Presenting at a HUB Division "Railfun Night" yesterday, there was a question about my signals. The best answer was taking the time to move this old RR-Line thread over:
26-May-2015: My layout doesn't have much hidden track, and only one out of sight turnout. When I'm running by myself I can either see the points or see a control lever that shows the position of the one I can't see. But my guest operators sometimes lack that local knowledge. Also, not everybody's eyes are up to seeing HO scale point position at a distance.
201312Eastern4.jpg
I've been thinking about signals ever since I got the basic control system working. This shows my guess of how B&M practice would have been applied to my Bexley Tower interlocking (the lower right corner of the track plan image):
BexleyPanelSchematic2015.png
I've read articles about CMRI-style signals, kept up with Ralph DiBlasi's signaling thread on RR-Line (not saved AFAIK; he was signaling his large HO Lehigh Valley at the time) and operated with them on home layouts and the Hub Division's modules. But I didn't want them at home:
1. For mainstream installations, if the computer isn't running, power turnouts can't be thrown.
2. At startup, those I've seen take time and effort to get the computer's view of turnout state to match the layout's.
3. Local control can be done, but each sub-panel takes a computer and extra implementation work beyond that required for a single CTC console.
4. Each computer is a point of failure, and requires at least a little system admin work from time to time.
What I want is lineside control, where logic at a given interlocking does most of the work. In my area/era, that logic got its orders from nearby signal towers and occasionally manual pushbuttons on the relay cases. Some commercial firms offer local logic boards which can be configured to reflect a particular track layout. But the more complex their options are, the more they look like programming a modern hotel room clock. Or sending text messages from a flip phone. Not the right UI for the complexity of the task. And no source code, so I'd have to work around any idiosyncrasies. And some of them didn't have components to handle the multiple-head searchlight signals my prototype used.
What I settled on is a mix of relays, commercial detection components and commercial signal head drivers. With cheap automotive relays, I ought to be able to keep the cost below $50/block, less in single-direction double track. If the cheap relays don't stand up to my low-current application, I write off their $4/turnout cost and spend $25 for a better grade.
I maintain and use a layout-wide cab control system, with one block toggle position connected to the DCC buss. A plus for my lineside signaling scheme is that I should be able to indicate turnout position regardless of whether the block that it's in is being powered by DCC or DC. And if I find a power-agnostic detector that doesn't require common rail, the rest of the logic should work with it.
My B&M Eastern Route layout thread is here: https://modelersforum.com/index.php?topic=6843.0 (https://modelersforum.com/index.php?topic=6843.0)
27-May-2015: The first interlocking I'm going to do is the east throat of Bexley yard. It's a single switch entering a main track whose normal
direction is eastward (left to right in the XTrkCAD diagram).
EBexleyPanelv1.png
SDxx will be Circuitron SD-3 bipolar LED 3-aspect signal drivers. For the Dx detectors, I was experimenting with OC-1s from Olin's Depot (no longer available 10 years later). This interlocking isn't an ideal application, as I can only reasonably feed two blocks from this point, leaving the other two detector units idle. Mast signals will be searchlights from Oregon Railway Supply (no longer available AFAIK). Dwarfs are probably going to be kit-built or scratched.
To do the full logic, I'd have needed two more SD-3 cards. I've simplified it so when the turnout is thrown for the yard, the eastward main's home signal will go Red and the two dwarfs will go Yellow (Call On). The bottom line is about $45, not including the switch machine.
29-May-2015: Now, components that I hope will make this a lot easier and cheaper than it might have been 20 years ago:
IMGP4091_v1.JPG
A $3.71 (2015 price) automotive 4PDT relay with LED indicating the coil is powered. I bought a 10-pack from an Amazon vendor in Hong Kong. All the coils read between 117 and 119 Ohms, so they'll draw 0.1A at 12 VDC. When I went to ring the contacts out, I found one side of one pole bad on the first (red X). I will report how the rest turn out.
IMGP4092_v1.JPG
Depending on how far I go with this, I might need to know point position on between 3 and 5 old Kemtron-style twin coil switch machines that are already installed. I really didn't want to solder multiple wires to the contact fingers. Spending an afternoon researching slide-on connectors turned out to be a lot easier than removing and re-installing the switch machines:
These FDD2 units for 0.187 blades cost $12 for 200. I'm using 0.25" Kapton tape to insulate them. If I need both a machine's contact sets, I'll probably cut off the crimp portion and solder wires to the slide body.
IMGP4089_v1.JPG
An NJI #1295 'Dwarf Signal - SA Type' (alas, no longer available today) dunked in acetone and re-kitted. With a bi-color LED that fits, I might need a dozen on the layout, and a couple of their 2-head cousins.
Edit Long After: I wasn't able to file a 1.8mm LED package down enough to fit into NJI single dwarf head castings. The 2-light castings are enough larger that I could upgrade them to bi-color. I scratchbuilt 9 single SA dwarfs as described here: (TBD)
29-May-2015: Explanation I should have included with the purple part of my schematic above:
The OC-1's occupancy detectors are NPN open collectors that I'll set up to ground their output when occupied. They can sink up to 25 mA. Circuitron says the SD-3 draws "a tenth of a milliamp" on its Red and Yellow inputs.
So Detector 13 (block 13 in diagram above) will pull down SD13wR (Signal Driver West 13 Red input) and SD11wY (11 West Yellow) when the block is occupied. This is the standard way to wire Circuitron SD-3 drivers.
The 'from switch machine' lead will have +12 when the turnout is reversed, so the relay will pick up. This will ground SD13wR, changing the eastward home signal's top head to Stop (red). The home signal's two other heads are fixed red and their wiring isn't shown. Other relay contacts will change the two dwarfs from Red to Yellow (not strictly prototypic).
If Digi-Key had bi-color Red/Green 1.8 mm LEDs (I'm sure they did), I couldn't navigate to them. I found 2-lead red/green at LED-Switch dot com (also gone in 2025). I was also able to find Red/Yellow, which I used in yard entrances.
I tested the rest of the relays: two each a normally-closed contact which didn't conduct. But taking them apart and pulling a strip of paper through the offending contacts fixed them. (in 2025, the Hong Kong relays are still working fine).
IMGP4094_v1.JPG
Since East Bexley only needs one 'make' and one 'transfer', I spent a while seeing if I could do it with the switch machine's DPDT auxiliary contacts (there's another 'transfer' but I'm using it to power the frog). As my rusty basement refugee Kemtron demonstrates, even simple modifications don't give this kind of slide-on connector access to all the machine's contacts. Rather than strain things to avoid the relay, I'll do it as I drew above, with one 'make' on the switch machine to power the relay.
I can't do the dwarfs till the LEDs arrive, but I can build the home signal, wire the relay, detector and signal driver and maybe see
'first light' over the weekend.
1-Jun-2015: This interlocking needs a new 3-head searchlight 'home' signal. The location means the ORS plastic platforms & railings would be both protected and not looming in the foreground of photos, so I didn't use my out-of-production Free State Systems etched parts.
IMGP4101_1.JPG
Oregon Rail Supply had been selling inexpensive signal kits since before 1990. Production got irregular after the founder passed in 2015. I've built several, but 3-head kits like this hadn't been in production since sometime before 2010 - possibly because it's so fiddly getting 6 wires down the 3/32" brass tube mast.
