EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: pilotplater on March 31, 2019, 02:07:20 am
-
HI!
I am a pilot/flight instructor/hobbyist, and one of the things I do semi-regularly is take people flying who want to propose to their significant others. I want to design and build an LED sign that would fill a ~5 acre area. Resolution doesn't have to be great, I'm thinking have it display at least the text "will you marry me?", and ideally have it controlled centrally so the text could be customized with someone's name, and animated.
One of my first thoughts was battery operated modules using esp8266, but it would have to do some fancy ad-hoc connections to act as a shift register to send all the info out, as there's no position that will likely be in range of every receiver.
May end up going super simple route and just purpose-build the pattern to say "will you marry me?", and forget the dot matrix alltogether, but if anyone has some really cheap/reliable solution to interconnect hundreds of modules over a ~400m distance that'd be helpful. I don't really want to untangle a mile of TTL cable ;)
-
Some sort of high power transmitter and a bunch of receivers would probably be the simplest solution. Probably 900MHz for ready made modules, or FSK over 27MHz CB radio. You could even put a directional antenna on the transmitter and transmit packets multiple times while sweeping the antenna.
Another solution is to have a lot of transceivers in a mesh, programmed to repeat packets to its neighbors. Some careful analysis would be needed to make a repeating algorithm that works without getting stuck in a loop, involving stuff like limiting the number of time to repeat each packet and consideration of timing to avoid collisions.
-
Rather than full dot matrix, you could just determine the ones you need to use for the animation and make those - like animated neon signs. Should cut down on the number required by quite a bit. The name customization would be tough then, though.
It sounds like you're talking about at least a hundred independent devices, and while I agree wireless is probably the right route, power is going to be tricky and setup is going to take a while too.... and given the time constraints and requirement for reuse, I think weatherproofing and reliability will be a big issue. It may be worth adding a light beacon to a commercial, ruggedized, solution, just so you have a good option for keeping it powered, controlled, and weather-resistant.
If you were to design yourself, maybe a mesh network style system would be a better choice than regular wifi, but I still think making a setup you can use reliability for even a year or two is going to be a big technical challenge and is going to involve a lot of gathering up modules and replacing batteries.
-
Perhaps a lora modem and solar panels for power with rechargable batteries for night time (that charge during the day)?
Sent from my SM-G960F using Tapatalk
-
DaJMasta - Definitely interested in commercial solutions, do you have one in mind for a similar purpose?
-
To make it visible from the plain, you'll need lots of power. Since it's a grid, getting power/signal from the nearest node with 3-5m cable laid in conduit is much cheaper than any sort of battery/solar/wireless solution.
-
Yeah I agree that cable may be the way to go. If I'm running a cable, there's no point in any of the nodes being terribly intelligent. Could possibly still do some cheap micro in each one to act as a shift register for that node.
Not looking for a permanent installation on the first try, which is why the cable/conduit is an issue, but I think aligning everything on the ground may be more difficult in any case.
-
Make sure to remotely control the sign from the plane to turn it to a Sad face in case she says No
::)
-
Make sure to remotely control the sign from the plane to turn it to a Sad face in case she says No
::)
hahahaha
-
Hmmm, I had assumed this would be a night display and that could be battery powered, but I guess the certifications are probably different, so you probably want daytime use. In that case, just do it with cables, and since the cables are for power delivery, you just need a controller at each power distribution node and the display dots can be 'dumb'.
For example, if you have 10 nodes over your 5 acres so that you can keep any individual node to light run under 100 feet (if that's how the space works out), then each node can control them individually (a big pain), or can control a few channels worth of lights, with each channel turning on the ones required for an individual animation frame using a relay or something. I'd drive the light "pixels" with a reasonable AC voltage (maybe 24-48VAC) to keep the cable losses to a minimum, but then with 10 main nodes or so, you can have larger power sources for each, or you can connect them up to some grid power source.
With such a setup, you could still use your preferred wireless tech with the base stations, but you could probably program them and just let them run since it would only be 10 units worth, provided you had a synchronization mechanism to keep the animations from drifting apart. The base station would probably be custom, but then I'd look for a commercial waterproof light with a spike to drive it into the ground or something, and a similar outdoor rated cable with waterproof connectors. That way the connectors and lights are all done already, with just the base stations to worry about. Once you know specifics of the lights you're using (and it's probably worth testing them by hooking them up and flying by to see if they're reasonably visible), you can determine acceptable cable losses, cable expense, and site dimensions to tell you how many nodes would be needed.
Still a large scale fairly complicated project, but I think centralized control would be a lot simpler unless you can use low power (only nighttime operation) and need individually configurable pixels.
-
Why don't you just place a projector somewhere up high and then correct for sideways projection (since it will be looking at an angle towards the field). I think there are projectors that use some form of laser light that can produce a sharp picture from a distance. At least it works on buildings. However for your application I'm thinking a projector mounted up high like on a cell tower or a tall building. Or perhaps on your plane?
