LED room

Thread Starter

mafiya

Joined Feb 8, 2011
8
Alright so I'm getting an apartment, and want to make the main "living room" area into something like this (see youtube like) except in a smaller scale, maybe the ceiling or one wall. What would be required to do this? LEDs, controllers, etc. I'd have a laptop dedicated to controlling it, and playing the music.

Youtube link:

http://www.youtube.com/watch?v=V9AySS3Ff2Y&feature=related

Other specifications would be that it would be controlled with the music, if anyone has any ideas on what I would need attached to the LED controller to make it go with the music, if there is an open source project to do this? Any ideas that people have are greatly appreciated. If anyone knows exact specifications on what the materials would be to build it.

Also if anyone has a link to projects like this, that I could learn about that would help me do this, and this would be my first large electronics project/working with LEDs (if that is how this is done).

Any comments would be appreciated. Also spending limit on this project is about 1000-2000. Though would like to keep it around 1000 or lower.
 
Last edited:

thatoneguy

Joined Feb 19, 2009
6,359
You'll need a solid number for number of LEDs, the $1000 limit puts that number pretty low, if you price RGB LEDs and the cost of wire (not insignificant anymore). Number of control lines is tripled due to the need to control red, green and blue independently.

For control, that gets a bit difficult. Lots of output in addition to many shift registers would be needed, interfacing the pattern with a notebook computer would require some advanced programming and experience with embedded control systems.
 

SgtWookie

Joined Jul 17, 2007
22,230
You'd need a number of fairly high-performance microcontrollers that were either USB or Ethernet capable, and either have hardware PWM with lots of channels, or you'd have to perform PWM in software, which would be a lot of overhead.

If you were using Ethernet, you would need a bunch of routers to talk to all of the uC's.

If you're wanting to fill up a wall or the ceiling, this project is going to cost a lot more than you're planning on, take much longer to finish than you think it will, and annoy your landlord quite a bit more than you thought it would.
 

SgtWookie

Joined Jul 17, 2007
22,230
You'd probably find it cheaper, far easier, and a whole lot more reliable to fill up a wall with HDTV sets and use them as monitors for a high-end PC.

Sure, you can get Google books in English. The book you were looking at was on a Norwegian server. Try a server in the US or the UK.
 

nickelflipper

Joined Jun 2, 2010
280
That video was really interesting, and makes you think about how it went together, and some silly questions. Like, what kind of man hours would go into programming something like that? And do you think multiplexing is going on there, or full brightness with individual control?

It looks like a limited color palette was used. Lots of processing time could be saved with say a 565 color scheme (ie. 5bits red, 6bits green, 5 bits blue). Another idea is just use 16 or 24 colors total.

Consider a wall of pixels 12'w x 8'h. Pixels to be 2"dia. with 3" on center spacing, similar to the video. That would make a screen of 48 x 32 pixels, making for 1536 rbg leds, or 4608 individual leds.

I bet that a $1/pixel would be optimistic.
 

Thread Starter

mafiya

Joined Feb 8, 2011
8
@SgtWookie/@nickelflipper: Do you have a cheaper solution that would have a same or close to the same results?


Also what is the price range of this? BASTLI

Also this is the software they use there/also the large fixture controller: http://www.madrix.com/

@nickelflipper: so would i need 4608 DMX channels?
 
Last edited:

wayneh

Joined Sep 9, 2010
17,496
You'd probably find it cheaper, far easier, and a whole lot more reliable to fill up a wall with HDTV sets and use them as monitors for a high-end PC.
Ditto. A cheaper approach would be to PROJECT the pixels. Even a single projector these days would have more than the resolution needed, but you might need more than one just to cover the area. Maybe if you customized the optics, you could "zoom in".

Of course, reflected light or an HDTV won't look as intense as the LED wall.

If this project were much smaller, say a 8 X 8 = 64 LED "pixels" array, how would each pixel be wired? Don't they use a transistor array for this?
 
mafiya:
If I understand correctly that one dmx512 channel is equivalent to one byte of data (i.e. color value in this case), then yes, that would be 4608 channels. I do not know if dmx is appropriate, or commonly used for led displays.

Here is pdf on a XMOS ethernet solution. Many other controller schemes could be valid.
 

Thread Starter

mafiya

Joined Feb 8, 2011
8
If you looked at this site: http://www.madrix.com/ it has the controllers for it, and also does music syncing to the lights. As you see in the original video.

@nickelflipper: could i get links for some of the materials i need?
 

Thread Starter

mafiya

Joined Feb 8, 2011
8
In short this is a very complex and expensive project. Your $2000 budget is woefully inadequate.
Here's one.
http://www.luxurylaunches.com/gadgets/barco_nx4_led_wall_costs_22_million.php
I'm not looking at something that big, and that's LED screens, all i'd want is the LED setup for it. Anyone here know what there doing with LED light setups make me a diagram of parts needed?

