggling to find a suitable staging approach. This will be part of the discussion in the next design post.

        Monday, February 15, 2010

        Choosing N scale track: Flex Track and Switches

        Edit from the future (2021):  The price situation is very different from what it was when I wrote this 11 years ago.  Peco is no longer at a disadvantage.  Please check your own prices.

        Last year I posted an overview of types of N scale track. In this post I'll focus on one type of track in particular: flex track, and the related topic of switches. In theory any switch can be mated with any type of flex track, but in practice there are limitations. For many brand combinations the work to connect the two types of track is onerous and kludgy. You might, for example, connect two incompatible types of track in a few places for, say, a transition from staging to the main layout or from a main line to a branch line, but not as something you'd do for every switch. In addition, there is an appearance question, as certain lines of switches look out of place when connected to certain lines of flex track. The purpose of this post is to cover just those brands of N scale flex track and switches widely available in the U.S. Regarding the below references to track "codes", see my earlier post on that topic. Atlas Code 80. The granddaddy of N scale track, now distinguished from other track lines by its cheap price (good) and unrealistic appearance (not so good). A wide variety of switches are available, but nothing larger than a #6. There is also a wide variety of crossings and sectional track. Some people still use Atlas code 80 because they have experience with it and know it works. Others choose it for staging and hidden track, where appearance is not important. In addition to appearance the common complaint about Atlas code 80 is lack of switch reliability. This can be addressed with certain switch prep procedures, as I've covered in an earlier post, and in addition Atlas has improved the switches in the past couple years. Atlas code 80 can be easily mated with all Peco track. Atlas Code 55. Probably the most popular choice for N scalers using flex track in North America today. This is because it is price competitive with the other options, except Atlas code 80, yet the appearance is much better. There are also a wide variety of switches, crossings, and sectional track available in this line, including a #5, a #7, and a really cool looking #10 switch. Atlas also just announced two curved switches. There are still no slip switches, but given the number of recent extensions of the code 55 line that may change. Criticisms of Atlas code 55 are: 1) the plastic molded "spikes" which hold the rails in place are oversized, 2) because of (1) cars which run the Micro-Trains "pizza cutter" wheels (that is, those with a high flange profile) can't run on Atlas code 55, and 3) although the switches have a nice wire for the frog, it tends to corrode and lose contact, so you still have to add your own frog wire. (See earlier post about frogs and switches.) None of these criticisms are severe. For (1), while the spikes are oversized if you look closely, from a distance it's not noticable. For (2), most people who use Atlas code 55 are likely to use lower profile wheels which avoids the problem. And for (3), this can be taken care of easily during installation. Atlas code 55 mates reasonably well with Micro-Engineering code 55, so some people like to use M-E flex track with Atlas switches. Micro-Engineering (M-E). Micro-Engineering prides themselves on making high quality products, both in terms of appearance and reliability, and that is a good description of their code 70, 55, and 40 lines of N-scale track. They offer these with weathered or non-weathered rail, and with concrete ties for the larger two sizes. They also are the only manufacturer to offer a bridge tie version of their flex track, which is important because bridge ties are very different than normal track ties (roughly twice the tie density, and a different shape). All this variety allows a modeler to include different rail sizes for different areas of the layout, as is typical in the prototype. M-E was historically the most expensive track but that situation is no more. The UK firm Peco raised prices extensively in the 2000s, in part due to the declining dollar, and now is more than 50% higher than M-E. Based on my recent price samples M-E now seems roughly in the same price range as Atlas code 55 flex track (Atlas has also had a few price increases on track in recent years). The only complaint I've ever heard about the track itself is minor -- a few people say it is harder to bend than the other brands. The main complaint about the line of track, other than the price, is that the switch offerings are so spartan: just a #6 switch for code 55 and 70, and if you want a code 40 switch you have to make it yourself. Because of the superior appearance of track but lack of switch choice many modelers use M-E code 55 with Atlas code 55 switches, and the two rail sizes match up reasonably well. However, if you want the epitome of best appearance for your track, the choice is probably M-E plus hand-laid switches, and for that M-E also offers supplies for those who build their own switehcs. Peco code 80. For a while the Peco code 80 line was considered the best choice in track, before the other sub-80 lines were introduced. It is considered somewhat better in appearance than Atlas code 80, and their code 80 switches are considered much more reliable than Atlas code 80. Because the Atlas and Peco rails work well together it used to be common for modelers to use the cheaper Atlas code 80 flex with the better Peco code 80 switches, and some Ntrak clubs still use this as their standard. Peco also offers a wide variety of switches, including the only slip and curve switches in code 80, as well as crossings, sectional track, and even a derail. (Note: there are actually two compatible lines of Peco code 80 track -- the cheaper SEtrack and the more expensive Streamline. I'm only covering Streamline in this review, as SEtrack is very hard to get in the U.S.) Their switches include a "spring" for snapping to one side or the other, which can be removed to install a slow motion switch machine. In code 80 they also offer "electrofrog" and "insulfrog" for all switches. Complaints about Peco code 80 switches are primarily from the American market because the UK N scale standards vary every so slightly from US standards, so that depending on the equipment you run you may have trouble with derailments or, if you use DCC, with occasional shorts. These seem pretty rare, but you do find them mentioned on comment boards, and there are standard suggestions for fixing this problem if you have it. The other issue with Peco, as noted earlier, is that the prices have skyrocketed in recent years and as a result they have nearly priced themselve out of the U.S. market. In HO, Peco has released American-style switches that are considered among the best you can buy, yet they have managed to keep the prices within reason. There have been rumors of a similar line for N scale, but nothing has yet been announced. Peco code 55. This line of track is highly unique, and so it has a longer write-up than any of the other lines. When Peco introduced this line in circa 1990 it was widely considered a huge advancement and the preferred choice for N scalers. Today, most modelers now see it as having been surpassed by M-E and Atlas, but there are still some unique advantages of Peco code 55, so it is not yet completely obsolete. When Peco code 55 was introduced there were a lot of commonly-held concerns about the viability of a small N scale rail. Peco addressed this by using code 80 rail but burying the bottom .25" into the ties. The track itself is very sturdy -- more than any other of the flex track offerings -- and the plastic tie "spikes" or "clips" are prototypically small, as they are used only for appearance, not to hold the track in place. There is a fake rail bottom at the top of the ties, then another one at the actual bottom of the rail. Peco code 55 track comes in wood or concrete, but just wood for the switches (AFAIK no one has yet come out with an N scale concrete tie switch yet). Peco 55 has almost the same large variety of switches and crossings as code 80, but all are "electrofrog". There is also a double-crossover with four built-in switches. Like code 80 the code 55 switches have springs which can be removed for slow-motion switch machines. Proponents like Peco code 55 for solidity, reliability, and wide variety of switch choices (this has become less of an advantage over time as Atlas has expanded their code 55 line). Detractors, and there are many, point to the high cost and appearance issues. Appearance issues start with the tie spacing, which matches a European prototype, not the U.S. The wood ties have a fake wood grain that is much too prominent. The switch mechanisms look nothing like prototype switches. The switches themselves are classified "small", "medium" and "large" instead of the more traditional #4, #6 and #8. This is because they do not follow the common U.S. practice of having the diverging rails go straight through the frogs, but instead are curved for the entire length of the switch. On the plus side, this difference means that the radius used for Peco code 55 switches is much wider than that use for competitive brands, and that the trains motion in going through the switch is more fluid. The "large" switch has a radius of 36", which is very broad (see earlier discussion of curves). But it is yet another deviation from the common American prototype. Finally, there are questions about code 55 switch reliability (there are concerns that the flangeway is too wide, leading to derailments), although there are many users who claim they work perfectly all the time. Given the cost issues Peco code 55 is usually not considered for new layouts any more. But thre are still a few willing to work through the appearance issues who find the variety of switches a compelling reason to choose this option.