IMGP4104_1.JPG
After applying some profanity followed by Floquil Bright Silver, I set it aside and unpacked the 1.8mm LEDs that came in today's mail:
IMGP4110_1.JPG
I gave up on trying to bend the leads to fit the stem of the signal casting as the ?Chinese? had done with the original - they'd probably started with axial leads. Instead I soldered 30 ga. Kynar to the cut-off stubs.
This photo used all the magnification I had, 48:1 if I read the labels on the old 100mm macro lens right.
6-Jun-2016: Thursday the signal project made a good deal of progress: first light at the Robinson Rd. (East Bexley) home signal. But I didn't take a picture and Friday made mistakes which undid most of it (6 LEDs burned up by two different mistakes). Here's one of my homebrew styrene dwarfs installed:
IMGP4119_v1.JPG
And here's one Mieke (I was dating her then; she made several contributions to the Eastern Route) took of me under the layout:
IMGP4121_v1.JPG
Good thing she wasn't around to take pictures when things went to hell.
8-Jun-2015: The final (I hope) version of Robinson Rd. Interlocking passed an out-of-place test this evening:
IMGP4128_v1.JPG
+/-12 VDC and ground enter at the top left. The Circuitron SD-3 signal driver is to the left. The three terminals above it will control the two dwarf signals from the 4PDT relay at center. A Chubb-design DC Optimized Detector I bought from BurleyJim is at right. The two terminals with jumper at the top center go to the track. The DCOD is pulling down (electrically, it's grounding) the Red pin of the SD-3 and it's indicating Stop on the Oregon Rail Supply dwarf I'm testing with.
The wires from the switch machine contact to operate the relay aren't present; I tested them Friday. I also need to install a diode across the relay's coil to protect the switch machine contacts. The wires to the active head of the home signal and the Yellow input to the SD-3 will enter the assembly via slide-on connectors.
Lessons learned so far:
1. No matter how tiresome it is to get the 30 GA wire from the signal properly gripped in a terminal block, it's worse to fix a signal in place because you can't pull heat-shrink splices back through the baseboard.
2. The DCOD is more economical for single installations, but also works better with the 'Remote' blocks on my layout because they don't care whether the Master block is set to DC or DCC. I'll eventually use 5 more DCODs for Remote blocks, with Olin's OC-1s for the main panels.
3. If I'd built/debugged the dedicated 2 VDC LED power supply first, I'd have saved myself 6 LEDs and a signal handrail and the whole interlocking might be in service now.
4. Test everything at every step, but don't do initial tests of control electronics with with layout-quality signals.
5. Stranded 30 GA DCC decoder wire is lousy for soldering to tiny LEDs.
9-Jun-2015: The eastward home signal is now indicating both turnout position and track occupancy well enough that operators should be pleased. It won't show Yellow until I install signals in Newburyport.
I started designing the end of double track/drawbridge interlocking at Newburyport, so I haven't finished the 2 VDC power supply. The fixed LEDs in the home signal are waiting on that. The two dwarfs that only indicate turnout position need that, plus me finding a reliable set of steps that will make a 1.8 mm LED fit into them while still displaying both Red and Green.
10-Jun-2015: This morning, the signal engineer will submit the draft plan for Newburyport to the Superintendent:
NewburyportPanel201506.png
After I posted, I noticed I'd cut off the SD-3 labels. From the left, they are SD12e, SD17w, SD16w and SD16e.
Operationally, I need signals in Bexley more than I do Newburyport, but designing this gave me a better understanding of how to protect non-powered turnouts and use diodes to partition the logic. The diode on this diagram is keeping westbound Block 14's detector and turnouts from affecting the eastbound signals.
I am cheating a bit, as the 17w stagger block will show double yellow when 16w0 is red, instead of yellow over red. And because block 17 is signaled EB only, 16e shows bottom yellow (call-on) when the double track turnout is normal, regardless of 17's occupancy. Either could be fixed with another SD-3 driver.
Anyway, this felt good to finish, as the LED power supply was still frustrating me.
10-Jun-2015: Feel free to ask questions. I was figuring this out on the fly and anything that isn't clear might not have been thought through. Better to fix it on paper than after it's wired.
Notation (it seems obvious to me, but I devised it):
Block IDs are 2-digit numbers. 09 through 16 figure in the Newburyport plan.
Turnout IDs are block_id T sequence_from_west, so 14T2 (the Hytron spur) is the second turnout in block 14.
Signal IDs are block_id end sequence, so 16e is the home signal at the east end of block 16. 16w0 is the home signal at the west end, and 16w1 is the 'against the current of traffic' fixed red dwarf.
Signal driver IDs add SD to the signal ID, e.g. SD16w0 and SD16e.
Relay IDs add an R to the turnout ID. Only power operated turnouts have relays. Others have a contact on the mechanism to ground appropriate SDs' "Red" inputs.
Detectors ground their output to indicate 'occupied'. Signal drivers operate off grounds at their Red and Yellow inputs. Turnout control is completely independent of the signals.
I'm following prototype practice generally, which is why you'll see electrical block boundaries without signals. It is well known that the B&M didn't do anything like it was made out of money. If a possible signal location wasn't going to earn its keep, the function was implemented on the signal to the rear. This is why there isn't a 14e; 16e does that job. There was little operational reason for EB moves on the WB track at the end of double track, so the dwarf was fixed red. If you needed to pass it, you called the drawbridge operator for permission.
There will be some shortcuts, mostly limitations on displayed aspects and not bothering to display track occupancy on most of the dwarfs. Any current train crew likely to operate here will have been using NORAC rules for at least 25 years, if not their entire career, so my pre-NORAC RR shouldn't trouble their reflexes. If a retired B&M signal maintainer should comment, I'll be happy to listen; it's all modular so $20 for another SD-3 and a few hours designing the changes and implementing them won't be a big burden. But right now I want to prove feasibility and get the basic system up and running.
16-Jun-2015: There will be one change to Newburyport before I start implementing it: I'm being inconsistent about 0-based (signal) and 1-based (turnout) numbering. I haven't decided which to change, but working on the bi-directional double track for Bexley brings it out.
Pete (Orionvp17) asked about 0-based numbering. It's a computer programmer thing - when you're numbering items, is the first one #0 or #1? Most people would say '1', but if you're using the number as an index into an array (a bunch of identical objects stored next to each other in memory), '0' saves a couple of math operations every time you touch an object. That may matter a lot if the objects are heavily used. Here, there's no benefit for counting zero-based vs. one-based, and mixing the schemes reduces consistency and clarity. So I converted signal numbering to 1-based.
21-Jun-2015: At the previous week's NRHS convention in Rutland, VT I worked on the Bexley signal plan during some idle time:
BexleyPanel201506.png
7T5 and 7T4 are SPDT contacts on the hand-throw mechanisms entering the yard and enginehouse. rDC and yDC are the positive and negative feeds for the bi-color dwarfs I'll use.
This does about 90% of what I want, but I don't have enough contacts on the crossover relay (R6X2) to provide 'call on' (Bottom Yellow) indications on signals 05W, 06W1 and 06W2 at the interlocking's west entrance. I'll sleep on it before embarking on a redraw.
