The Heybridge Railway, 1889 to 1913

[simond]

I used warm white LEDs in my toplights. I‘m not really convinced, and may convert to filament in due course. My loco shed has LEDs too, and they are definitely too blue-white for 1930, but came with rather nice conical shades, and resistors, fifty to the box, probably from China, but I can’t find the link any more.

The toplights are powered from the track, using “American” pick up, wheels shorted on one side, bogies are opposite polarity, as this avoids the drag of lots of pickups. Easy on the JLTRT plastic bodied vehicles, more hassle on brass Blacksmith vehicles. There is a rectifying diode (I run DCC) and a capacitor in the toilet space in the vestibule, the capacitor keeps the lights from flickering due to dirty pickups, wheels or track. The lights are permanently on, if I want them off, I can always isolate the carriage sidings, when I build the carriage sidings…

There are construction details & component sources on my Porth Dinllaen thread on RMWeb.

 

[magmouse]

Thanks, Simon. Working in the theatre, we used to use French enamel varnish (FEV) to tint the colour of filament light bulbs. You can also get specially made lamp dip from Rosco to do the same job - probably better. You might be able to tint the blue-white LEDs in your loco shed with some yellow FEV. You system for carriage lighting makes sense, and I may go down that route in the end. I fancy the extra control of DCC, but I may not when I have to have a decoder in every carriage and the price starts to mount... Having said that, I am not planning much coaching stock, so it might be OK.

Nick.

 

[simond]

I’ve pondered the use of varnish to tint the output of LEDs but never actually tried it - I did use stained glass “paint” for nav lights and signal glasses, though it darkens with the heat if painted direct on the bulb.

I suspect that LEDs will not work - essentially they’re single wavelength devices, so they only output one colour - an incandescent bulb outputs “all” colours so you can filter out what you don’t want, but I fear you can’t filter out what ain’t there! You’ll just make what is there less bright.

With incandescent bulbs, you can adjust the brightness by adjusting the voltage, but the colour will change together with brightness. LED brightness is controlled by restricting the current, but the colour doesn’t change.

Atb
Simon

 

[Richard Gawler]

I am glad I kept my Ladybird books.


From "Light Mirrors and Lenses", and photographed in daylight.


From "How it Works, The Camera", photographed under LED lighting.
The bottom illustration is "Looking at white light through subtractive filters".

Translucent amber paint (as painted onto headlamps years ago) makes white light warmer by reducing the intensity of the blue component. With black and white film, a yellow filter increases contrast between sky and clouds. Subtractive filters work for all sources of white light, sun bulbs and LEDs. The amber paint ought to work in a model coach, though I agree an under-run "white" grain of wheat bulb may give a better colour.

I ought to sort out coach lighting before attempting to paint any passengers.

 

[simond]

"White" LED

Spectrum of a white LED showing blue light directly emitted by the GaN-based LED (peak at about 465 nm) and the more broadband Stokes-shifted light emitted by the Ce3+:YAG phosphor, which emits at roughly 500–700 nm - from Wikipedia - Light-emitting diode - Wikipedia

The Incandescent light emits a continuous spectrum whose peak is certainly in the infra-red, and this explains the typical "light bulb" colour. My point is that the LED junction itself only emits one colour - in this case, the very sharp peak at 450 nm, which is very blue; the hump from around 525 - 625nm is generated by phosphors embedded in the plastic body, glowing in response to the LED - they are green-yellow-orange, all together with the blue, they look white-ish.

I guess you could filter out the 450nm blue which would leave you a dim yellowish light, which might suit our purposes quite well. How effectively one could filter the blue is another question. I guess experiments are called for!

It would be very good to find LEDs that look credibly like an oil (or gas, or even early yellowish incandescent) lamps. Their life is much greater than any model railway is likely to need, they are usually quite cheap, they are very small, and in general they do not run hot, at least at the likely power we would want.

 

[Richard Gawler]

Sorry Simon I did not know about the phosphors or indeed the construction of a white LED.

The light bar from my Dapol coach has five surface-mount LEDs, white ones. Right now, two have blobs of amber headlamp paint drying off. I will report back after dusk, when I can best control the ambient light in the room.

If I can obtain the effect of coach lighting by an oil lamp, I might be happy. If I model individual oil lamps I'm not sure if I will be a great deal more satisfied.

 

[michael mott]

I have a desk led lamp that changes color from a hard bluish light to a softer yellowish light, my question is how is this accomplished and could it be done to the coach lights?

Michael

 

[magmouse]

I guess you could filter out the 450nm blue which would leave you a dim yellowish light, which might suit our purposes quite well. How effectively one could filter the blue is another question. I guess experiments are called for!

Thanks for the background info. Reducing the blue spike should be possible with a yellow filter - the fact that it is a spike should mean the specific yellow colour won’t make too much difference. The colour of what you are left with is hard to predict - often ‘warm white’ LEDs have a pink tint, which is the main reason I don’t want to use them.