          Monday, Monday

          Did manage to clear my work desk of half-finished projects, as I'd hoped. The last one was finishing up the BNSF decoder install. The problem was simply that one of the motor wires was folded up and subsequently too short to reach the decoder.

          Alas, once I got the BNSF dash 8 running it was apparent that the motor squeaked and needs cleaning. Back on the problem log -- but at least a little closer to working.

          As a side note, while my son Daniel was doing some math I picked up an old project to convert an Athearn F59 PHI locomotive (in full New Mexico Rail Runner dress) to DCC. I wanted to test the decoder with the NCE tester I mentioned in the last post. Well, decoder is fine, problem is with the Athearn board, so I sent an email off to their help line. We'll see what happens. I'd like to get a small HO layout set up using the Bachmann roadbed track just so Daniel can play with it a few more times before he goes to college.

            Sunday, February 14, 2010

            Best Laid Plans ....

            I have not failed. I've just found 10,000 ways that won't work. -- Thomas Edison

            Of course Edison was describing his attempts to invent wonderful things. I'm only trying to find a way to make my track look realistic. But I seem well on my way to finding 10,000 ways of not making my track look realistic.

            I suppose I could give up ... but I expect to live with this layout for decades so I'd really like to get this part right.

            Part of the problem has been lack of correct materials. I drove around Saturday trying to find Matte Medium ... it's so prevalent in Model Railroad books that I thought I could find it somewhere local at one of the various "hobby" stores, but nada. There are a few other suggested ingredients, but that one was key. Oh well -- I have to fly to visit a customer Tuesday and I'll drop by Caboose Hobbies along the way. Then make another attempt at track scenery later this week.

            For what it's worth I did make some progress. I've figured out the measurements, including slope grades, for the track sides and found a really good mix of ballast. And I have a test sample track-with-roadbed-and-subroadbed just waiting for experimentation when I get back from my business trip.

            On the plus side I also finally -- and I do mean finally -- got the darn Kato RDC to work with the Digitrax DN143K2 decoder. I bought this for my son for some event -- one I can no longer remember -- in 2005 and first tried to add the DN122K2 decoder (DN143K2 predecessor) back in early 2006. It was a miserable failure. From time to time I've tried again, sometimes destroying the decoder in the process. This is a really nasty installation because the internal electrical connections are so finicky, but yes, today I called my son over, handed him the throttle, and said "see if this works". It worked great, with all the lighting functions, and he spent the next two hours operating trains. Chalk that up to one more experience point with decoders.

            Speaking of decoders, I've also been fighting with a DN163A0 for the BNSF Atlas dash 8 -- the first N scale freight locomotive that I bought on that fateful day back in 2002. The lights come on bright and perfect, in the appropriate direction, but the motor just sits there. As part of the debugging process I finally opened up the NCE decoder tesk kit I bought from Mike last year, as an upgrade from the free decoder test kit Digitrax provides with their starter kits. It was a useful exercise. At this point the problem is isolated to the connection between the motor and the decoder. The motor is fine, the decoder is fine. I want to get this off of my work table so would like to resolve it tomorrow.

            One other bit of fun -- the Amtrak Superliner cars frequently derail going over one Peco switch, but none of the other Peco or Atlas switches. Occasionally, but not very often, other cars derail on this switch. Naturally this switch is in an inconvenient spot, right next to the box and far from any of the layout sides, so I have to crawl in-and-out frequently during the diagnostic process. The NMRA N scale gauge has been helpful, but not conclusive. I've found that the Superliner cars have wheels that are slightly narrower than the standard, which is why they derail more often than most. As to why the switch causes this I still haven't found the problem. At first I thought it was an overly-wide flangeway between the frog and the guard rail, but I'm not sure that is any wider than other Peco switches. Now I'm focusing on the points, but still nothing conclusive. Another learning experience that will hopefully help later on.

            My goal for tomorrow is to clean off the work desk of projects and be fully ready for restarting on the track ballast problem when I get back from the business trip later this week.

              Saturday, February 13, 2010

              Egad, what a CAD

              This post is about Computer-Aided Design (CAD) tools for model layout planning. The title actually comes from an obscure, one-act melodrama that is occasionally put on by high schools, more commonly back in, oh, say, the 1970s. Don't ask me how I know that.

              I managed to avoid CAD for layout design for a long time. I downloaded a few demo products back in 2003 but after realizing they all had steep learning curves I stuck to graph paper, pencil, erasers, a ruler and a compass.

              However, after I completed the high-level layout design and made my first attempt at a detailed staging design I realized that I was going to need CAD to complete this layout. I'd been working at it for about 2 hours when I realized that the approach I was taking simply could not work, and that I'd need to shift the entire track arrangement around. Then it occurred to me that this was likely going to happen many, many times given the nature of what I was trying to fit into the space. A tool that would allow me to shift whole track groups quickly from place-to-place, trying to find the best arrangement, was going to be essential if I was to complete the design in the same decade.