Or anyone have a few threads that were on here about this? Maybe smaller scale that could be transfered to a larger scale.
 

Wendy

Joined Mar 24, 2008
23,415
As a guess I suspect $1K-$2K will make a relatively small screen, from 3 Ft square to 6 Ft.

The circuitry is relatively straightforward, it is the programming that is the stone cold bear. If you want to see what I mean try programming the effects you want on a simple monitor. This is cheap to do, but would illustrate the amount of work you would have to expend for the µC programming nicely. Actually it would be about 1/10 the effort than it would be for a LED panel of similar size.

I've done a lot of work with LEDs and special effects on small scale, and I've written my share of software, but I find that room very intimidating. You have bitten off a very large bite, but I wish you luck.

As a thought, look for a company that makes LED billboards. They are used as giant video displays. If you can make something for your monitor it could be adapted for the billboard display, and that is established technology (basically the same technology as far as I can tell). Bet the price of one of these puppies is going to take your breath away.
 

SgtWookie

Joined Jul 17, 2007
22,230
Have you ever programmed a microcontroller, and made it light up a series of LEDs, or even one LED?

I have a feeling that the answer to that question is "no."

A recurring theme on here (and many other sites) is that an electronics beginner tries to take on a very large and complex project with great enthusiasm, only to become completely lost and disillusioned because it was too far beyond their level of experience and skills. The scope of this project would mean that you'd need some pretty advanced skills.

Mik3, one of our long-time members, made a spectrum analyzer using a PIC18F4685 and a number of additional components to drive the LEDs. You might wish to have a look at it.
The thread where he introduced it is here:
http://forum.allaboutcircuits.com/showthread.php?t=14246
Blog entry: http://forum.allaboutcircuits.com/blog.php?b=116
PIC code for the project: http://forum.allaboutcircuits.com/blog.php?b=219

However, if you've never used microcontrollers before, or built anything larger than a small kit project, you'll need to start off with something much less complex. Building a number of projects that are relatively easy and don't require lots of time will help to build your knowledge, experience, and give you more confidence to tackle larger projects, along with improving your chances for success.
 

Thread Starter

mafiya

Joined Feb 8, 2011
8
Have you ever programmed a microcontroller, and made it light up a series of LEDs, or even one LED?

I have a feeling that the answer to that question is "no."

A recurring theme on here (and many other sites) is that an electronics beginner tries to take on a very large and complex project with great enthusiasm, only to become completely lost and disillusioned because it was too far beyond their level of experience and skills. The scope of this project would mean that you'd need some pretty advanced skills.

Mik3, one of our long-time members, made a spectrum analyzer using a PIC18F4685 and a number of additional components to drive the LEDs. You might wish to have a look at it.
The thread where he introduced it is here:
http://forum.allaboutcircuits.com/showthread.php?t=14246
Blog entry: http://forum.allaboutcircuits.com/blog.php?b=116
PIC code for the project: http://forum.allaboutcircuits.com/blog.php?b=219

However, if you've never used microcontrollers before, or built anything larger than a small kit project, you'll need to start off with something much less complex. Building a number of projects that are relatively easy and don't require lots of time will help to build your knowledge, experience, and give you more confidence to tackle larger projects, along with improving your chances for success.
Thanks for the info, I'll start by building a small scale project, and yes you are correct, just a small time enthusiast. As far as the lighting goes, the program i linked is for controlling the LEDs, and it analyzes the music and makes the show that you saw in the video. I'll read though those threads and start off with working on the smaller projects, would you guys mind if i kept this thread for working on starting the LED project, and i can show my work from small projects to larger ones? This way, when someone else comes along they have a great place to read about a start to finish, from small to large projects. Or should i rather start a new thread for a smaller project?
 

SgtWookie

Joined Jul 17, 2007
22,230
It helps a great deal to keep threads focused on their original purpose/statement; otherwise things can get confused in a big hurry.

If the topic you have in mind does not directly correlate to this project, then go ahead and start a new topic/thread.

At the other end of the scale is having multiple topics open on the same project; this is also counter-productive.

As far as Mik3's spectrum analyzer - even that is quite ambitious for a "n00b". You'd be better off to start with a smaller project.

You might find some of Big Clive's project kits interesting:
http://www.bigclive.com/shop.htm

...as these are small, quick to assemble, uC's are pre-programmed, and are along similar lines as to what you eventually wish to accomplish. Odds of success are quite high with kits like this.

However, even before you start on a kit, you might want to obtain a couple of pre-drilled PCB's and "junk-box" type components, just so you can have some practice soldering components to a board. Soldering is not an "intuitive" skill; it takes patience and practice to get good. You'll likely destroy a number of boards/pads/components before you get reasonably good at it.

I shudder to think of what my first attempts at soldering looked like; huge wrinkled gray blobs of slag. :p It wasn't pretty.
 
Top