Which arduino board to get? (for large led matrix)

I am kind of in the same boat as some others in here, but I didn't want to hijack someone else's thread.

For my first "project" I have a number of components that all need addressing. The first part of that is probably the most hindering on what board I need to get.

I really want to be able to build a fairly large LED matrix, either using pvc piping to shield each individual LED or ping pong balls.

Something in the range of 10x20 or perhaps a little bigger or slightly smaller. The point being I don't only need an 8x8, and I would prefer not to have to use an external chip to address the led's, but if it's a $10 shield add on and it makes my job easier, I am fine with that.

In regards to that, is the power provided to an arduino enough to power that many LED's? So how does one power that many LED's?

I have looked at the rainbowduino, and the lightuino, but I am not sure if those are overkill, or exactly what I need, or they are missing other features I need.

I have also seen some RGB led's that would be nice to use. As I understand it, that would mean I would need 4 wires in some form or fashion going to every LED, in a matrix.

If my math is correct that means at minimum 50 output connections. 10 to red, 10 to green, 10 to blue and then 20 in the other direction.

The other features I am looking for, is the ability to do some motion sensing and/or light sensing, but that's probably the easiest part, since from what I understand at most that would take 3 analog inputs.

In regards to that what would be the best method to make a simple object sensing. Similar to a ball going into a basketball hoop toy game counter of some sort. I was thinking light beam break, but since the I am not sure how sensitive that would be it might not work.

The last component, which can come later is the ability to run a web server either via wifi or ethernet to control the arduino. I can add an ethernet shield later on, but I would like to know which if any specific one supports basic web server and backend scripting with the ability to control the arduino directly.

I am new to this for sure, but I am fairly confident in my ability to learn, and create something, but I just don't want to spend $50 and then find out I needed to spend another $60 to get the features I needed.

The seeduino mega seems like the ticket, has 70 i/o points, so it would give me room to grow. Any help is appreciated.

If 10x20 means 200 LEDs, and you want them to be RGB leds, that is effectively 200 LEDs. There is no way you can individually address 600 LEDs with any Arduino or clone, with external components.

The least of which will be an external power source for that many LEDs.

You can use shift registers to control up to 8 LEDs, but you will need 75 of them for 600 LEDs. You wont find them on a shield for $10, either.

Motion sensors are generally digital devices. Something is moving or nothing is moving. How much stuff is moving, or how fast, is not reported.

Light sensors need one analog pin, plus power and ground.

There are plenty of threads on the forum concerning ethernet shields. Look at some of them to see what issues people are facing (firewalls, routers, dynamic IP addresses, etc.) See whether those issues are a concern for you. Then pick the ethernet shield that causes you the fewest problems.

For my first "project" I have a number of components that all need addressing

If it is your first project then do something that wouldn't challenge a pro.

If my math is correct that means at minimum 50 output connections. 10 to red, 10 to green, 10 to blue and then 20 in the other direction.

It is not just a matter of outputs you also need to drive them at a refresh rate fast enough to make it look like they are not flickering. That would be say 50 times a second for each row so that is 500 times a second, or one refresh every 2 mS. Also as each LED is only on one tenth of the time they will not be too bright.

Well, this is off to a great start... /sarcasm

I don't mean to sound unappreciative of the advice, after all this is my SECOND post, and both of you have well over 4000 posts, so you've got me beat, by a million miles.

The problem with not wanting to do something that won't challenge a "pro", is that I have an idea for a project, a functional thing, not just some toy, and it will serve a purpose, albeit a fun purpose, but there's some other purpose to it as well. I see a lot of "test" projects, that don't really serve much purpose, don't show much imagination, and overall don't really fit into "that's worth me doing", at least for me.

Here's a mockup of what I am after, for all intents and purposes it's a recycling collection "basketball" game, with an led sign. It's not terribly complicated, but the principal is to be able to see the sign from a distance. (fixed image)

I did some looking around and people seem to have gotten a 'sure electronics' large led display to work with arduino's. Is this another component of my project that's going to "challenge" me?

If I choose to go with that, as opposed to my unrealstic custom made led matrix, will I need anything beyond an arduino mega, or a seeeduinoa mega. Or is that overkill?

It sounds like a motion sensor might be what I am after. If I bury it low enough in my "ring", it might work fine, but perhaps I am wrong about that. The problem however is that this whole "rig" is going to be mounted to a frame, and dragged behind a bicycle, so if it's going to sense any slight movement in the bag, maybe that's not going to work then. Will need to think about that, but I am willing to experiment with a couple of different things. Even though I think getting that just right will be pretty difficult, I feel like the display is actually the most difficult aspect of this, but maybe I am wrong.