              There are several CAD tools available for model railroaders. I'm sure I don't have a comprehensive list but CADrail and 3rd PlanIt seem to be the most popular choices. There is also a free tool from Atlas which apparently is acceptable if you are using their sectional track, and I've heard there is a model railroad version of Xtra CAD somewhere, but it didn't turn up in my quick Google search.

              Which is the best? Heck if I know. Like almost everyone else I just picked one, learned it, and since it works I haven't bothered to try another since the learning curve is steep for each tool. I've read a few reviews of the various tools in Model Railroader, but they are careful to avoid making any one tool look better than the others.

              So, I picked 3rd PlanIt only because when I tried the demo I was able to locate quickly the templates for the N scale switches I was going to use. Not the most scientific approach, but at least I knew it was going to work for me. I've found it a good tool, but if I had to recommend one now I'd suggest you look at CADrail only because I can't get the 3rd PlanIt people to respond to my emails.

              The learning curve is steep even if you are adept at computer use and even have CAD experience, say from home design software. Actually I can't imagine how someone who has limited computer experience could possibly learn this tool on his/her own. Yes, part of the learning has to do with getting used to the concept of working with objects segregated by layers -- only some of which may be visible at a given time -- and part of it has to do with getting used to thinking in three dimensions. But most of the learning has to do with getting used to setting up the track and the connections between the track, including topics like easements, parallel track, and connecting track that doesn't align to the nanometer.

              In my case I spent maybe 3 weeks getting the lower deck designed, then did the upper deck over the next 2 weeks, then re-did the lower deck now that I really understood the tool.

              And once you are skilled at using the tool you'll find, if you compare notes with another skilled user, that your approach and his/hers are completely different, yet both are effective. This is because the tool has so many options that it's likely no two people use it in exactly the same way.

              So, thus far all I've said is that I used 3rd PlanIt to design my layout, I am skilled with the tool, and it took me a long time to learn how to use it. But the real questions are: do I consider it worth the effort, and if so what are the advantages? The answer to the first question is an emphatic yes -- I wished I learned it years ago and saved the time I spent doing smaller layout designs on graph paper. The answer to the second question is that these are, IMHO, the advantages of using a CAD tool:

              1. Speed. Once you are adept you can put down the initial version of a track section at least as fast as you can using graph paper. After that is done subsequent adjustments can be done in mere seconds or minutes. You can save literally hours or days with a single redesign effort.
              2. Precision of design. On paper, no matter how hard you try to be exact, you can too easily "fudge" the angle of a switch or the radius of a curve, and deceive yourself about the workability of a track arrangement. With CAD what you design is what you get. "Fudging" is only possible if you cheat by, say, forcing two tracks together that don't really fit. If you use the standard connect tools that won't happen.
              3. Precision of implementation. CAD allows you to know the exact location, down to 1/32nd of an inch if you like, for a given switch or start of a curve.
              4. 3D and overlays. It is a challenge to figure out how elements fit together in a three dimensional space using graph paper. With CAD you can easily select which layers to view, allowing you to see how different elements fit together without distractions from other elements. Futhermore, most CAD programs have 3D views so you can confirm it all works.
              5. Precise calculations. Will that siding be long enough? What is the exact slope of this grade? What is the exact radius of this curve? All these and more are quickly answered with CAD, but often can be calculated only approximately, and after a lot of work, with graph paper.
              6. Easy printing and sharing of designs. A nice side effect of CAD.
              In addition, there are other features of CAD which I haven't used, such as testing out scenic views or actual train operations, which many people find helpful.

              The only caution I'd give to someone using CAD is that it really helps if you have experience building from CAD before you decide that your CAD design is "final". This is because CAD's extremely precise calculations can mislead you about how the design will translate to reality. My plans, for example, are calculated to the 16th of an inch. When you start working in your layout room you'll find that real life isn't always like that. Floors aren't always level and walls aren't always plumb (in fact, they usually aren't). The peninsulas you build may not be exactly perpendicular from the wall. You may also find that if you carefully plot two locations of a line from the CAD to your benchwork, that when you extend the line outwards it doesn't seem to correctly line up. In part this is because 1/16th of an inch is really small -- your pen mark may be that wide! You probably will need to check measurements again and again at different locations, and also be prepared to adapt your CAD design slightly based on what you encounter on the benchwork. Once you've done this once or twice you'll take it into account when doing your next CAD design, and leave yourself a little slack room in the design.

              So, I used CAD for the iNdoor layout design, and then later for the Garden railroad as well. In future posts I'll describe those designs.

                Goals for weekend and beyond

                My main goal for this weekend is to give scenic treatment (i.e. track paint and ballast) for one short section of track on the main layout. This will be a fairly involved process, but over the past few days I've made progress on testing various methods and researching what the prototype looks like, and I think I'm ready to go.

                During the past few days I added a couple design posts, worked on the track scenery techniques, started preparations for the DS64 tests, and repaired a small number of locomotives with help from my son. In the process I felt like I was getting too scattered, and needed a clearer focus for the work.

                I decided the key thing I want to emphasize is the "fun" factor for myself and my helpers in the family. We want to run trains, yes, but right now we also want more variety in terms of the trains we run and the places they go. This means adding more track and making the portions of the main layout that are currently off-limits usable again. Ideally the first part of this effort will be connecting the middle tier staging to the main layout because we need that space to store more trains on the layout.

                Now, I could just proceed in this manner without thinking about track paint or ballast. Indeed, most modelers recommend deferring the ballasting of track for near the end of the process, after the major scenery projects are over, due to the mess and the tendency to "spill" things onto the ballast, forcing it to be redone. In addition, some modelers recommend not ballasting until all track is complete and you've run a few operating sessions, because there is a good chance you will want to change some part of the track arrangement after you experience how it operates.

                However, the truth is while I have theories about how to scenic the track I don't really *know* because I haven't done it for this layout. I want to have completed the track scenery for at least one section that includes switches before proceeding with new track, because there is a good chance that I'll adapt my track-laying techniques as a result.

                Therefore the goal for this weekend is to scenic a short section of double mainline track without switches, then next week I'll try an adjacent section with a crossover.

                Once that is done I may decide to ignore the ballasting until all track is in place. Or I may decide that it looks so cool that I will want to complete it, at least for areas where I know the track arrangement won't change. But either way once I've completed this I can proceed to more track-laying with confidence that I know how the scenic details will be added later.