It would be very good to find LEDs that look credibly like an oil (or gas, or even early yellowish incandescent) lamps. Their life is much greater than any model railway is likely to need, they are usually quite cheap, they are very small, and in general they do not run hot, at least at the likely power we would want.

I am probably more than averagely fussy about the colour of light - the result of spending a big chunk of my career as a stage lighting designer and in lighting education. That’s why I am investigating the use of incandescent (filament) lamps - run at low brightness, they give the warm, continuous-spectrum colour of a flame-based source like an oil lamp. Run at lower than rated voltage, their life will be huge - you don’t have to under-volt a filament lamp by much to greatly increase its life. Heat may be an issue - the horse box I am working on has a brass roof to act as a natural heat sink, but I might need to make specific arrangements to get rid of heat in an all-plastic model. Cost-wise, my lamps cost £1.46 for 10 from Rapid, so hardly breaking the bank:

KL T1 12V 60mA Wire Ended Lamp

My biggest concern is the additional power needed from the DCC decoder - an under-volted lamp still draws around 50mA, and most decoders only produce 100-150mA on their function outputs. I will be investigating using a simple circuit to increase the available current, for coaches with several lamps.

Right now, two have blobs of amber headlamp paint drying off. I will report back after dusk, when I can best control the ambient light in the room.

Richard - I’m looking forward to seeing what your experiment produces.

I have a desk led lamp that changes color from a hard bluish light to a softer yellowish light, my question is how is this accomplished and could it be done to the coach lights?

Michael - most likely your lamp has two sets of LEDs, one with a low colour temperature, similar to the warm light output of a traditional filament lamps, and a second with LEDs producing a higher colour temperature, cold white light, similar to daylight. The two sources will be mixed to vary the balance of the light output from warm to cool. I have a similar lamp with specially designed LEDs that produce accurate colour rendering of objects illuminated by the lamp. It’s intended for colour-critical work such as photography.

The difficulty with this approach in a model is that you need the light fitting to mix the two sources together, so you don’t get multi-coloured shadows. That takes up space, which we generally don’t have.

Nick.

 

[Yorkshire Dave]

I’ve pondered the use of varnish to tint the output of LEDs but never actually tried it - I did use stained glass “paint” for nav lights and signal glasses, though it darkens with the heat if painted direct on the bulb.

I suspect that LEDs will not work - essentially they’re single wavelength devices, so they only output one colour - an incandescent bulb outputs “all” colours so you can filter out what you don’t want, but I fear you can’t filter out what ain’t there! You’ll just make what is there less bright.

With incandescent bulbs, you can adjust the brightness by adjusting the voltage, but the colour will change together with brightness. LED brightness is controlled by restricting the current, but the colour doesn’t change.

I use both yellow and red stained glass paint on warm white LEDs to get closer to the yellow/orange incandescent effect, and throw an additional resistor into or use a higher value resistor in the circuit to restrict the current.

The only thing with using incandescent bulbs in enclosed plastic models is making sure the heat generated can escape otherwise distortion may/will occur to the roof.

 

[Richard Gawler]

Here are some real oil lamps, photographed using the camera I use for my model shots. These have been converted from camera RAW to jpg using an iMac, but there is no deliberate editing here.







At a glance, I think we have got to use incandescent bulbs to get anywhere near this but I will post some photos of LEDs later.

 

[ChrisBr]

Question from someone who knows nothing about the subject (but is very interested in the conversation) would the type of oil used impact the colour of the flame (to any significant effect) and if so do we have any understanding of the oil used then and now?

 

[magmouse]

Thanks for posting these, Richard - it’s always good to have a real-world reference point. One thing we need to be careful of is how the camera and/or the processing software decided to set the colour temperature as it turned the raw data captured by the camera into a human-viewable image. Did you make deliberate choices about that, or leave it to the software to deal with?

The reason I ask is that the ‘hot-spots’ in the images where the flame is vary in colour between photos. This may be because they are over-exposed in some cases, and over-exposed areas in digital photos tend towards white, irrespective of the original colour. Or, it may be the processing software is seeking a good colour balance, and making the picture with a lot of blue objects/background warmer than the pictures with more red content.

Question from someone who knows nothing about the subject (but is very interested in the conversation) would the type of oil used impact the colour of the flame (to any significant effect) and if so do we have any understanding of the oil used then and now?

Probably, Chris, yes. The fuel will have some effect on the temperature of the flame and so the colour (the visible part of the flame comprises soot particles heated to white heat so they glow). I don’t know too much about oil lamps, but the folks I know involved in historic reconstruction of stage lighting with candles have gone to some lengths to recreate authentic candles using original methods and materials.

Whale oil was preferred in the 19th century for lighting, and some other applications: Whale oil - Wikipedia. I don’t know though if this included coach lighting, or if it is different in light output to other fuels in any way that would be significant from a modelling point of view.

Nick.

 

[ICH60]

The temperature in Kelvins of the light , gives you the colour


Sadly LED's only go down to 2700K.

 

[Richard Gawler]

Here are three photos of the lighting in my Dapol Stroudley coach taken today. The two LEDs in the guard's compartment I have overpainted with Granville 'Amber Light' paint, the three in the passenger part of the coach are unmodified.