50FPS is an insanely fast refresh rate, 24fps (otherwise known as what the majority of films are shot in) would be the maximum I would need, if that even, and not every led would be lit ALL of the time. Besides that, it sounds like a fairly unrealistic goal. I just saw quite a few people building matrices with larger baffles around the LED's it provides the illusion of a larger LED panel, without the need for a ton more LED's. At this point I would be fine with starting off with something smaller (in the less led's sens) like this: http://www.youtube.com/watch?v=QXNxg_AjSUw But it would have been nice to go with something just a little bigger, but if it's not possible then so be it.

Anyways thanks for the advice, I am looking to get started, so definetly some advice on a respective arduino for the overall project concept, so I won't say "I wish I would have..." when I get to the final stages.

Well, you need more outputs than any arduino has so you're going to have to use some components.

An arduino mega would mean you'll have to worry less about conserving pins and memory.

Shift registers take a serial input and output it across a number of pins, most commonly eight. So you get your arduino to tell it 11110000 and the shift register turns the first four pins on and not the second four. It's been a bit since I used one but they take three pins to control - you could nest these in order to create a matrix. If your matrix is limited in what it displays you could do it in far, far fewer components. That is, if you only need it to display a few set patterns rather than have every light individually controllable.

Motion sensing through a hoop is easy - you can use an IR or laser trip-wire. When the light goes out something has passed by.

I've got a Sure Electronics matrix (a small 8x32 one) and it took about ten minutes to set up, and I'm a novice! Someone else had done the hard work and made a library for it, so it was easy sorting the code and wiring up was really simple (much easier than wiring an LCD display). Use the search to see if anyone's used the specific component you have in mind, and see what code is available for it.

I paid under £10 for my matrix, I'd say it's well worth getting one (or a few - you can chain up to four together) to have a play with to see how it works and how you want to do your big matrix.

I've got an ethernet shield too. The hardware is easy (well it would be, it's a shield) and the software isn't too bad. I've not done huge amounts with it yet, but I do have a physical hit counter in my lounge that clicks on one every time my web site gets a hit.

24fps (otherwise known as what the majority of films are shot in) would be the maximum I would need

The 24fps is for the persistence of motion. On a movie film each single frame is flashed on the screen three times, giving a flash rate of 72 flashes per second. The persistence of vision (not motion) is at least 32 Hz, and many people can see flicker faster than that. That is why you need 50Hz refresh rate to make sure but you defiantly need over 32Hz.

The problem with not knowing about a subject is that you can often stray into areas that appear to be easy but are in fact hard. That is why you are asking here isn't it? So we are telling you that driving a large matrix is tricky and the results might not be as good as you think.

Grumpy Mike, I totally understand.

Much like many things in my life, I am an expert on pretty much NOTHING, so I digress to advice, which is why I am even asking.

That's also why I am looking into other solutions, like the pre-made modular led matrices, that seem to have a way to interface with the LED's using an alternate method, that I may not have the skills to make right now.

My electronics experience is limited, but I am willing to learn, just need a push in the right direction.

This is interesting albeit a tad expensive... http://www.synopticlabs.com/blog/?p=97

but cheaper than the other similar solution, making an rgb led matrix with blinkM's which just for an 8x8 matrix would cost well over $600...

The rgb led strip can be found on eBay for around $120, still pricey, but awesome.

On that article they seemed to be powering the entire strip with only 5v... So that would tell me the arduino is more than capable of accomplishing that.

With 160 LEDs one could make an 16x10 led matrix that measures 50mm by 80mm which is a pretty good size.

I saw that someone actually did just that, so it's possible. Anything else I would need to be concerned with in regards to that strip?

Here's a mockup of what I am after, for all intents and purposes it's a recycling collection "basketball" game, with an led sign. It's not terribly complicated, but the principal is to be able to see the sign from a distance.

I get the idea conceptually but what do you want to display on the sign? Probably simple text and patterns - right?

On that article they seemed to be powering the entire strip with only 5v... So that would tell me the arduino is more than capable of accomplishing that.

The question is less about voltage, more about the current the arduino can provide.. If you have an electric car whose motors work on 5V, do you think you can drive it at 120km/h when you have the motors hooked up directly to an arduino? Powering a couple of LEDs is possible through the arduino, but that is it... You'll definitly be needing a seperate power supply for the LEDs.

(Warning: I don't really know exactly what I'm talking about)

You could use a bunch of 32 bit shift registers (http://www.youtube.com/watch?v=DMP1BQHCg6k), a hell lot of amplifying transistors, and an arduino mega and you should be ready to go. You can also try charlieplexing them. Now, you'll need a high current input cause 200 LED is a lot of current even if each of them only takes 20ma.