Sent from my SM-G960F using Tapatalk
-
Why don't you just place a projector somewhere up high and then correct for sideways projection (since it will be looking at an angle towards the field). I think there are projectors that use some form of laser light that can produce a sharp picture from a distance. At least it works on buildings. However for your application I'm thinking a projector mounted up high like on a cell tower or a tall building. Or perhaps on your plane?
Sent from my SM-G960F using Tapatalk
https://www.google.com/search?q=projector+from+building+to+ground&safe=off&client=ms-android-samsung-gs-rev1&prmd=ivsn&source=lnms&tbm=isch&sa=X&ved=2ahUKEwi_jdGml63hAhVhoYsKHXf_BlAQ_AUoAXoECGUQAQ&biw=360&bih=612
Sent from my SM-G960F using Tapatalk
-
Why don't you just place a projector somewhere up high and then correct for sideways projection (since it will be looking at an angle towards the field). I think there are projectors that use some form of laser light that can produce a sharp picture from a distance. At least it works on buildings. However for your application I'm thinking a projector mounted up high like on a cell tower or a tall building. Or perhaps on your plane?
Sent from my SM-G960F using Tapatalk
Been there done that, at night.
In daylight with today's technology it would take stupid amount's of power and $ to project anything readable.
You have better chances with even the cheapest (coarse pixel) Chinese video walls.
-
is some sort of acrylic letters with backlit would be a cheapest , i think a single panel of acrylic 1.5 mm 260 cm x 80cm around $40-60
static letters, nothing fancy
-
There's always the old-school way. Gather up a lot of the victims friends and family and give them all contrasting poster-size boards and some same-day training from some guy with a megaphone who has comms with you. Even if a few of the people screw up their pixels, it should still work.
-
The conditions in which it needs to be visible will really determine the feasibility here. Maximum solar illuminance can be up to 120,000 lux, and since five acres is about 20,000 m^2, that means you have about 2.4 gigalumens to compete with over the entire area. Making the background dark will certainly help provide contrast, but even with a 1% reflective material, that only takes you down to 24 megalumens. You can reduce that further by only targeting a limited viewing angle, and using lenses and/or reflectors to send your light in that direction, but remember that lenses and reflectors work both ways, so you'll need to take care to make sure that sunlight isn't focused down into an LED-death ray inside your contraption.
Obviously, something that only needs to be visible in the dark is going to be much easier, and at that point it's mainly a question of what sort of visible size each pixel needs. Maybe you only need a couple of PAR lamps every meter or so, or maybe you need something that illuminates the entire ~square meter area. You probably will need to do some tests on a small scale to see what sort of visual size and brightness you want before you really go any further on the project.
-
The single largest question on this project is "Daytime or nighttime?"
If daytime, then keep it simple which will make it easier and cheaper to build. First thought is an automated version of Nusa's suggestion: Rotating flat panels, each of which represents a pixel, that have contrasting colors on either side and can be flipped remotely. This is exactly the technology used by several roadside sign designs prior to the advent of high brightness, high efficiency LED's. Each pixel was a brightly colored object that could be revealed or obscured. It's like CMOS... nearly zero static power consumption. Just scale up the size for your project. The active element could be anything... a solenoid could work, and require power only during switching with the panel axle having a slight detent to retain its position with power off.
As for electronic architecture, I'd keep all modules identical and run a serial data line where each module reads its bit. That way servicing becomes easy, just swap in any module and use a dipswitch to set its X and Y "address". You don't have to keep track of which wiring harness goes where, etc.
This whole thing could run off a car battery.
-
Flip dots are WAY more expensive (like 10x) than LEDs. A single pixel can cost up to several dollars whereas LEDs are cents. If you are willing to spend the money, you can even get lit flip-dot displays. Otherwise standard modular matrix displays are yout best option. The controller can range anything from a dumb repeating ROM computer that just sets and repeats to a full computer with video. But either way, way cheaper than flip-dot.
-
Agreed that conventional flip dots are more expensive, but the power requirements for a daylight-visible emissive display are enormous, and providing/distributing that power does not come cheap.
I'm wondering if there's some existing thing which can be adapted to make large reflective pixels. Window blinds would seem promising, spray one face black, the other white (or maybe one side in some sort of dirt colour to blend in with the background as your pixels will not be butted close together), and then it's just pulling strings which has a modest power requirement.
-
Unless the "marriage proposal on a private plane" market is much larger than I think, I doubt that this will be commercially viable... To approach this from the other side, do you have a budget you are willing to commit to this?
I think we are looking at a large, mains-powered system, which you will need to build in a way that keeps it safe from wildlife, weather, and plant growth. More likely than not, you will also end up spending more time than you like on maintenance -- replacing failed components, fixing water ingress, mowing the lawn or meadow. (With the installation getting in the way of mowing...)
If you are considering a stationary (non-programmable) solution, do you need to be able to switch it on and off at all? Couldn't you simply ask a local farmer to mow or seed a field in the right pattern?