                  Thursday, February 11, 2010

                  Wiring Standards part 6: Power cabinet and Loconet

                  I created a number of posts last year about the wiring standards for the layout. This post extends on those by describing standards that have been developed and/or refined as a result of the power cabinet work last weekend.

                  Here is how the cabinet looks most of the time:


                  The goal was to have a clean appearance with everything easily accessible for operations and maintenance. Regular operations can be performed by opening the cabinet to start the layout and get the throttles -- at which point the doors can be closed again until operations are over. For maintenance the cabinet can be pulled away from the layout, since it is on rollers, to access the wiring in back of the cabinet or underneath the layout.

                  Here is the view with the doors open (in this view the cabinet is pulled away from the layout):



                  The top shelf is currently for storage of throttles and of Digitrax manuals. I've earmarked key pages that get frequent use, like CV tables, how to program consists, Ops mode programming steps, etc. As the number of throttles grows I'll likely need to have a separate place to store them.

                  The second shelf has the DCS100 command station on the left and the power strip on the right. The power strip is screwed lightly into place and all the power supplies needed for the layout are plugged in place. If you click on the photo for the larger view you'll see they are also all labeled with their purpose. In this setup, under normal operations the starting and stopping of the layout is done with just the main switch on this power strip.

                  The DSC100 has various wires attached to it but I've tried to keep those organized too. The power and ground outputs, red, black, and green, go directly from the DCS100 to the barrier strip on the left wall. Any device that needs connection to the DCS100 is connected to the barrier strip, not to the DSC100 directly. All power wire is 14 gauge, and as with the rest of the layout pairs are twisted together and labeled.

                  The DCS100 also has a programming track set up in front of it. This allows for quick decoder programming during an operating session if for some reason Ops programming won't work. It can also be used at other times, but for normal locomotive work I'll probably use the Digitrax Zephyr at my work desk.

                  The empty space on the second shelf is reserved for a booster, which may be needed when I start building the upper deck of the layout, or for a second power strip.

                  The third shelf is for power management, as can be seen from the labels in the larger picture. On the left is the PSX4, which provides power to districts 1, 2, 3 and the auto reverse (AR) sections that are controlled by the PM42 (the middle device). These are lightly held in place and have padding protecting them from the shelf itself. The empty space on the right is reserved for the second PSX4 for the upper deck districts. As I plan to have 8 power districts plus a district for the PM42 I will also need a PSX1. I plan to place that in front of the PM42 and wire it to the PM42.

                  During operations if we get a circuit break or if we otherwise have to view the power management LEDs for problem diagnosis we simply open the cabinet to see them.

                  The bottom shelf is currently occupied only by the main power supply for the DCS100. (It's a DCC Specialties device, recommended by Mike Gleaton.) The rest of the space is reserved for an additional power supply for a booster, if it turns out that is needed, and for transformers for the lights and switches.

                  Here is a look behind the cabinet:


                  This does look a bit messy with all that wire, however if you look closely you'll see I've tried to tame the mess a bit by setting up hooks around the back of the cabinet for routing the wires. Plus, the wires are all organized into labeled, twisted pairs. However, because the cabinet must be allowed to roll away from under the layout there has to be some slack in the wires that connect the cabinet to the layout. All these wires can be disconnected, if need be, by plug or by undoing the electrical pigtails, and all are labeled to allow easy reconnection. However, the slack wires do make it look messy. The wires for power districts 3 and 4 and for the upper deck AR section have already been deployed, so that attaching these later will be a simple matter of connecting them to the wire ends at the back of the cabinet, instead of having to add wires inside the cabinet.

                  One other minor wiring standard emerged from this work, and that had to do with the Digitrax UP5, UR91 and UR92 front panels. These are distributed throughout the layout on front fascias to allow plugging in of Digitrax throttles anywhere. UR92 is the two way radio receiver panel, the two UR91s are one-way radio receiver panels, and the UP5s are basic panels.

                  These panels have two wired connections. One is Loconet, a 6-pin cable. The other is a single 18 gauge power wire. The single wire, together with the loconet ground wire, allows up to 10 panels to share a single power supply. The UR92 doesn't have the power wire because it requires a dedicated power supply. The power supply for the 10 other panels is connected to the UP5 nearest the power cabinet.

                  One standard I'm following is to configure the loconet in a branching configuration, avoiding loops, per the Digitrax recommendation. The other is to label the power wire "UP5" every few feet to distinguish it from other wires. The Loconet cable is not labeled, as it is clearly unique.

                  The last point is placement of the panels. Usage-wise, radio throttles can connect to any panel, not just the radio ones. The one-way throttles use any panel to plug-in to select or dispatch locomotives, but once that is done the throttle is disconnected and communication is automatically done via a one-way receiver. For this reason the two UR91s panel/receivers are placed in the two different rooms, and roughly in the middle of the rooms. The two-way throttles don't need to plug in at all unless they get confused, however they do need to be initialized when they start up by plugging into the UR92 for two seconds. For this reason the UR92 is placed on the right side of the cabinet, so that you can plug in at the same place that the throttle is stored. That also is roughly the center of the layout, which should minimize any chance of radio interference. You can see the UR92 panel in the upper right of the first photo with this post.

                    High level iNdoor layout design

                    There are as many different methods of layout design as there are model railroaders, but one thing most methods have in common is to start big and work down to the details. John Armstrong had a method for the first-pass, high-level design that he called "doodling by the squares". You can find this described in his Track Planning for Realistic Operation book and also as the first article in the Kalmbach book 48 Top Notch Track Plans (IMHO one of the best track plan compilation books out there, even though most of the articles are from the 1980s.)

                    While I didn't use the exact method Armstrong proposed, I did follow his general approach. In this post I'll cover the development of my iNdoor layout design from the start of design to the point where I transitioned to detailed design and decided to move to CAD (computer-aided design).