28-Jun-2015: I was feeling happy about progress on dwarf signals ( <TBD> ) so I went to work on designs that use them:
DrawPanel201506.png
My staging is open and easy to see, so I'm not bothering with detection. I don't really need departure signals either. The east end signals will only indicate Stop or Approach, but the westbound home signal will change between High and Middle Yellow depending on the crossover position. The west end's signals are covered in the single Newburyport drawing since they don't interact with the east end's.
The mechanism slide switches on 01T2 and 02T2 are using all their contacts for the swing blocks & frogs. Adding a couple of SPDT microswitches to their mechanisms shouldn't take an unreasonable amount of under-layout time, though.
And an oops: The dwarf at the west end of the East Lead should be 'D3e', not 'D2e'. I've fixed the plan's track schematic and diagram for 01T2 but I really should build some more dwarves instead of fixing the screenshot above.
9-Jul-2015: I finally got the first interlocking working: Robinson Rd., the east entrance to Bexley yard. This was after burning up the two dwarfs one more time (wiring error, should have documented the actual interlocking board better):
IMGP4474_1.JPG
Here's the removable board with the SD-3, relay and DCOD detector, as installed. Next I wanted to start on Newburyport, but I couldn't finish it till I got more Oregon Rail Supply parts.
Thank you for responding. Great info here.
Thanks, Dave. Back to 2015:
19-Nov-2015: Back to work on the signals: East Bexley interlocking proved helpful at my last two op sessions, so it's not just for show. And over the past 4 months I'd accumulated everything I need to complete the north end of the room:
NewburyportPanel201511.png
I'd re-drawn Newburyport to show inter-signal cabling. The longer segments can be done with telco 'quad'. The six wires needed from Newburyport West across the Merrimack River to Draw will be recycled ethernet cable. The lines from block detectors D14, D16 & D17 come from the Draw panel because the Olin's Depot detector assumes common rail and I have to put it on the feed side of my anachronistic block toggles. While I'm building the signals and working out how to make them removable using DIP in-line sockets, I'll add the wire colors to be used for each circuit to the drawing. And fix the 'connection' dots so they show on screen captures and printouts.
30-Dec-2015: I'd been wanting to get back to work on my signals, and Sunday's op session emphasized the benefits I'm hoping for. So I spent the last couple of days implementing the Newburyport diagram above. I wasn't finished, but one bit of engineering has been demonstrated and I've figured out how to address long-haul wiring.
I'd always considered it important that signals react to the position of turnouts. So back when I was building my hand-throw turnouts in Newburyport, I used DPDT slide switches. One transfer contact for the frog, the other for signals.
IMGP7339_v1.JPG
Today's wiring got me Stop (three red) at Robinson Rd. when this turnout is reversed, and Caution (high yellow) when a turnout in the next block is reversed. The center pole is DC ground, the 'closed when reversed' goes to Block 15 Occupied (see below).
If I was doing this 20 years ago, I could have scrounged any quantity of telco 25-pair cable from office renovations. With some 50-foot scraps, I'd have used different wire colors for every circuit. But what I had today was telco 'quad'.
IMGP7343_v1.JPG
Now I'm enlightened about how prototype signaling worked. The same positions on the crossarms would often carry different circuits in advance of a signal vs. in the rear. I'm doing the same with quad under the layout: To the left, Green is Block 15 Occupied, to the right Green is Block 14 Occupied. Yellow carries Block 17 Occupied through.
I'm noting all this on my printed drawings, to be copied onto the machine-readable files Real Soon Now.
2-Jan-2016: Over the past few days I had installed my first OC-1 quad DCC-only detector from Olin's Depot (upper left):
IMGP7347v1.jpeg
It had to go on the feed side of my block toggles, so I replaced the old wiring. Here it's indicating occupancy in Block 17 (EB main through Newburyport depot). The open-collector output is grounding the Oc17 circuit the same way my turnout contacts do. I haven't played with its adjustable occupancy threshold yet.
IMGP7350v1.jpeg
Looking east from Bexley yard, the 'high yellow' (Approach) aspect isn't ideal, but using bi-color LEDs saved me enough time and money that I plan to get used to it. And I wasn't considering Approach Lighting; At the time I thought that would double the complexity (not true, I didn't know about Rob Paisley's driver boards, but in 2025 he's only to be found in the Internet Archive, it seems).
NewburyportSignal20160101.png
This is about the level of documentation I thought I'd need down the line. I may revise this one more time as I build signals - as shown, 'stagger block' 17w will display 'double yellow' (Approach Slow) when 16w1 shows Stop. The correct B&M rulebook aspect is 'yellow over red' (Approach). RR employees who might notice that won't be operating my layout very often, but it'll probably happen. And I'll know. I can do that if I use a 2nd SD-3 driver and a diode or two.
3-Jan-2016: Pete (Orionvp17) said this was over his head. I replied: If you think of the Circuitron SD-3s I'm using as black boxes which turn a bipolar LED red, yellow or green, everything else just grounds the proper Y or R inputs to display the aspects I want. This diagram shows how the 17w 'stagger block' can show Green/Green, Yellow/Red and Red/Red:
NewburyportSignal20160101.png
The 'High' head defaults to green, shows yellow when block 16 is occupied and red when block 17 is occupied. The 'Bottom' head defaults to green, but shows red when either 16 or 17 is occupied. The two diodes keep blocks 16 and 17 from affecting each other anywhere but at this signal driver.
Next I built the actual 17w signal.
4-Jan-2016: I'd been procrastinating building signals while I wired; today I reminded myself why. The 'stagger block' signal 17w has taken me two or three hours already and is far from complete. Maybe the second will go faster...
Oregon Rail Supply's 'searchlight' parts come closest to what the B&M used, but I need to make some changes and some of the boughten parts are both more fragile and more clunky than those I can make myself. ORS heads & brackets are good, given that I'm settling for T-1 (3mm) LEDs instead of scale-diameter lenses. ORS platforms are better than I can do myself, but I sincerely wish I'd bought more of Free State Systems' etched ladder and platform sets before they folded.
IMGP7353v1.jpeg
ORS' handrails only work at the top of a ladder, and the B&M had platforms with handrails at each head. Detail Associates .010 x .018 flat brass bar to the rescue. Happily, I didn't melt any of the plastic while soldering the brass:
imgp7430v1.jpeg
Here it's ready for ACC on the metal/plastic joints. LEDs, a mounting socket and paint the next day maybe.
5-Jan-2016: A little more engineering to blog tonight:
imgp7438v1.jpeg
I'd seen photos of people using IC sockets as signal connectors, but hadn't bookmarked the pages :( I purchased several Mil-Max '8-pin IC socket 0.3" spacing, 0.1" pitch' (DigiKey part ED90048-ND). I won't ever need more than 6 contacts, but the 8-pin parts can't really be cut down. Many different IC socket sizes are available.
I wanted to shim the signal up above the sockets, so I tried out my solvent cements on the plastic body: Weld-On acrylic cement bonded the two blocks of .100 x .125 Evergreen styrene visible between the pins just fine. Once I'd soldered the #30 LED leads into the sockets, I used MEK to bond the styrene signal base to the blocks. Then I soldered two pairs of telco 24 gauge to the appropriate pins on the fixed socket.
