More on Layout Lighting
I’ve been thinking about a number of things related to the layout this month, but mostly about lighting the layout itself. The current layout is lit by a mix of my original track lighting system (using compact fluorescent bulbs) and the newer fluorescent tube valences. The photo above shows the old track lighting in use, with the first valence installed in front of it, but not yet wired up.
That experience convinced me of the merits of fluorescent tube lighting. It also convinced me of the need to build the lighting valence as part of the benchwork, rather than trying to suspend it from an irregular ceiling. As you can see above, the heating ducts caused some difficulty in attaching the lighting units in this part of the basement.
When I built those valences, a couple of years ago, I used 4’ T8 tubes. T8 is the modern replacement for the older, larger, T12 tube, and probably the best choice today for layout lighting (you can also get them in 2’ and 8’ lengths). There’s an even smaller T5 tube, available in lengths up to 36”, which may be preferable for space-constrained multi-deck layouts. But T5 tubes and fixtures cost more, and are harder to find in high-CRI form.
The Quality of Light
And that “color quality” is the most important thing about fluorescents. The older ones had serious problems with the color, creating a greenish cast on photographs and looking rather poor in person. Those often had a Color Rendering Index (CRI) less than 70 on a 0-100 scale. To get good color, you need a CRI above 80 (and maybe even higher). Fluorescents today are easily available with CRIs of 82 - 85, and some are available around 92. For comparison, incandescent and halogen bulbs always have a CRI of 95 - 100, but the heat produced by enough of them to light a layout would be a big problem. I’ve found CRI 82 lights to be fine, but it’s another one of those subjective things and you may, quite literally, see things differently.
Note: if a fluorescent light (tube or bulb) doesn’t have a CRI rating, it’s going to be below 80. Incandescent and halogen bulbs, on the other hand, never have a CRI rating listed on packaging because they’re all above 95 by definition.
Another issue to consider with fluorescents is the type of “ballast” used inside a fixture. In a fluorescent light, the ballast is the circuitry used to drive the light itself. In older lights these were relatively simple. Modern electronic ballasts designed for high-CRI lighting shift the frequency of the electricity from the normal 60 Hz used for utility power in North America to frequencies in the kilohertz range. This prevents flicker, and also results in more even lighting. So it’s important to use a fixture with an electronic ballast. This is likely to be the default in a T8 or T5 fixture anyway, but it’s something to verify before buying a lot of apparently cheap fixtures. If you have to replace the ballast, that gets expensive.
Because electronic ballasts work at kilohertz frequencies, and fluorescent bulbs do emit some Infra-Red (IR) light, they can interfere with IR remote controls (like some hand-held walkaround throttles). If you use something like that, best to test it with a fluorescent of the type you’re considering before investing too much.
Another problem with fluorescents is that, in general, they aren’t dimmable. There are dimmable ballasts for tube fixtures, but they’re very expensive if you can find them at all, and still rather problematic (you need to replace all bulbs on a circuit at the same time, for example). Compact Fluorescent Light bulbs (CFLs) can be found with dimmable ballasts (in a CFL, the ballast is built into the bulb), but I’ve had problems locating any dimmable ones with high CRI ratings. If you use fluorescents you’ll need to accept lighting being “on” or “off”, without any option for a “twilight” setting.
After my experience with CFLs in my track lighting, I’m reluctant to use them again. It’s hard to get floodlight-form CFLs with good CRIs. And placed as close to the layout as track lights need to be, even “flood” CFLs tend to produce bright spots unless you use a lot of them.
However, even if you place fluorescent tubes behind a valence to light the layout, you need some lights for the rest of the room. This is where down-pointing flood lights may be the best solution. Some people use fluorescent tube ceiling fixtures for this, but to my mind this makes the room itself too brightly lit.
If you’re using larger “can” ceiling fixtures (e.g., PAR38 bulbs in large-diameter fixtures), then you can get fairly bright CFL bulbs with good CRI levels and a variety of “colors”, although you’ll likely have to order them online; selection at my local home stores is very limited. My problem was using smaller bulbs (PAR20), which have less selection available. And the benefit here is that as LED prices come down, those CFLs can be replaced with high-CRI LED bulbs. Something that isn’t true of fluorescent tubes.