                    Rarely, if ever, do you start a layout design with a completely blank slate. Usually you have something in mind already, and ideally you already have a thorough list of the "givens and druthers" (to use Armstrong's term) of layout *musts* and layout preferences. For me I had my layout concept, the parameters of the layout room itself, a basic understanding of staging requirements (at first not quite as detailed as what is in the linked post, but pretty close), and a list of design standards:
                    1. Minimum radii: 36" on the mainline, 24" for slow speed yards and station approaches, 18" for staging

                    2. Aisle width minimum: of 3' (later expanded to 3'10")

                    3. Track spacing: as close to 1" (13' 4" scale) as possible given spacing needed around curves

                    4. Maximum freight train length: 12', with 16' for a few through trains

                    5. Maximum passenger train length: 10' for distance trains, 5' (8 cars plus engine) for commuters

                    6. Maximum grade: 1.5% (later reduced to 1%)

                    7. Maximum reach: 2' from fascia to rear train track

                    8. Maximum shelf width: 2' 2"

                    9. Minimum switch size: #8 main layout, except a few #6 in industrial areas; #6 in staging
                    In addition to the quantifiable standards listed above, there were certain other design goals:
                    1. No duckunders for normal operations and viewing (ok for occasional maintenance). If a duckunder situation appeared inevitable, use a liftout bridge or gate, but ideally not even that.

                    2. Movability. I have no desire to move house ever again, but just in case we do move I wanted to build the layout in a manner that would allow it to be broken down to 4' sections for moving.

                    3. Double deck with helix. If you have the room a single deck is nicer because the layout height can be optimal for your situation and the overall illusion of the model works better. But, the benefits in terms of railroad length made, for me, a double deck a given. For layouts featuring long, steep grades you can have multiple decks without a helix, but that isn't an option for this layout.
                    So, armed with my design standards and preferences I went to work. I started by drawing the room boundaries and features on graph paper with bold lines, then made many photocopies. This way I didn't have to redraw the room with each iteration.

                    The first question was where to locate all the things that I wanted to include in the garage apartment: the layout, my home office, a gathering area (sometimes called an operator's lounge), and modeling workspace. I decided that the north room would be for trains, the smaller south room for everything else. The early goal was to avoid if possible extending the layout into the south room, but that wasn't a requirement -- I kept that as an option if needed (and of course that is where the main staging resides now).

                    The next question was what to do with the built-in desk in the north room. My first drawings tried to fit a layout in and around the desk, but very quickly I decided that the desk had to go. I would move it to the opposite corner in the south room.

                    Then came the main question: where to locate the layout benchwork itself. A common approach is to start with an around the walls design and then figure out where to put in peninsulas to use the middle space. I tried that with many designs, comparing the options, and also tried a number of designs which left one or two walls open. I tried variations with diagonals, curved benchwork, and a variety of other shapes. This wasn't so much that I didn't have a reasonable design in mind, I just wanted to make sure I'd explored all the options.

                    For this layout the biggest design challenge is the four-way, "X" junction of the two main lines. By this I mean a junction like one shown in this sketch:


                    Junctions in general are a complex topic in layout design, with three-way "T" junctions being simpler than "X" junctions and junctions to a branch of the same railroad often being simpler than junctions between railroads, as the latter usually needs some provision for interchange.

                    Junctions are fun to include in a layout because of the complex trackwork and the interesting operations that result from interchange possibilities (an interchange yard is often referred to as a "universal industry" because you can justify any kind of freight car via an interchange). In addition, if the junction is level you have additional operational interest in directing the traffic across the crossing. For my layout I wanted the junction mainly from the point of view of Union Station. To me one of the coolest things about large stations is the way that there are many trains departing at nearly the same time from different platforms, taking different track routes to different destinations. (As opposed to a station like in San Francisco, where trains leave roughly hourly in one direction -- toward San Jose. Much less interesting operationally.) In theory you don't have to have a junction to model multi-destination operation because you could have the trains departing from the station, go into staging, and you can pretend that there is a junction somewhere on the other side of staging. However, to me that wouldn't be really satisfying -- I wanted to include at least one major junction on the layout.

                    As part of the design process I skimmed all the model railroad magazines and books in my library looking at how others have modeled junctions. One common way is to have a "passive" junction. That is, the track for the crossing railroad is included on the model but it's not used and it doesn't go anywhere. Often the "passive" junction will include an active interchange yard that is fed from a staging yard.

                    Once you move into "active" crossings -- especially if you want all 4 legs of an "X" crossing to be active -- you run into a challenge trying to avoid creating a duckunder. To understand this, consider an "X" junction with tracks coming from 4 directions, which I'll call N, E, S, and W (see above sketch for an example). Suppose you are operating a model train from the N direction to the S direction, and walking along side it as is normal practice. There is no way you can accompany your train from the N side of the junction to the S side of the junction without somehow navigating around the track on either the E side or the W side.

                    This topic was addressed in two articles in Model Railroad Planning 2000. Paul Dolkos covered the topic in detail in one article and suggested that if the crossing angles were sufficiently acute you could avoid a duckunder by keeping two legs of the crossing (say, N and E) together on one side of the shelf and the other two on the other. This approach was used in the Hutchinson, KS layout I linked to earlier, and a sketch of this approach is shown here:


                    You can see that the junction can be fit into a single 2 foot wide shelf using this method. I seriously considered using this method but ultimately rejected it. One problem is that I wanted passenger trains to be able to go in all 4 directions, which would not be possible given an acute angle and my minimum radius. But the real problem is that the acute angle would not look right. Granted, there have to be things on your layout which don't look exactly like they would in real life, such as curving a freight yard around a room corner, but to the degree possible I wanted to minimize those visual "giveaways" which spoil the illusion.

                    So, I tried different designs using a 90˚ crossing (or at least reasonably close to 90˚), including putting the crossing in the middle of the room and at various diagonals. Eventually I came to the same conclusion a lot of other modelers did, based on the layouts that I saw in the books, which is that the optimal way to model a real-looking, active, 4-way junction is to put it along a wall. Here is an example of this in sketch form:


                    In this method two adjacent legs of the junction (let's call them N and E for this example) are positioned along the wall facing each other. As they near each other they curve toward the (approximately) 90˚ intersection. The other two legs, S and W, then go out towards the middle of the room.

                    The reason this can work is that the S and W legs won't go very far before they transition into staging. Practically speaking, unless you have a warehouse-sized space for your model layout you're not going to be able to model a long stretch of all 4 legs of the junction, so we pick the two which go into the center of the room as the legs to keep short. Going with this approach I came up with this concept:


                    To be honest I wasn't initially 100% thrilled with this approach either, but it seemed to be the best compromise. In the first draft of this concept (not shown) I located the peninsula along the bottom wall in the drawing and was troubled by the way the N and E legs left the crossing then both curve around about about 135˚ to get to the next wall. The 36" minimum radius (and given the space, probably a wider radius for both curves) helps somewhat, but 135˚ curves are still iffy. I addressed that in the above sketch by locating the crossing on the west (right) wall. This way one of the legs had to manage only a gentler 45˚ curve to then head towards the end of the layout, and staging or a helix. For the other leg, by moving the junction as far as possible toward the south end of the West wall (the top end of the drawing) I could break the 135˚ curve into two lesser curves and improve the appearance.