IMGP7355v1.jpeg
Plugged together, the two IC sockets are 3/8" on a side and 1/2" deep. I drilled two 1/8" holes about 1/8" apart, then cut a hole to fit in my 1/2" Homasote roadbed. Here I'm feeding the leads down the holes.
IMGP7358v1.jpeg
The signal needs a finial and paint and it's lit by the Diode Test function of my VoM, but it's a lot closer to done. It won't take a lot of wiring to hook up the SD-3s, maybe tomorrow.
6-Jan-2016: RR-Line members Orionvp17 and BurleyJim had complimented my documentation: Thank you, Jim and Pete. The documentation is organic to my 'design, THEN build' process. Life is too short (and my back & knees turn 60 in a week) to do this by crawling around under the layout, muttering 'ok, I need another wire from here to over there' and shouting '%$#@&^' when I burn another few LEDs up. I learned from building the first interlocking, so Newburyport's design got revised several times. Implementing Newburyport's signals pointed out several errors in my first design for the bi-directional double track in Bexley, so I'm working on that in parallel.
Bulletin Order #4, January 6, 2016:
Eastward Automatic Signal 17W at MP 36.0, Newburyport, West has been placed in service.
IMGP0945_v1.JPG
IMGP0949_v1.JPG
Block 17 occupied, Stop.
IMGP0947_v1.JPG
Past the end of double track and onto the drawbridge: Caution. The yellow might improve with adjustment.
IMGP0945_v1.JPG
Into the staging yard and Clear again. I still haven't put the mileage number on this automatic signal to distinguish it from an interlocking signal. B&M didn't use a flat plate I could letter, instead they attached metal numbers to brackets off the mast. Hard to execute, and then the mast has a delicate part attached below the lowest platform. Exactly where I put my fingers when I need to install or remove a signal.
8-Jan-2016: I was asked about the socket mount: I'm building a couple of 3-light interlocking signals with better photography of the socket work. Because I'm just plugging one socket into another, the pins aren't held strongly - a light impact will knock the signal right over, which I believe is what I want. Somebody who knew more about IC nomenclature might be able to find a 'dummy IC' with pins shaped so they'd be gripped more tightly.
I'm concerned to find that the Oregon Rail Supply website (www.oregonrail.com, printed on packaging) is not resolvable in DNS today. I know the founder passed in Aug. 2015; at the Portland National Train Show, I met his widow at the ORS booth. They had product to sell and she was intent on continuing. There isn't anything else on the market that suits my B&M prototype as well. I thought I had my needs covered, but I just found out the signal heads in part #102 have smaller targets and are molded differently from those in the 'searchlight' range (#125, #127 & #130).
Later that day I got a 3-light interlocking signal ready for LEDs and assembly to the socket:
IMGP7366_v1.JPG
The upper sockets haven't been modified; the lower left one shows where I remove a couple of molding pins which interfere with the .100 x .125 styrene shim blocks.
IMGP7368_v1.JPG
IMGP7369_v1.JPG
This one will be quite close to the audience, and ORS hadn't provided a long enough piece of ladder in their #130 kit. I used Free State Systems (out of business) etched ladder and platforms, soldering all the brass/brass joints. This required care heat-sinking to avoid melting the signal head brackets that had to go between solder joints. And of course I mis-oriented some parts the first time...
11-Jan-2016: A few more shots as I wired up the socket on the 3-head interlocking signal:
IMGP7371_v1.JPG
Here I'm starting to insert the LED leads into the IC sockets.
IMGP7375_v1.JPG
Here, they're all in, ready to solder. I used a 15W iron and Kester .020" rosin core 2% silver solder. And my binocular magnifier. And then I tested all the LEDs again.
Finally, I used solvent cement to bond the ORS base to the two shim blocks. Then I ACC-ed the signal heads in place. Next is wiring the fixed socket. But I started digging around for my bag of heat-shrink tubing (which I was using last week) or my wire-wrap tool (which I haven't used this decade, but did see last year, just can't recall where). No joy on the wire-wrap tool 9 years later. I'd probably have to get a new one off the 'net.
12-Jan-2016: The socket was Mill-Max Manufacturing Corp. '110-44-308-41-001000', 'ED90048-ND' at Digi-Key. The ladder visible here is an out-of-production etched part that Free State Systems used to make, their '8001-1'. Pack of 8 with platforms, nice etching, long enough for any configuration I'd need, but now only on eBay. The 'stagger block' on the previous page uses the punched ladder that ORS includes.
I'm sacrificing flexibility for simplicity in implementation, so this particular signal uses 3 different LEDs to display Medium Clear, Medium Approach, Restricting and Stop: The top head is fixed red, Digi-Key #160-1704-ND. The middle is Oregon Rail Supply's stock bi-color LED. The bottom is a "3 mm red/yellow w/c 2 lead LED" from "led-switch" (since closed). "w/c' meant 'water clear'.
LED Switch also had LEDs similar to ORS', though not with the translucent white 'diffusing' ORS uses. I could probably get a closer match at Digi-Key, but haven't felt the diffusing was a major priority. It probably would help the fake Yellow.
20-Jan-2016: Work continued, but I'm not satisfied with a compromise I made early on. Most dwarf signals and a number of 3-light home interlocking signals will have a "Call On" aspect (yellow, or bottom yellow on a 3-light). I'm driving these from +/- DC voltage using a transfer contact on the associated turnout's relay. An example is R16T1 (relay) driving the bottom head of 16e (home signal) in the Newburyport diagram on the previous page.
This is OK for movements into unsignaled territory. But at Newburyport Draw, signal 16e shows "Call On" even when an EB train is in the block, and I want to do better. I would like to be able to drive a single bi-color LED based on whether or not an occupancy circuit is floating, or grounded. I could do this with a Circuitron SD-3, but they're $16 and about 2x4 inches. After a couple of fruitless hours searching the web for 'bipolar LED driver' and the like, I checked Paul Mallery's chapter on signals and found no help.
In 2025, I'm still hoping EEs or serious electronics hobbyists suggest a circuit: one input (ground, float) +/-12 VDC power, 2 outputs (one per LED lead) switch polarity based on the input? Must be smaller/simpler/cheaper than an SD-3.
Bulletin Order # 12, January 24 2016:
Eastward 3-light interlocking signal at MP 37.4, Newburyport (End of Double Track) and Westward 3-light interlocking signal at MP 37.7, Merrimack Draw have been placed in service.
IMGP7407_1.JPG
The prototype had 500' between Merrimack St. and the end-of-double-track turnout, so their home signal had a conventional foundation. I had to imagine how they'd have addressed putting it atop the retaining wall, too close to the track centerline for a standard footing:
IMGP7404_1.JPG
I set the socket in Wood Putty and built a styrene box to install around it, like they'd poured a pier set into the top of the wall. I'll put some rust spots on it to suggest anchors going deeper into the fill.
IMGP7382_1.JPG
Here's the other end, showing the Call On aspect indicating the turnout is lined for the eastward track. I'm living with this for the moment: A friend I've known since MIT had a similar signal project at the North Shore club in Wakefield, and offered to work with me on developing a board using a low- power H-bridge motor control chip, possibly one of the LN293x series. Alas, that never got off the ground..
