Proof of concept for Model Train Museum

[TenBall]

I am a total newbie to all of this but not to LED lighting, as I work for a company that produces trade show booths from 100 sq ft to over 5000 sq ft. I've been watching various YouTube videos for the past few weeks and the more I watch the more questions that are coming up. I also see the potential to use this style of lighting for trade shows.

Let me start at the beginning. I'm a member of a model train museum in Tucson, AZ. The main building is 60' x 100' currently lit with 4' fluorescent tube fixtures. After several of our members have toured other model train clubs and museums we would like to incorporate some of the things they have seen. The idea is to remove the existing fluorescent fixtures and replace with RGBW pixels. The ballasts are becoming more difficult to find and if we have to replace the fixtures we might as well do something cool.

First I need to sell the board of directors on the basic concept. Due to the size of the project it will need to be done in several stages.

Stage one/Proof of concept: We want to be able to simulate going from dusk to dawn and back to dusk again in a time frame of 15 to 20 minutes. On occasion we would like to add in a thunderstorm. For special events like birthdays we would like to have it say "Happy Birthday person's name" maybe with a train graphic or animation. As stated before the ceiling is 60'x100' and am planning on using matrix panels 4'x8' with a pixel count of 288 per panel. Total number of panels is 168, total number of pixels is 48,384 or 968 strings of 50 RGBW pixels. Strings will be 5V WS2811 RGBW pixels. I want to use 5V strings to conserve energy/power consumption. The reason for RGBW is that when the museum isn't open to the public and leaning or work is being done in the museum we want to be able to use the single white chip to save power consumption. I know this will require more power insertion, but worth the energy savings.

Questions for stage one/Proof of concept. What controller(s) should I use to handle this size of matrix with the number of panels? From my research a Falcon F16V3 with 2 expansion boards should be able to handle the number of pixels, right? Or should I get a F48 differential controller and several 4 string differential smart receivers? I want to purchase the primary controller to be able to show the board of directors what can be done and not have to spend money on different hardware if I get the green light on the project. Can a low voltage switch be installed so that with the touch of a button the pixels switch from the dusk to dawn pattern to white lights? Or to the special "birthday sequence?" Is there anything I have missed?

Stage two: We would like to be able to add some special effects. One special effect would be is someone takes a photo with a flash that an optical sensor picks it up and inserts a bolt of lightning across the "sky" and then an audio file of thunder plays. One of the clubs did this with an Arduino, but was connected to a few white LEDs that pulsed quickly for the lightning.

Stage two questions: Can this be done with Xlights?

Stage three: Incorporation with a home automation system. (I know I'm probably dreaming) We have a number of members that are 60-70+ years old and don't always do well with technology or simply forget to turn off the light, so it would be nice to be able to turn them off remotely when the property is viewed via the security cameras.

Stage 4+: Increasing the number of pixels in the matrix, adding additional DMX instruments, additional items outdoors for holiday season while riding narrow gauge train around property. (By this point I should have a good grip on the hobby and what would be needed.

I understand this is a terrible time to be asking all these questions and for help but the museum is closed due to covid and not sure hen we will be reopening. I want to make the proposal while it is closed and easier do the work.

Thank you for all your help in advance.

Russ aka TenBall

[keithsw1111]

I don?t see anything there that could not be done.

[TenBall]

I don?t see anything there that could not be done.

I'm glad to hear that my crazy ideas could all be done.

Any suggestions of controller to use to handle the number of pixels? F16V3 or the F48 Differential What are the pros and cons of each?

[keithsw1111]

The F16s wire all strings back to the controller. The F48 lets you run cat5 cable to receivers which can then drive 4 strings. This gives you more ports but not more pixel driving. It can however simplify power management as you power through each of the receivers. Without the specifics it really is impossible to say which would be best. Generally people like the F16 where all the strings are close (say within 5m) of the controller or F48 where they are more distributed.

[dkulp]

I do want to point out that a single F16v3 or F48 can only drive 16,384 pixels.   I also believe that assumption is RGB pixels (so 48K channels).       If using RGBW pixels, the channel count is obviously 33% higher and thus the pixel count is lower. You'll need 4 F16's or F48's to drive your full display. 

[Gilrock]

Hey I'm in Tucson and visited the museum years ago.  I would do anything I could to avoid RGBW it just doesn't work that great in xLights without extra work.  I've never looked does it really save power?  I would think turning on the red, green, and blue at 33% would be nearly equal to white at 100% in power draw and light output although the pure white would probably look nicer.  Maybe alternate lights could be mounted for when only white is needed.  I've helped folks out locally some with advice and some I've gone to their displays so I might could help but not until after Dec 7th....going up to the mountains tomorrow.

[TenBall]

The F16s wire all strings back to the controller. The F48 lets you run cat5 cable to receivers which can then drive 4 strings. This gives you more ports but not more pixel driving. It can however simplify power management as you power through each of the receivers. Without the specifics it really is impossible to say which would be best. Generally people like the F16 where all the strings are close (say within 5m) of the controller or F48 where they are more distributed.

My thinking is to have a matrix of panels with the panels 8' (2.4m) parallel down the 100' (30.4m) so that the matrix would be 12x14 panels.

Because of the distance between the controller and the strings I guess it makes more sense to use the F48 and place them in the corners of 4 panels (strings) that would place 6 differential controllers at the intersection corners of 4 panels goring down the longest dimension for a total of 6 and 7 down the short dimension for a grand total of 42 differential receivers, that would require four F48 controllers. CAT5 runs would be roughly 8' (2.4m), 24' (7.3m), 40' (12m) between controllers and receivers.

[TenBall]

Hey I'm in Tucson and visited the museum years ago.  I would do anything I could to avoid RGBW it just doesn't work that great in xLights without extra work.  I've never looked does it really save power?  I would think turning on the red, green, and blue at 33% would be nearly equal to white at 100% in power draw and light output although the pure white would probably look nicer.  Maybe alternate lights could be mounted for when only white is needed.  I've helped folks out locally some with advice and some I've gone to their displays so I might could help but not until after Dec 7th....going up to the mountains tomorrow.

Gilrock,
I will send you a private message and maybe we cam meet in person after you return to town.

[jnealand]

That sure seems like an awful lot of lights in that ceiling.  I have lots less than that and I can light up the neighbors yard to be too bright.  Just sayin'.