                    At this point there were still a few big design issues to take care of. 1) Locating the helix. 2) Width of the peninsula, and what to do with the remaining interior space. 3) Where to locate the branch line. 4) Desire to get maximum use of the layout space, and thus to extend the double deck over the peninsula. 5) Where to locate staging.

                    Locating the helix (item 1) was just too obvious -- the west alcove. I had briefly considered that as a potential dispatchers office, but a) a dispatcher would be much more comfortable in my home office in the south room rather than in this cramped space and b) this was a perfect spot for an oval helix between the two levels.

                    In terms of item (2) I tried minimizing the width of the peninsula in an effort to somehow cram in yet another peninsula, then realized that a simpler approach would be much cleaner. I'd use a wide peninsula, one that took up the available space in the middle up to the limits of the minimum aisle width. This would result in a very wide space, almost 9', for the peninsula, which is too wide normally. However, I'd create a walled-in "box" in the interior which would also serve as the support for a second deck on the peninsula (addressing item 3) and the upper deck peninsula would then be the obvious place for the branch line (item 4). Voila! I had figured out a workable high level design. The sketch now looked like this for the lower level:


                    Note that at this point I am not identifying any of the layout design elements (LDEs -- using Tony Koester's term from Model Railroad Planning) except Union Station on the sketch -- those will be located during the detailed design phase. Also note that there are two main lines identified, the N-S (north-south) which runs for most of the lower level, and the E-W (east-west), which is part of the lower level, but via the helix moves to the upper level where it is the only main line:


                    The last question was what to do about the one remaining issue, the location of staging?. My initial thought was that the tracks which terminated on the lower peninsula would have staging under the peninsula itself. The two tracks which terminated on the lower south wall would have hidden U-turns and then have staging underneath the lower deck. My plans for the upper deck staging were vague -- maybe thinking of trying to squeeze staging in immediately under the upper deck. The helix area was an obvious spot for a staging entrance, but I wanted to avoid any sort of bridge from the south wall tracks to the helix.

                    Realizing that a workable staging solution was essential that would be the first topic I tackled when I started the detailed design.

                      Wednesday, February 10, 2010

                      In the (Wednesday) Evening

                      With trains running there is some definite momentum. Fixed three more items from the problem log today and made progress on the never-ending concrete-tie project.

                      The first item was to fix Loconet cables. Without going into details, Digitrax Loconet cables are 6-wire phone cables that users typically make themselves. When I was doing this back in 2007 I didn't realize that it was important that the wire orientation, left-to-right, be the same on both ends of the plugs. If you ignore this requirement Loconet itself still works, but later when you try to use RailSync it will fail. So, today I surveyed the Loconet cables on the layout and corrected the plug ends where needed.

                      The second item was to move one of the UR91s (the Digitrax one-way radio transmitter, of which I have two) from the top of "the Box" to a spot along the fascia on the east side of the room. This required setting up the Loconet cable. In addition, I have 8 vanilla Digitrax UP5 panels at various points in the layout and two UR91s, all of which are powered by a single Digitrax PS12 power supply (up to 10 such devices are supported on one power supply). This requires a single 18 gauge (or so) wire connecting all the devices to provide a "return" path for the current that comes in via Loconet. This return wire is labeled UP5 -- and I made sure that the labeling is now extended to the entire layout.

                      The third item was the Amtrak P42 locomotive problem mentioned yesterday -- where it "clicked" from time to time during normal running. Turned out this was a piece of debris that got lodged up high in the gears above one truck. The bad news is that in the process of debugging this I broke a piece of plastic, so now the loco runs great but if you pick it up from the track the truck is apt to fall off. So, one problem fixed, but a new item to order and install a replacement part for the same locomotive.

                      For the concrete ties I found that I still don't have the colors right, even after adding dullcoat. Sigh. I also tried a few more methods to "sand" down the fake wood grain without good result. I think I need a visit to an art supply store to seek out more options here.

                      Grainy Concrete

                      More progress. One of the goals of the week was to address three of the issues on the problem log. Already two are done.

                      The new DT402D (two-way radio) throttle was having an issue where at times it would seem to lose communication with the base, so we'd have a runaway train. I read the manual and made two changes. First, there is a throttle setting for maximum battery conservation -- that comes set by default and I turned it off. Second, the manual said that if you hit the Loco key once you dispatch (lose control of) the locomotive. Odd -- I'm not sure if that was the same behavior on the DT400R throttle -- but on that one it wouldn't have mattered while running wireless because changing locomotives required plugging in. So I now know to avoid that. Not sure if the problem is solved, but I now have it in "monitoring" state.

                      The other issue was to resurrect the Kato Amtrak P42 Genesis locomotives that I bought in 2002. Back in late 2002 or early 2003 I tried a method of track cleaning that someone recommended -- using metal polish. VERY BAD IDEA. It left a partially-insulated coating on the track and on the wheels of the engines that were unfortunate enough to be used for this experiment. I then compounded the mistake by installing decoders in the P42s in 2006 without first verifying that they were working fine with DCC. Now I had two non-working locomotives and didn't know whether the problem was DCC related. I set them in my project box where they stayed until last night.

                      I brought out the old Digitrax Zephyr, the Digitrax low-end starter set, which I bought in 2005 and used until I upgraded to the Radio Super Chief (now replaced by the duplex version) in 2007. I kept the Zephyr because it could be used for my son's HO set and also used for testing locomotives at the work desk, which is what I used it for last night. When I put the P42s on the programming track the decoder could barely be read for one and not at all for the other. Then I tried putting them upside down in one of the N-scale foam cradles I have (a must for any N-scaler) and tested again, this time with electric alligator clips connected directly to the wheels. Much better connectivity -- both decoders were easily read and I could actually get the motors to spin in track mode. So, using this, I took a wheel-cleaning brush (also a must for any N-scaler) and used that while the moter was running, then followed with rubbing alcohol and a q-tip to clean off any left over gunk. Afterwards both ran beautifully on the track, although one of them developed a "click" after two laps that I'll try to look into today. It was fun -- for the first time in about 7 years we ran the Amtrak superliner, and the first time ever with the MHCs that I bought in the interim.