IMGP7402_1.JPG
The westbound half of the interlocking has an SD-3 and a DCOD, set up to detect only the part of block 16 west of signal 16e. This avoids having 16e go red as soon as a WB train pulls out of its staging track, before the train reaches it.
IMGP7401_1.JPG
The eastbound half has the relay for turnout 16T1 and the SD-3 for signal 16w1. I've since added a diode across the coil to protect the switch machine contacts.
[edit: picture taken before signal 16w was installed, as indicated by the test LED hanging from the SD-3's output terminals]
Next, I worked on the westbound track, which meant beginning the occupancy detection in the Bexley Enginehouse area.
26-Jan-2016: MarkF and Orionvp17 complimented my work. Thanks, Pete & Mark.
My career included both 'out where the streetcars don't run' software engineering and 'explaining to co-workers so I don't have to do it all myself' teaching. From a process PoV, I've broken signaling up into manageable (IMO) chunks local to individual signals and interlockings. I did circuit diagrams for those who work that way. The photos of the interlockings show how they look in real life, with narrative replacing sticky labels and wire tags all over the place (you may see the labels & tags when I get to the bi-directional double track part).
3-Feb-2016: Detection was in place on 8 of my 16 main line blocks. Signals protected operation over 5 of them.
I finally got the Bexley CTC (Extending from Saugus Jct. to D'Arcy Ave., under control of the Train Director at Bexley Tower) drawn sufficiently to my satisfaction. And I started wiring it.
BexleySignal20160203.jpeg
First came the simple interlocking at the end of double track (D'Arcy Ave.). Then the somewhat more complex one at Bexley Enginehouse (east portal of the tunnel). I'll probably have to revise the drawing to make room to document all the circuits between Bexley Enginehouse and Saugus Jct. (I split it into Bexley Tunnel and Bexley Tower). It was clearer in XTrkCAD, but I ran out of screen.
10-Feb-2016: While my unregulated phone/DSL company was waiting for the most economical time to repair their downed cable, I worked on the Bexley CTC (under control of the Train Director at Bexley Tower) design. And got far enough into it that I felt comfortable starting construction of the Enginehouse area:
BexleyEnginehouseLogic.png
These interlockings are westbound-only, for the end of double track at D'Arcy Ave. and the west entrance to the yard just before the tunnel. Signals governing eastward movements through blocks 7, 9, 10 and 11 are at Bexley Tower, west of the tunnel:
BexleyEnginehouseTrack.png
The easterly 'pole line' carries 3 circuits: occupancy indications for blocks 9 & 10 coming from the Bexley panel, and block 13 coming from East Bexley.
Between Bexley Enginehouse and D'Arcy Ave. there are 6 circuits: Occupancy for blocks 9, 10, 11 and 13, a circuit triggering Yellow from driver SD9e1 and a 'turnout reversed' from Bexley panel to control relay R9T1.
Let's consider D'Arcy Ave:
SD9e1 gets its Yellow input from Bexley Enginehouse. Red comes from Oc9 OR'ed with R9T1 'normal' (for eastbound moves). The diode keeps the turnout position from changing the state of Oc9 visible to other interlockings.
The top SPDT (transfer) contact on the relay does turnout position. When it's reversed, Oc11 is grounded (activated) preventing EB moves from taking the wrong track.
The second SPDT selects whether dwarf 9e2 gets + or - voltage, changing it from red to yellow.
After posting I got back to work on the Bexley Enginehouse interlocking and its yard exit dwarf signals.
Bulletin Order #17, February 9 2016:
Westward 3-light interlocking signal and dwarf signal for against-current-of-traffic moves off the eastward track at MP 12.5, D'Arcy Ave. have been placed
in service.
14-Feb-2016: I got the Bexley Enginehouse interlocking working that afternoon. It wasn't completely functional until I finished the one to the left (RR west) of it, as Enginehouse needs relay R7T2 to tell it the position of that turnout, and thus whether to display the aspect on the Middle or High head of signal 7e1:
BexleyTunnelE_Logic.png
Here's relay R7T3 and drivers SD7e1H (high head) and SD7e1M (middle) with power and a few of the external inputs wired:
IMGP7418_v1.JPG
The diode that's installed implements logical OR of the turnout position and block 7's occupancy (Oc7). Turnout position shouldn't affect Oc7 because other signals need it unmodified. The two diodes in my hand below OR turnout position+Oc7 with Oc8 for SD7e1M's Red input. Oc8 isn't affected either:
IMGP7419_v1.JPG
This is how the interlocking left the bench, with the Red inputs for the two SDs swapped :erm: and some details of the logic driving the yard exit dwarves (7e2 and 7e3) not yet worked out:
IMGP7422_v1.JPG
Here's how it looks, installed and correctly wired:
IMGP7429_v1.JPG
I need to work out how to build 2-light dwarf signals before I do the next interlocking, which will also give my back/knees a break from crawling under the layout in the cold.
22-Feb-2016: As I posted in my 'Prototypic dwarf signals in HO' thread, I fitted my NJI two-head castings with bi-polar 1.8 mm LEDs without problems. Then I found a couple of logic issues with the Bexley signal plan posted above. I have resolutions, but didn't post again till both ends of the tunnel were properly interlocked.
IMGP7440_1.JPG
This stagger block will replace the current WB single-head block signal on my Rowley River module. It's needed as an approach signal to the D'Arcy Ave. interlocking, and I've got a scheme to display a couple of fancy aspects on it. But first I need to analyze the output of the Logic Rail Technologies 'Signal Animator' that'll still be used when traveling with the Module Group and make sure I don't let the smoke out of something.
The Eastern was upgraded from semaphores to searchlight signals in the late 1930s. But true to its roots, the B&M re-used everything they could, including many base boxes that used to house semaphore mechanisms. Sunrise Enterprises offers both single and double (seen here) HO boxes in pewter - drilling this one out to 3/32" required an extra-long drill and patience.
IMGP7436_1.JPG
FSS platforms (bottom) were sized for a scale diameter mast, somewhat less than 1/16". Having made a bit of a mess fitting them to the 3/32" masts used by ORS components, this time I thought first:
1. Heat the ring portion of the platform red hot (I used my gas stove) to anneal the brass.
2. Expand the ring to 1/16" using that hole in my MicroMark punching plate. I used a round nose plier jaw as the drift.
3. Cut the outside portion of the ring with nippers.
4. Expand to 3/32" using the next bigger hole in the punch plate (in progress in the photo).
A finished part is lying in front of the punch plate - much neater than I'd achieved before.
Free State Systems closed their doors for the last time around 2020. What I used here was from '8031 ladders & platforms'.
28-Feb-2016: Some recent progress wasn't perfectly executed:
I found that D'Arcy Ave. and Newburyport End of Double Track weren't using the proper logic to generate Yellow. Easy to fix but I didn't have a spot on the terminal strip so I had to run the wire direct from the relay to the 'pole line' along the layout's framing.
I found out that a 'positive logic' AND is equivalent to a 'negative logic' OR. Luckily, the 'ground is active' components I'm using look like they'll let me use Diode-Resistor Logic to generate the Yellow output from the two active heads of the Enginehouse home signal. I'll show the picture once I have it working.
