A thread from RR-Line about how I built the (relatively) prototypic dwarf signals on my B&M Eastern Route layout:
9-May-2013: My B&M Eastern Route layout (https://modelersforum.com/index.php?topic=6843.0) (https://modelersforum.com/index.php?topic=6843.0) needed a half-dozen dwarf signals for locations like entering single track from double against the direction of traffic, yard exits etc. Post-WWII the B&M generally used SA-style single head dwarfs. There was a cast metal kit (originally Century's), but I had been looking at NJ International's 129x dwarfs and
bought a pair (1295s) (2025: NJ International hasn't had inventory of these for several years). The problem is, their LEDs are single color, and I want red/green. The factory LED didn't come out with just a little tugging.
Lou (Desertdrover on RR-Line) replied: James, If you have craftsman in your blood, make your own dwarf signals from
styrene; http://www.nyx.net/~jpurbric/railroads/dwarfs (http://www.nyx.net/~jpurbric/railroads/dwarfs)
Interesting, Louis: John Purbrick is an old friend from my TMRC days. That method will certainly work, I was just hoping I could substitute money for time on this sub-project, so I'd have more time for the signal controls...
[edit] And no criticism of NJI; I knew I was buying single-color LEDs, the packaging was clear enough
29-May-2015: I returned to this topic, as I'd started my Lineside Signal project: https://modelersforum.com/index.php?topic=6882.0 (https://modelersforum.com/index.php?topic=6882.0)
Closure on this; last night I dunked an NJI #1295 "Dwarf Signal - SA Type" in acetone:
IMGP4089_v1.JPG
The paint came off in less than 30 seconds, revealing a nice 2-part investment cast brass head. A little gentle prying and pulling on the LED with needlenose pliers re-kitted it. This morning's project is to ID the LED package used and see if I can get a bi-color to replace the original. This may take a while as none of the first two Google pages on "led package type" show this, though I've seen it sold. Of course I didn't get anywhere till I went to specialists...
1-Jun-2015: More closure: I got red/green 1.8 mm LEDs that fit the lens opening and housing at led-switch dot com, which has vanished. So at the moment, there's a slog through Digi-Key or a similar site, then testing to see if their dies are close enough together.
IMGP4110_1.JPG
30 ga. Kynar wire soldered to the stubs of the cut-off leads. I'd have needed magnification to do this even when I was 18. But now I feel confident about making street lamps using Microlumina or Ngineering surface mount kits.
In response to a question: It's a trade name for this: http://en.wikipedia.org/wiki/Polyvinylidene_fluoride If you've ever done wire wrap (I've both configured computers and wired parts of my layout with it), it's standard for that. It's a thin, tough insulation that stands up to being threaded through rough openings in brass tube fairly well.
16-Jun-2015: This is what an NJI #1296 red/green 2-light dwarf looks like after an acetone soak and a little wiggling:
IMGP4124_v1.JPG
In B&M interlockings, these were equivalent to the bottom two searchlight heads on a 3-head mast home signal. For my application, I'll try filing another of the 1.8 mm red/green LEDs down to fit in place of the green (upper). I'll leave the red (lower) LED soldered in place.
MarkF asked about PRR position-light dwarf signals. I had found Integrated Signal Systems earlier:
http://www.integratedsignalsystems.com/ho-scale-parts-2/ (http://www.integratedsignalsystems.com/ho-scale-parts-2/)
27-Jun-2015: After burning up a number of 1.8 mm red/green LEDs trying to get them into the NJI single heads, I pondered a bit. My relay-based signaling scheme wants bi-color LEDs but I can simplify the aspect logic by limiting the single head dwarfs to red/yellow (Stop/Call On) and not making any promises about occupancy as you enter signaled track. You phoned thetowerman to get the switch thrown in the first place, didn't you?
I hadn't found 1.8 mm LEDs in red/yellow, but several parts vendors sell 3 mm (T-1) in modelers' quantities. So I tried building a couple of styrene dwarfs per John Purbrick's directions that Louis linked to:
IMGP4454_v1.JPG
The left signal (base installed) was built per John's directions. The right only has one layer of .020 styrene on the housing. Aside from me thinking it looks better, applying the 2nd layer of styrene was problematic: If I got solvent cement on the bent part of the 1st, it broke. I had 5 failures to 1 success, but if I make another attempt, I'll try tube model airplane cement for the 2nd piece.
9-Jul-2015: My wiring error burned up a pair of these dwarfs, but because I hadn't followed John P's directions exactly I was able to get them back on-line in about an hour. I didn't glue the LEDs into the heads; instead I relied on the thick coats of black, then silver paint to keep the press-fit from coming apart. Also I'd used little terminal strips to connect them.
IMGP4475_1.JPG
The 30 gauge wire is pretty fine for this kind of strip, but with a little fiddling I could get the screw to grip. So after I was done swearing about letting the smoke out, I undid the wires, pushed up on the square styrene till I could pull the signal out from above and twisted the LED gently by the remains of its leads till it came loose.
[After almost 15 years of living with cheap Euro-style terminal strips, I've resolved to never buy the kind where a screw is expected to catch the wire against a brass piece. I need to find the kind where the screw pushed one brass piece against another, to clamp wires that are anywhere between them]
18-Feb-2016: My signaling work reached the part of the layout that needs 2-light searchlight dwarf signals:
IMGP7460_v1.JPG
These are 1.8 mm red/green bicolor LEDs from LED-Switch (regrettably gone - experiment to ensure the red & green dies are close enough to make yellow before gluing it all together). The left one has been filed down to fit the NJI 1296 head. I've done two and haven't wrecked either junction with the filing (unlike with the single-light dwarfs last year). Alas, I haven't yet found similar-sized red/yellow LEDs.
IMGP7463_v2.JPG
Here's one installed. After soldering the #30 wire as close to the LED body as possible without shorting the leads, I bent the leads over to opposite sides and cut/filed the excess off.
I'll have four wires coming down from the head instead of NJI's three. Once I see how it looks painted and installed, I might try the next with #38 magnet wire. But I might regret signing up to push that through a hole in the plywood and then somehow terminate it below.
17-Mar-2016: A couple of as-installed shots of the NJI #1296 two-light searchlight dwarf I added bi-polar LEDs to:
IMGP7517.JPG
As presently wired, the top head can display Red/Yellow/Green because it's driven by a Circuitron SD-3. The bottom governs entry to Bexley Engine Terminal, so it's driven by contacts on the switch lever and only displays Red/Green. The base is filed from .040 styrene.
IMGP7521.JPG
My four #30 wires are more conspicuous than NJI's original three. I'm thinking more about #38, at least as far as something intended to look like a junction box next to the signal on the roadbed surface.
The laserboard 'joint bar' is painted yellow as a 'gap marker' for operators.
19-Mar-2016: Dave Silvernail commented:
QuoteHere is a little trick to pushing wires through plywood base. Find the smallest brass tube that the wires into, even if it is tight, drill a hole through the base and feed the tube through pulling the wire with it. I use this trick on modules that have 2" of foam with plywood underneath. I have even used this trick on a
10" foam base under a mining scene on a friend's module.
That's the end of the thread as of Feb. 2025.