                      This method definitely goes into the engine cleaning arsenal.

                      Otherwise, I also worked on the main task for the week, making the Peco track look like it has modern concrete ties. I put 4 layers of paint on the ties hoping to cover up the fake wood grain and found that even then some of the wood grain shows through. Rats. (Hence the title of this post.) Probably will need to somehow sand off the wood grain. Will be looking at that today.

                      Otherwise it looks pretty good -- I need to make the color a little more grey, but before I do that I want to add rust to the rail sides and the rail clips and spray some dullcoat on it today and see how it looks after some weathering.

                      For the rest of the week I will continue to focus primarily on the concrete ties, but in between will continue to address the issues on the problem log.

                      Tuesday, February 9, 2010

                      Tuesday Afternoon

                      Another post title based on a song title. But the real purpose of this post is a progress update.

                      Item #1 from this week's task list is done. I have to remember to run the CMX once per week for track cleaning, and over time I'll learn how to optimize this exercise. Once optimized I'll write a post on it.

                      For item #4, I've ordered the track, per my post from last night. That's all I wanted to accomplish for this task this week.

                      I've also made progress on item #2 and on one of the items on the problem log. I'll write more on those later this week.

                      In addition, I added one item to the problem log. The Kato SD90/43MAC (CEFX 118) which has the new decoder runs very fast at speed setting 001. I suspect that I may have accidentally modified the speed table CV when I was resolving the address problem.

                      Monday, February 8, 2010

                      Comparative shopping

                      As a follow-on to my post last year about N-scale shopping, here are the results of a price search I did tonight for the remaining track I'll need for the lower deck staging. I needed to order 23 sections of code 80 flex track and 4 #6 code 80 remote switches (2 left, 2 right). While I find pricing on items like locomotives to be very close between the best sites, for some reason there is a lot of variation on track.

                      Wig-wag: $2.83/piece; 15.98/switch
                      BLWNscale: $2.95/piece + $7.95 flex track handling charge; $13.35/switch
                      Wholesaletrains: $2.41/piece; $14.22/switch

                      I'd forgotten to list wholesaletrains on my post last year, but in digging through past emails I found that I'd ordered most of my Atlas flex track from them, and switches from BLWNscale.

                      Wig-wag is usually super competitive and are consistent on discounts, so I'm guessing that they don't get quite as good a pricing on track from Atlas as other sites that also sell a lot of HO product. BLWNscale apparently is not interested in the hassle of shipping flex track, but its a mystery how they can offer switches at such a low price. I'm going with wholesaletrains for the switches this time because I'm ordering so few, and a separate order from blwnscale would cost more in additional shipping charges than I'd save in product.

                      And when the Trains come marching in ...

                      The title is a punny reference to the latest Super Bowl winners. It was a fun game. I was making progress on the layout at the start but by mid-3rd quarter I just focused on the game.

                      Regarding the weekly task list from last Thursday we got most of the targeted stuff done, plus a few things not on the list. #1, the switch machine issue, is resolved temporarily in that the crossover switches do work automatically with only one Tortoise and the derailments have stopped. The solution didn't take long -- it was the same concept I was trying in 2007 but this time with better materials. However, it's now only a proof of concept -- when I start the project to automate all the switches I'll re-do this solution in a more solid manner.

                      #2 is done, as I noted on Saturday, except that I need to finished writing up my post on the power cabinet. #3 is done except that I need to start the weekly routine of using the CMX car. The only major goal I didn't complete was that I didn't get started planning #4.

                      In addition to the planned tasks 3 other things were done. I resolved #5, as I reported earlier. My son successfully did his first decoder install on his own, a DN163K2 "plug and play" Digitrax decoder into a Kato SD90/43MAC. Worked on the first try. The only minor glitch here was my fault -- the locomotive number on the shell is 118, but when I assigned that as the DCC address it wouldn't run. Eventually I worked out that 118 is considered a 2-digit address in the Digitrax system (the cutover to 4-digit addresses starts at 128), and knowing that we were able to get it running immediately.

                      The other bit of work this weekend came as a result of a problem we ran into on Sunday. One really nice thing this weekend was that Daniel and Emma joined in a mini-operating session with me which lasted about 2 hours, only ending when I said it was bed time. Emma did well figuring out the DT400R throttle and is getting pretty good at making sure the switches are set correctly. On Sunday she asked to start again, but after a while a short occurred.

                      As I investigated the short I found that the PSX showed both Power District (PD) 1 and 2 as shorting, which seemed wrong. We stopped playing, but when I got back to the layout later I started a mini-project to investigate whole layout wiring.

                      As it turned out the short was caused by an engine going over a switch that was set incorrectly, something I didn't think about because we don't use that switch right now. But, it was a useful exercise. I found that PD1 and PD2 were not fully isolated, because when I cut the gap in the Peco code 55 rail I didn't cut deep enough into the ties. I also learned, to my surprise, that when the layout has no power one of the rails (the "left" or "black wire" rail) is still connected electrically across the PDs and the ARs, but the other rail is isolated. This must be the standard way circuit-breaker devices like the PM42 and PSX operate, but I didn't expect it and as a result I was confused for a while about the problem symptoms.

                      One other point I've been aware of for a while is that in PD-1 the rails are not completed isolated, but instead register something like 100 ohms of resistance between them. Not enough to prevent operation, but it bugs me. I'm adding this to my problem log to work on again.

                      Another problem added to the log is to investigate why the D402D (new two-way radio throttle) loses connectivity from time to time. My first step will be to actually read the manual.

                      Finally, as a result of this investigation I also ended up fixing some of the track around the electrical gaps.

                      So what tasks to work on this week? Here is the new list:


                      1. Track cleaning.

                        Get the CMX car running over all track on a weekly basis.

                      2. Track scenery.

                        This is the item noted in last week's list. This is a potentially big project, with goal of being able to make the Peco code 55 look like realistic, modern, concrete-tied track.

                      3. DS64 deployment.

                        Also noted last week. This project is to figure out how to get the stationary decoders working with the switch machines.

                      4. Extending middle tier staging.

                        First step is size and order the needed Atlas code 80 track. Once it arrives the project will be to build the track from the middle tier staging to where the hidden track meets the sceniced main line track.
                      My goal for the end of this coming weekend is to have a happy solution for item #2. This is a big task that will require a lot of thinking and trial-and-error, and my experience is that I get more done if I also have some smaller tasks as part of the same timeframe goal. So I'll also target item #1, ordering the track for #4, and to resolve 3 of the items on the problem log. I don't think I'll start the DS64 until I have an answer for #2.