The 2-light stagger block I built is now serving as the westbound distant signal for the end of double track at D'Arcy Ave. But it's using the Signal Animator which doesn't detect occupancy. Full integration is at least one more interlocking away.
I bought some of Rob Paisley's 556-based bi-polar LED driver boards: (vanished from the 'net sometime after 10/2023):
Plus: They do approach lighting, which is the way almost all B&M automatic block signals worked. Plus: They're smaller and cheaper than Circuitron's SD-3.
Minus: I can't tell whether they're displaying 'yellow' or 'red' on my Oregon Rail Supply LEDs unless I'm watching as the aspect changes.
I don't see a way to adjust the 'yellow' color, so I've emailed Mr. Paisley. The SD-3's yellow is screwdriver-adjustable.
And finally, I know automatic signals should have numbers - it's how crews could tell a permissive block signal from an absolute interlocking signal. But the B&M didn't settle for just a painted plate:
BM_AutoSignal76.png
Model Shipways offers etched brass numbers and letters (the Eastern used a 'P' prefix) in 2 mm and 2.75 mm sizes, but their image doesn't enlarge so I'm not sure what I'd get for my ten quid.
27-Feb-2016: A 'Caution-ary' Tale (as in the yellow 'Caution' aspect of RR signals):
IMGP7460_v1.JPG
Yellow displayed by Oregon Rail Supply's stock bi-polar LED driven by a Circuitron SD-3 and adjusted to my taste.
IMGP7467_v1.JPG
Yellow displayed by ORS stock LEDs driven by a non-adjustable Signal Animator from Logic Rail Technologies.
IMGP7471_v1.JPG
Yellow displayed by Ligitek Electronics Inc. part LHG 2063 driven by a Signal Animator. Note how far apart the red and green dies are?
IMGP7474_v1.JPG
Yellow displayed by an ORS stock LED driven by one of Rob Paisley's 556 Signal Driver boards. I can probably learn to tell it from Red, but I don't think I can expect that from my operators.
I exchanged a number of emails with Mr. Paisley on this matter, including scope traces.
3-Mar-2016: The eastern half of Bexley's signal plant is now in service, including CTC on the single track and the yard exit dwarfs which I hope will help my operators.
BexleyTunnelE_Logic.png
Most of my hobby time over the past 10 weeks has gone to signaling, so I feel like a break. I'll clean up the layout, have an op session or two and work at a slower pace on the issues in my last two posts.
Regarding Rob Paisley's 556 signal driver boards, he thought if I spend a little quality time under the layout with my oscilloscope, he may be able to advise me how to tune a better 'yellow' output.
16-Mar-2016: The Eastern Route's signal plant worked flawlessly throughout today's op session. The operators liked it, and there was a good deal less 'feeping' from my PSX DCC circuit breakers in signal territory. But I'm pretty busy through April, so I'm not sure when the signal crews will return to Bexley and Lynn.
30-Nov-2016: I had no thought of signals when I built all the main track on my layout. I knew I liked them, but I hadn't thought about the technology or infrastructure. Part of the price is visible below:
IMGP0920_v1.JPG
I had used 3PDT slide switches for two locations where I wanted a manual turnout to control a 'swing block' ('X Section' in Paul Mallery's book). One pole powers the frog, the other two select which block's power is fed to the swing block. But if these turnouts were to be protected by signals, I needed another pole.
I found 4PDT slide switches with a reasonable operating force. But they were larger and had a longer throw than the 3PDTs I'd originally installed. So I spent a number of hours under the layout rebuilding two turnout mechanisms.
6-Dec-2016: I finalized the Riverworks (Draw staging east end) interlocking yesterday and built it this afternoon:
Draw_panel.png
Eastbound and Westbound are separate and almost identical in this instance. EB uses a two-light dwarf with the bottom head fixed. WB will get a three-light home signal with RYG on top and RY on the bottom for entering the 'non-signaled' staging tracks 4 and 6 (someday, when I take up my soldering iron and build what the B&M called a "bracket arm").
IMGP0923_v1.JPG
I could have built the two sides on separate boards, but it was a little simpler with only one set of +/-12 VDC terminals. Next is to mount it and build/install the remaining signals.
NJI_2HeadRepair.png
Only the top head of this NJI dual dwarf needed a red/green LED, so I'd saved the manufacturer's red LED in the bottom position.
All was well till a lead pulled off when I was installing it on a styrene base.
I went outside and split some wood, but came back to it late in the afternoon. Luckily there was somewhat of a blob of solder on the terminal that lost its lead, so I was able to re-solder without disconnecting the other leads.
The westward home signal is a bit up in the air. I've built a number of ground-mounted 3-light home signals, but this one would have to go between two tracks in an area with frequent switching.
imgp1866_v1.jpg
I don't recall any commercial offerings of what the B&M called 'bracket arms', but I could build one. Probably out of brass, because I've seen the beating plastic signal bridges have taken on other people's layouts. But that won't be done in a day... (or even 9 years so far).
11-Dec-2016: I got a little prototypic eye-candy out of extending the signals through Draw staging. The Newburyport West stagger block shows:
IMGP0945_v1.JPG
Clear
IMGP0947_v1.JPG
Caution and Approach Medium.
NewburyportWestApproachMedium.png
The signal driver for the top head (17wH in the diagram) used to get its Yellow Input from the signal in advance (16w) being Red. Now it gets Yellow when either the next signal OR the one after it is Red. The two diodes connected to the YI line implement that:
IMGP0943_v1.JPG
With the black band away from the Yellow input, it will get pulled down when either diode sees Ground on its line. Same thing is happening with the Red input of the lower head (17wM).
After I got that working I built a 4-detector board for West Lynn, but pre-snow shopping and the distractions of being a small-time author prevented me from installing it this evening. Tomorrow it should make a nice excuse for not driving in the expected frosty slop.
16-Dec-2016: Last winter (near the top of this Modelers Forum page) I mentioned that the westward end-of-double track home signal (9e1) at D'Arcy Ave. wasn't doing Middle Yellow right. It's complicated because the next westward signal (7e1 at Enginehouse) has two active heads, High for the straight route at the west end of the Tunnel, Middle for diverging. I could have done it with two transfer contacts, but that gets complicated quickly.
Happily, I (a non-EE) had found out that Diode-Resistor logic could AND the Red inputs for the 7e1H and 7e1M together - only if both are Red should 9e1 show Yellow. Last night I'd gotten far enough with Saugus Jct. that I needed to draw two of these AND gates. So I spent today figuring out how to implement one that worked:
DiodeResistorLogicAND.png
If either 7e1Hr or 7e1Mr is logic high (more than about 6 VDC), the diode passes it faster than the 100K Ohm resistor can draw 9e1My
down. Only if both are logic low (0-2 VDC when I've measured it), does the 100K resistor succeed in pulling 9e1My down (active).
The first I built was outright wrong. I experimented with the 2nd till I got the resistor right - 10K was way too small, 82K
worked, 220K was too big. In the middle, I chased a flaky in the occupancy logic at the other end of the tunnel - Problem Disappeared
In Diagnosis is an unsatisfying way of closing a ticket if you know you're in line for it next time it pops up.