EDIT: if you can't find any 32 bit shift registers, make your own with a very large bunch of D flip flops.

I get the idea conceptually but what do you want to display on the sign? Probably simple text and patterns - right?

Yes, numbers, scrolling words, patterns/shapes. Something like my symbol for my company, a recycle symbol animating (if that's even going to be possible), arrows animating to the right, to the left. Anything else I decide I want to show.

The "game" is for a night time parade like event that happens monthly, and people tend to liter, and this will be a fun way to influence people to not throw their recycling on the ground.

Powering a couple of LEDs is possible through the arduino, but that is it... You'll definitly be needing a seperate power supply for the LEDs.

The LED Strip comes with a power supply. But I am interested in powering them with a car or motorcycle battery? Or a lithium-ion battery... I need to make this somewhat portable, which is why I am also interested in the power usage/consumption. How would I caclulate the power usage of that led strip? Or the LED matrix that Sure sells?

I think I've written off doing anything beyond using a pre-made matrix, or using programmable LED's like the blinkM's, or the LED strip I posted above. It's beyond my abilities, at the moment.

Now back to the main question is the barebones arduino going to handle this or again am I going to regret not springing for a mega (or the seeeduino mega which has another 20 extra i/o pins), or some other specialized arduino like the rainbowduino...

Secondary question, based on my mockup, what is going to be the most reliable way to check if an object has passed through the ring? Since the object (in this case, cans, glass bottles, etc.) will be variable shapes, and have a wider area to pass through. Unlike say a door, where someone walking by pretty much can't avoid "breaking the beam", but maybe I am over-estimated the wide beam pattern of some of the options out there.

I have looked at: IR sensor LED sensor Super sonic sensor photoelectric PIR Motion Sensor

I am just not sure what would be best, especially since this entire thing is going to be moving, so a motion sensor might detect motion when there is none. AAAHHHHGGG! Experimentation is necessary, for sure. But what would be my best option.

BTW, thanks for all of the responses.

Afraid I can't help you in regards to power consumption, I'm a coder, not an electrician =P (allthough the arduino has sparked some interest ;) ) Though the use of a seperate power supply is definitly the right choice, lithium-ion stuff packs the most power per volume / mass.. but I'm not sure it could compete with a car battery for instance, I don't know.. perhaps somebody else could help you with how long an estimate of LEDs could be powered with either choice.

Now based on what you want to do, it does sound like you could do it all on a duemilanove, assuming the number of IO pins can be solved with shift registers or another way to adress them all (the recently posted centipede shield might be suitable?). Doesn't look like you'll be needing tons of PWM, multiple serial devices... so I'd take the duemilanove, which has slightly better support.. and lots more shields available =P

Would a modified version of this board suit your needs?

Peggy 2 Light Emitting Pegboard Kit

It uses up to 625 LEDs, can be configured in a RGB pattern, it's open source and arduino compatible.

Here is the link:

http://evilmadscience.com/tinykitlist/75-peggy2

I'm currently working on a project very similar to the one beeker has in mind. With lots of ideas but a very limited background in electronics, I've had many of the same questions and countless others.

I'm a dj and club/event promoter and like to create a unique atmosphere for every show. I have a more LEDs than I care to count from use with on a previous show and now desire to have individual control over when each LED is on/off. Thinking I could accomplish this task using the Arduino I recently purchased, I set off to make it happen.

My initial ideas were very grand, and needless to say have been scaled back quite a bit. I've now decided to make a "simple" LED matrix to use as a backdrop to the dj. I have a bunch of 4-inch PVC pipe laying around that I will use to accomplish this and have been cutting it into 5-inch segments.

I have had many problems attempting to teach myself the circuitry of a project like this, and so I decided to order a Peggy2. I received it yesterday and will be putting it together later today. Instead of soldering the LEDs directly to the Peggy2 board, I plan to run some (many) leads to essentially expand its size. I've chosen to use a matrix that is 16 wide by 8 tall to get a desirable size, and hope to get multiple LEDs in each.

I still have alot of questions and may reply later in a quest for answers if I can't figure them out myself, otherwise I will respond with pics/video.

If anyone is aware of a potential problem I face, I'd be very appreciative if you would let me know.

Trevor

you mean like this

http://blog.makezine.com/archive/2009/09/giant_peggy-based_led_display.html?CMP=OTC-0D6B48984890

Wow, that's great. I like that one a lot, but it's more like this one, only larger:

http://www.youtube.com/watch?v=hrTe-2u2LMs