                      Saturday, February 6, 2010

                      In the year 2525

                      The title has absolutely zero relevance to anything, except that today I snapped the 2,525th picture on the digital camera I got in 2005 and it reminded me of the song from long ago.

                      The real purpose of this post is to provide an update on the progress regardingthe goals for the weekend. Item number 5 was resolved, per the last post, and yesterday and today I dealt with items 2 and 3. Item 2 is completely done -- the power cabinet rewiring and organizing is complete, the Digitrax panels are in place, and I've even fixed the incorrect Loconet cables in the staging area. I'll cover this in a new post on the Wiring Standards theme, either today or tomorrow.

                      The cabinet and wiring work took many, many hours (6? 8?). A lot had to be changed and I wanted the final result to be as "done" as it can be given that there are still some unknowns about the remainder of the layout. My son has observed that there are certain "fun" activities in model railroading that he loves to participate in, starting with running trains but also including repairing rolling stock. Other tasks, he observes, like construction, are not as fun. The power cabinet task was definitely in the "not fun" category, but now that it's over I'm very glad it's done, and it will make future operations and maintenance much easier. The best part was that everything worked the first time I tested it, even the middle tier AR section. Sometimes -- rarely -- things happen that way.

                      I've also cleaned the remainder of the middle tier staging track so that I can store rolling stock, part of task 3. In order to complete task 3 I have to bring out the CMX cleaning car and start the weekly track cleaning process. I hope to do that tomorrow.

                      Here are a couple of pictures from today. The first can be titled "Earthquake!" as it shows the staging area after I put a bunch of staples in underneath to hold in the loconet wire. Each staple really shook the cars on the layout:


                      The second pictures shows the staging area after all other work was done. You see cars are now organized on the middle tier and, if you look closely, you'll see the Digitrax throttle panels now installed on the front of the layout.


                      My next task is item (1) from that earlier post.

                      Thursday, February 4, 2010

                      Upper AR problem solved (?)

                      I was testing the Upper Auto-Reverse problem mentioned in the last post when it suddenly went away. I determined that the problem was that only one rail was switching polarity, not the other. Trying to isolate the cause I used the multi-tester on the PM42 connectors, but having forgotten I turned the MT off, touching the connectors caused the DCS100 to short circuit and the PM42 to switch polarity again (at least the indicator light changed). That reset apparently stopped the problem, as the PM42 now works fine again.

                      I'm putting this into the "monitor" category on the problem log as I don't understand the solution and thus the problem may reoccur. I'm also wondering if this was the same problem I had with that one "bad" PM42 section in 2007. Hmmm....

                      Tasks and more tasks

                      Had a short, but intense business trip this week. Fortunately successful, as the first site to launch this large new platform worked the first time. Lots and lots of hours so I will be taking tomorrow off to compensate.

                      Being away from the house meant no time directly on the layout, but I did have some down time to write the last three posts on design influences, concepts, and the layout room. I also fired up 3rd PlanIt again and redesigned the track from the middle tier to where it meets the main layout. I'll describe that in a later design post, but that had been a roadblock to any further track construction.

                      Now I'm back, somewhat rested, and am writing this post to organize what to do next. Here is a brief list of tasks needing to be addressed, in a rough priority order.

                      1. Resolve switch machine issue.

                        On the main layout the switches are all still manual. This is mostly okay, but three in particular are impossible to reach unless you go into "the Box" in the center. Back in spring 2007 I was running trains using the first iteration of staging and was finding it a pain to go into "the Box" to set the switches, so I took the two Tortoise switch machines that were left over from the 2002 layout and hooked them up, temporarily using a DC power pack and switch boxes from the same layout.

                        So what about the 3rd switch? Well, being cheap by nature I wanted to avoid buying another Tortoise -- which runs around $17 if you buy just one. In my defense, two of the switches in question form a crossover. One attribute of a crossover is that the two switches always move together, either both facing normal or both facing the diverging route. Thus, if you can set it up correctly you can drive them both with one switch machine. Further in my defense, I have about 15-16 (lost count) crossover configurations on the layout, so the savings of using just one switch machine on all of them would be over $200. And that doesn't count the two double crossovers, in which one machine can theoretically power all 4 switches.

                        I tried this in 2007 but never got it working. Then summer came, I stopped work on the layout, and in December 2008 I removed the first staging and started the replacement project -- as described at the start of this blog.

                        Now we're finding that the one switch that I was working on is a common derailment spot because the spring was removed as part of the installation process. I need to solve this, either by finding a one-machine solution or giving up and allocating two per crossover.

                        This item has to be done before I can start thinking about the DS64 project below.

                      2. Power cabinet.

                        I received the items from Mike in the mail this week. I now have what I need to set up the internals of the power cabinet in it's semi-permanent configuration. As part of this task I would also install the new UR92 on the layout facade by the cabinet and re-install the UP5 and UR91 panels onto the staging area (they were taken out when the old staging was dismantled). Also as part of that I should test the loconet cables for an issue I'll describe in a later post.

                      3. Track cleaning.

                        This includes hand-cleaning the remaining middle-tier track, so I can start using it for storage and test running of more trains, and also starting a regular weekly track-cleaning cycle with the CMX car.

                      4. Track scenery.

                        My attempt last week to paint the Peco ties to look like concrete came out worse than most of my previous attempts. Something I'll describe in a future post. But I really want to solve this before I lay any more Peco track, and that means it has to be before a lot of other work.

                      5. Upper AR issue.

                        The upper tier AR section has a funny problem whereby it works fine in the normal direction, but in reverse it simply loses power to the track. We work around it now, but during sessions where a switcher moves cars from one upper tier track to another this will be a problem.

                      6. DS64 deployment.

                        This is the stationary decoder I mentioned last year. Once the switch machine issue is resolved I should get the DS64 working with the Digitrax system and 4 switch machines. This is a necessary learning experience, a prerequisite for putting up switch panels on the layout facade or using traffic control software.

                      7. Extending middle tier staging.

                        Quite a bit to do here, but the first step will be ordering the needed Atlas code 80 switches and track, as I've run out. So I can get that started this week.
                      My goal for the end of this weekend is to complete items 1, 2, and 3 above and to have a plan in place to solve 4.

                      During the same time I expect my son will continue his work with car repair and soon, decoder installs and engine cleaning.