So now back to drawing my layout's last interlocking
20-Dec-2016: I checked several on-line model RR vendors for bi-color 3mm LEDs suitable for use with the Oregon Rail Supply
searchlight heads I've been using. Alas, I only found 4 of ORS' own #126 red/green LEDs in stock anywhere, and they're
$5 for two. Circuitron's red/green LEDs are $3.95 each.
Can anyone point me at a manufacturer part # with red & green dies (the light emitting bits) set close enough to make
a decent 'yellow'? Unlike the Ligitek 2062s I showed on 27-Feb-16?
Or has anyone like the results with a less expensive part, like those .LED-Switch used to have ? I have to order a few blue
LEDs for dummy masts (I'll explain that when parts are on hand to build the two 'bracket arms' I need), so it would only
be another $1.50 to try 10 of them. Life is short...
8-Jan-2017: The holidays, houseguests and travel left some room for progress on my signals. I completed the design for the last two interlockings and completed one:
IMGP0952_v1.JPG
Here's Saugus Jct. installed, with three of its four signals wired and operational. The easterly half of the crossover is Chipman St., still on the workbench being wired. Once it's in, four more mast signals remain to be built.
12-Jan-2017: Yesterday I finished installing the Chipman St. interlocking and connecting it to two of the three signals it drives. That was the last, so my signal project had reached another milestone.
SaugusJctSignal201701.png
The Saugus Jct. end of this diagram is different from the picture above; I didn't read my own documentation and so I had to invert the sense of relay R5T1. All I'd been able to get off Bexley Tower's antique key switches was a 'break' when the turnout was reversed. Because of the number of diodes involved, it was easier to shift two functions one contact set left.
BexleyTunnelSignal201701.png
Here's the final (at least through 2025) Bexley (Tunnel West, Bexley Enginehouse and D'Arcy Ave.) drawing, with wire colors, correct labels and a few obsolete ideas cleaned up.
Next I used parts on-hand to build the two 3-light home signals that will finish Saugus Jct. and Chipman St. I still don't have the brass bits for the 'bracket arms' required to finish West Lynn and River Works.
13-Jan-2017: The night before I implemented these interlockings. Chipman St. (below) is the most complicated I've built, and it's a bit of a mess, mostly because of all the diodes I needed to logically OR multiple inputs to each signal driver's RED terminal.
imgp4096_v1.jpg
The auxiliary terminals left of the top SD-3s are crowded, so are some of the main terminals and the relays' screw terminals.
ChipmanStDiodes201701.png
I count 24 diodes in this interlocking. I drew them in a matrix, and it would have been much neater to wire it like diode matrices for yard throat control. But I'd had trouble with a matrix I built from perfboard and bare wires, so I wired this directly.
Question: has anyone found a commercial circuit board (probably for prototyping) that's a good base for building a matrix? Copper leads going one way on one side, perpendicular on the other, lead and hole spacing allowing diode installation somewhere between 1/4" and 1/2" apart?
19-Jan-2017: More progress the past few days:
IMGP1049_v1.JPG
The eastward home signal for Saugus Jct. with soldering complete. Red/Green bipolar LED for the High head, fixed Red Middle head, Red/Yellow bipolar LED to provide Restricted call-on into the Bexley industrial area. Finished it this afternoon (photo in B&M Eastern Route thread), three more to go.
18-Jan-2018: My signal system has been almost complete for a year now (except for two bracket-arm masts I haven't gotten around to
scratchbuilding). It's operated reliably; the only glitches were a couple of cases where a fine 'pole line' wire wasn't being well-gripped by my bargain euro-style terminal strips. When I do a 'DC Day' and run my older equipment, I do find myself wishing I'd used only DCOD detectors, but the places where I did operate just as I wish.
18-Feb-2018: Sometimes "act now" is the right path; I should have bought the Model Shipways etched brass numbers (page 1) for automatic signals back in 2016. Now it appears Model Shipways has been swallowed up by another firm and the 2mm and 2.75mm letter/number sets aren't to be found on the website.
I can get 1/8" from miniatures.com (oversize, but maybe more legible) and 2mm from secretweaponminiatures.com. Anyone have other suggestions? Europe/UK included, as I have overseas family who would probably help.
27-Nov-2018: A few updates on my signals: Once in a while something on a contact gives an incorrect indication. All cases to date have been fixed by throwing the guilty turnout a few times, so it's usually one of the Kemtron-style switch machine contacts.
I was testing a prototype Spring Switch Controller from Azatrax:
IMGP3931_v1.JPG
This is the 'three wire' version, for use with Kemtron etc. twin coil solenoid machines. I added the outboard relays so the machine could be driven by my capacitor discharge supply. It has a rectifier but its built-in relays buzz a bit on AC. It's not on Azatrax's site today, but I understand it and the '2-wire' (Tortoise etc.) version should be available soon.
Like many other Azatrax products, it uses I/R LED and detector pairs to detect trains. I used the 'reflective sensing' mode, which was a little tricky where I had little room. It works well, but detection & throw take a moment: 2" from the 'branch' sensor to the rail gaps is OK up to maybe 40 scale MPH. I don't mind because the prototype limits speed to 25 MPH or less when pushing switchpoints over.
I had to update the Newburyport diagram to reflect this and the use of Rob Paisley's LED drivers instead of SD-3s:
NewburyportDrawSpringSwitch.png
I fixed several errors/omissions in labels too.
7-Jan-2019: I just got reminded I needed closure on getting a good 'yellow' out of a stock Paisley 556 signal driver card: LEDs I bought in 2017 from led-switch.com labeled '3mm Red/Green Diffused 2-lead' give a nice 'yellow'. led-switch is now gone. I've got to test before buying any diode signal lamps in quantity.
17-Jan-2019: I learned that Free State Systems, who made the nice etched signal ladders and platforms shown earlier in this thread, is no longer in business. No inventory remains.
24-Feb-2020: In October 2019 I prepared a Lineside Signals clinic slide-show for an NMRA Seacoast Division event. I was reviewing it and realized an illustration I'd made was worth posting here:
EasternTrackSignaled.jpg
This shows my layout's signaled main-line blocks.
3-Apr-2020: Rix hasn't had the complete CTC Knob kit the last couple of times I looked, but I ordered plates and knobs. There seems to be a spring which retains the knobs on the rotary switch shafts, but I can probably improvise that from street sweeper bristles or strapping off bundles of ties. Now that I have functioning spring switches, operations would be more realistic if I used knobs to set direction (East - Auto - West) on my two single track segments.
4-Oct-2021: At 8 PM Saturday Oct. 9 I gave my clinic about building this signal system at the Northeastern Region convention Mill City 21 in Westford, MA.
28-Oct-2021: I'd noticed a minor problem a while back: One light on a 3-head home signal never showed Clear (green), only Caution and Stop. I sat down (under the layout) last night with the diagram to diagnose it. I'd used female spade terminals for connections to the Circuitron SD-3, rather than solder, so first step was pull the Y wire off. It stayed yellow, so the problem was in the SD-3. I had another on the layout whose signal hasn't been built, so I swapped them.
Now both work. Maybe a bit of cut-off wire had fallen into it and moving it cleared it out, but a clear case of Problem Disappeared In Process of Diagnosis.