There are some LED “panel” lights, long strips in widths of 12” or 24”, that may someday replace fluorescent tube fixtures. However today they cost more than five times the cost of an equivalent fluorescent fixture, and don’t appear to come in CRI-rated versions. And you’ll need to replace the entire fixture when the light fails, as the current ones don’t use a separate “light”. Fluorescent fixtures are cheap enough that if I did someday replace them with LED panel fixtures, it wouldn’t waste too much money.
So my current thinking is PAR38 CFL floods in can fixtures for room lighting, and single T8 tube fixtures for valence lighting. That’s a rough plan and still subject to change; if things improve quickly enough I could use LEDs in the can fixtures.
There is one potential flaw: I’ve read that heat in recessed can fixtures tends to make CFLs, even CFLs supposedly designed for them, fail early. Some high quality ones (Phillips) are less prone to this. I’m concerned that LEDs could have the same problem, as I’ve had early failure issues with LED bulbs in smaller airflow-constrained fixtures I use for office lighting. This makes me less likely to use LED lights in can fixtures while they still cost significantly more than CFLs and haven’t had enough time in use to be proven.
The Color of Light
I mentioned “color” above. You may recall that when I first designed the track lighting for Sumida Crossing, I looked into replacing halogen bulbs (which I’d used on the old HO layout) with CFLs, and part of that was deciding what color of light to choose. Halogen lights are typically around 3000K, which is slightly “bluer” than an incandescent bulb (which may be around 2700K). And tungsten photo floods are around 3500K. All of these tend to be generically termed “warm white”, although there is some visible difference between them.
I tested a number of different bulbs, with different colors, including some around 5000K or 6000K (both are called “daylight”, although sometimes 5000K is separately called “natural”). I found 5000K too blue, and ended up settling around 4000K, partly due to the UV filter I used with my CFLs, which shifted the light from around 3800K to 4100K. And when I chose fluorescent tubes, I used a “cool white” tube rated as 4100K.
I’ve found 4100K to be a good color, and plan to use it on my next layout. I’m not alone in this, I’ve read of a number of others who use cool white fluorescents. Although some people do prefer bluer “natural” or redder “warm white” lighting. It’s one of those subjective things that has no “right” answer.
One benefit of using Cool White for layout daylight lighting is that most “white” LEDs on buildings will look a little blue, as they should for fluorescent or very bright modern headlights, and “warm white” LEDs can be used where there’s a need to replicate redder incandescent lights, as those will look “redder” in comparison to cool white.
The Ultraviolet Question
One concern I have with both fluorescent and LED lighting is Ultraviolet (UV) light. This is what causes colors to fade, which will affect both my models and the photographic backdrops I use. Both fluorescent and LED lights use UV internally, but should limit the amount emitted to relatively low levels. Unfortunately “should” is a word that’s been violated a number of times in the past.
What prevents the UV from getting out is the phosphor coating the inside of the bulb, which absorbs UV and emits visible light. If the phosphor doesn’t cover the whole of the inside (a problem in older fluorescent tubes at the ends) then UV will come out. Some CFLs were also designed in a manner that left UV leak, although ones with a frosted surface generally did not. High-CRI LEDs basically work on the same principle, with UV used to energize phosphors providing a mix of wavelengths in the visible range.
I might be able to count on modern lights doing a good job of containing UV, since any loss is inefficient (power wasted not producing visible light). With the current emphasis on energy-efficient lighting, manufacturers have an incentive to avoid that kind of leakage, and manufacturing techniques have definitely improved from past decades so its less likely to occur by accident. But one option is to place a transparent or translucent plastic diffuser panel in front of the light. Plastic will absorb any UV, at the cost of reducing the intensity of the visible light somewhat. It will also make the light less directional, and that’s probably good in terms of evenly lighting the layout (at least for valence lighting). I’m still on the fence about this issue though.
Summary
But to sum things up: after giving it a fair bit of consideration, I have a pretty good idea of what I want to use for layout lighting. It’s going to require a bit of work, between valence construction and can lights for the room. But I’ve tried to cheap-out on lighting a couple of times in the past, and have been disappointed with the results. It’s time to accept that this is an area that deserves some time and expense up front.