All of these photos are in my hobby room with no artificial light except the coach lighting itself.


An underexposed shot taken in broad daylight this afternoon.


A better-exposed shot about 30 minutes before sunset. I think the flare of the LEDs is coming from the glazing material.


A final photo about 15 minutes before sunset. There is quite a difference in the appearance of the interior colours; the model is finished in the same colours throughout.

I offer these photos to show the difference between the effects of bare and painted LEDs. Edit: The white balance in all three of these photos is from my own efforts to reduce a strong blue cast which has came from using the 'auto' setting in the camera.

To be honest, I have a feeling the moulded three-dimensional glazing rules out any chance of successful low-light photography with interior lighting. Edit: subsequent painting of the glazing proved this concern to be unfounded.

 

[Richard Gawler]

Thanks for posting these, Richard - it’s always good to have a real-world reference point. One thing we need to be careful of is how the camera and/or the processing software decided to set the colour temperature as it turned the raw data captured by the camera into a human-viewable image. Did you make deliberate choices about that, or leave it to the software to deal with?

The camera WB is always set to automatic. I suspect the iMac has some fairly advanced settings to alter colour space and what have you, but I've never altered there anything from its default settings.

All of my photos of coach ligthing here are "unmodified by me". As far as I know, all that has happened is the iMac software called Photos (the replacement for iPhoto) has used the camera RAW file to create a jpg, which I have uploaded here.

 

[ICH60]

The main issue is that LEDs do not produce enough energy in the red end of the spectrum, which the old incandescent lamps produce



You maybe able to filter out anything below say 625nm of a 2700K led which would make it warmer but low light output but that maybe what you need if you are producing gas lighting. Sadly I never seen a spectrum from a gas light so difficult to say. I fact I not sure I have ever seen as gas light to say what colour it was. Was it more white than we think? just of lower intensity? Oil light which will be the lighting in my coaches I have oil lamp so I can work that out by the old fashion method of looking at it Though what oil they used andwill they effect the colour

 

[ChrisBr]

This is the nearest to a spectral response for an oil lamp I could find courtesy of Google -

which supports Ian's assertion we need more reds to mimic the light...

 

[Yorkshire Dave]

To be honest, I have a feeling the moulded three-dimensional glazing rules out any chance of successful low-light photography with interior lighting.

It's unfortunate they've moulded them as shallow trays. They would have been better left as solid panes in Roco HO fashion.

I noted earlier you painted the inside of the glazing to match the interior of the carriage sides - it may be worth painting the outside face of the moulding a brown or mahogany colour including the edges of the panes to see the effect. If you use acrylics this can be taken off by just rubbing with a cotton wool bud if it does not work.

I have a street car with similar glazing and just knocked up this in 15 minutes (nothing scientific) and photographed under a 60w daylight bulb to see the effect. I think it will be less noticable with 3V dimmed LEDs.

Car interior



Car Exterior. I painted in between and the edges of the panes with mahogany, car body colour (Santa Fe red) and a black Sharpie permanent marker.



And what it looks like inside the car held under the workbench 60w daylight bulb. I think it can be disguised by weathering the panes and cleaning the centres leaving the edges dusty.

 

[simond]

Richard,

I think your amber varnish has transformed a very obvious LED into something that might pass for incandescent, so that’s a win in my book. I shall experiment with some yellow / amber stained glass varnish.

Thanks
Simon

 

[magmouse]

All of my photos of coach ligthing here are "unmodified by me". As far as I know, all that has happened is the iMac software called Photos (the replacement for iPhoto) has used the camera RAW file to create a jpg, which I have uploaded here.

Although you haven’t specifically set anything, Photos will have had to make a decision about the colour temperature of the lighting of the scene when interpreting the RAW file. Usually it will read what the camera decided from data embedded in the RAW file and use that. One way or another, a decision has to be made on how to interpret the RAW data and turn it back into colour values - whether that decision is made by the designers of the camera, the software or you, the user.

The main issue is that LEDs do not produce enough energy in the red end of the spectrum, which the old incandescent lamps produce

Yes, indeed. A further nuance in this discussion (as if there weren’t enough factors in play already) is that there is a difference between the colour a light source appears to be (e.g. “warm white” or “cool white”) and how the light from the source renders colours. The warm white LED may look very similar to an incandescent lamp when looking directly at the source, but the lack of the really long wavelength reds in the LED light may mean red materials (e.g. the seat covering in a carriage compartment) lose their richness of colour.

Sadly I never seen a spectrum from a gas light so difficult to say. I fact I not sure I have ever seen as gas light to say what colour it was.

As it happens, I was in Prague a couple of months ago and saw incandescent mantle gas lamps on the Charles Bridge:



The photo doesn’t reproduce the colour perfectly, but my sense was it is a whiter white than a filament lamp, with a slight acid yellow tint. In terms of coach lighting, it is important to distinguish gas with an incandescent mantle from the earlier type that just had a naked flame, and would have been similar in colour to oil - certainly a bit brighter and perhaps whiter.

Nick.