First component failure in 10 years of service: A DCOD (Chubb DC Occupancy Detector) on Block 1 would indicate occupancy on the built-in LED, but did not pull its Occupied output down. DCODs are designed for simple replacement, but I couldn't lift the connector latch without a tool. The reach was awkward, the latch broke off and as I was removing the card from the connector, one of the wires broke off its pin (no eye, no room to wrap the 18 GA power feed around the connector's pin). Couldn't re-solder the wire in place, so I disconnected all wires at the screw terminals and did the repair at the bench. The color-coded Occupied wires were easy to put back, but the non-coded wires required some trial & error. Next time I'll use craft paint to mark any wires that need it.
Showcase Miniatures ( www.showcaseminiatures.net ) is offering the old Century Foundry signal parts, plus some etched stainless parts and LEDs they apparently did in cooperation with Richmond Controls. About 50 years ago, I rigged up a Century (IIRC) dwarf with a fiber optic from a bulb under the layout, with a solenoid-actuated red/yellow gel mechanism. Anyone have more current experience?.
If you were doing this over starting from scratch now, would you continue with your discrete component approach, or would you try to do something with an Arduino and put the logic into software?
dave
I'd only use computers (with network connecting programmable nodes at each signaled location) if the era and traffic levels required CTC operation. The logic is pretty much dictated by the track arrangement; I've only had to change it twice: Both times replaced simple signal logic with something more prototypically accurate/complicated. People who build big layouts and scrap them soon after completion may have a different perspective.
I'm a wiring hack. If this were mine it would be on fire. I love those signals especially this one. Great work sir!
(https://modelersforum.com/index.php?action=dlattach;attach=103037)
James,
Way way over my head but still very impressive signal system.
Tom
Thanks, Philip and Tom. I think you overestimate the difficulty of this. But if you model RRs that never had any signals, the question is moot.
Quote from: jbvb on March 24, 2025, 03:50:37 PMThanks, Philip and Tom. I think you overestimate the difficulty of this. But if you model RRs that never had any signals, the question is moot.
That works for me :-) I have one planned signal, the ball signal controlling the SG/NG crossing, and I built that ball signal last year...
dave
Ralph's an accountant. But his 1970s Lehigh Valley needs signals, so he's done what was necessary (except learn to program in Basic AFAIK).
Illustrating two ways to attach wires to the pin connectors the used Chubb DCODs came with:
IMG_2386_v1.JPG
The plastic connector body isn't super heat-resistant; note soldering tip scar by the left pin.
IMG_E2388_v1.JPG
The DCOD's connector appears to be designed for wire-wrap. At least it worked very nicely before my wire-wrap tool disappeared. It's been gone 10 years; clearly the only way I can make it turn up is to buy a new one. Have to order online, wire-wrap is uncommon enough today that nobody I know can think of an in-person store which would sell me a wire-wrap tool.
:o
Quote from: jbvb on March 25, 2025, 08:59:57 PM...
The DCOD's connector appears to be designed for wire-wrap. At least it worked very nicely before my wire-wrap tool disappeared. It's been gone 10 years; clearly the only way I can make it turn up is to buy a new one. Have to order online, wire-wrap is uncommon enough today that nobody I know can think of an in-person store which would sell me a wire-wrap tool.
Micro-Center in Cambridge? (We had one close to us when I lived in Virginia, kinda Nerdvana. Almost as good as Fry's Electronics in California...)
dave
QuoteMicro-Center in Cambridge? (We had one close to us when I lived in Virginia, kinda Nerdvana. Almost as good as Fry's Electronics in California...
Maybe? I may be there for other reasons this Spring, but I should call them before I buy online.
Spent the afternoon working on the short circuit I was getting when the Spring Switch at the Newburyport depot was in the reversed (eastbound) position. Turned out to be a mechanical issue: the NJ International twin-coil switch machine is more than 50 years old and was getting balky. A few puffs of Kadee Greas-Em graphite powder fixed that (graphite lock lube would have worked too, my bottle was just farther away).
But while getting to that point, I documented the wiring I'd done by color, made the fascia bipolar LED indicator power feed independent of the Azatrax D2T power feed, switched the D2T from the +/- 12vDC bus to the 16 vAC power bus and generally organized things down there.
On the To-Do list goes: Apply graphite lube to the rest of the layout's switch machine solenoids; all are over the minimum age required to serve as President.
Is this what you're looking for? https://www.amazon.com/Wire-Wrap-Gauge-ELECTRONIX-EXPRESS/dp/B00BFYE0CY
dave
Close; 28-30 GA is thin for HO track current. But I could have already re-bought what I lost, for 22-24 GA, if I wanted to further enrich Mr. Bezos.
So I found one on FleaBay, due Friday (I like that billionaire a bit better).
The transient short associated with my Azatrax Spring Switch controller throwing to the "spring" position is found and fixed.
My design uses one 4 pole double-throw relay per power turnout (or crossover in one case). I installed protective diodes across the relays' coil terminals. A voltage spike is generated when you switch off a relay coil. It can eventually erode the contacts controlling the coil. I don't bother with cheapo control panel pushbuttons in panels, they're easy to diagnose & replace. But I put diode protection where I'd have to crawl under the layout to diagnose and fix.
This one diode failed in "short circuit" mode. Usually a burned up diode just stops conducting, but not here. I may have contributed to it by connecting the +/- 12VDC power backwards. Regardless, the dead diode was shorting one side of the power supply the whole time the switch was in the "spring" position, which somehow (not worth figuring out details) was making a PSX power district controller unhappy. Sometimes as many as three at once. Diode replaced at the bench, "interlocking board" reattached to the benchwork, wires put back in the screw terminals, signals do as they;re supposed to without FEEEPing.
Showcase Miniatures makes an HO B & O Signal Tower Mast Assembly: 2391. It's a good match for the straight-brace version (GRS I think?) of what the B&M called a "Bracket Arm"; a signal placed next to a multiple track route with a mast for each track. A signal bridge could have done the same job, but was more expensive. B&M rarely used signal bridges outside of bi-directionally signaled territory. If not all the tracks to the left of the Bracket Arm were signaled, short masts with fixed Lunar White (blue) heads were installed to indicate the unsignaled tracks.
The kit includes injection-molded plastic base and braced crossarms, 2 diameters of brass tube for the mast and etched brass and stainless components for the platform. The Union Switch & Signal version of Bracket arm had curved braces for the platform. Since they were also lattice girders, they'd be tough to scratchbuild.
IMGP5813_v1.JPG
Here I've cut and soldered the tubes to place the platform 22 feet above the railhead. I left the bottom extending below the base in case I needed it to attach the HO Break-Away Base: 2365. I bought them because the site is at my GE River Works, which is intense switching territory now and will get worse when I finish the GE in-plant track.
.IMGP5814_v1.JPG
This is where it will go. I cut away the platform casting's molding sprues (against the mast in these photos) before actually gluing it to the mast.
This signal was installed by Guilford at Ayer, MA sometime after 2000. Two 3-light masts each signal one track of a bi-directional main line to the right of the mast. It matches B&M-era prototypes except the platform is much higher. I think that's to keep it visible when the near track is occupied by auto racks or double stacks.