Giant full color led display

[jmaia]

daktronics pixel theory

Ok, no more sarcasm please. Let's take it seriously from now. I'm working very hard to build a big display. If someone want to participate with me let's make a private group and work together. My English is not so good but at the end I hope we understand each other.
Best regards.

 

[blueroomelectronics]

full color led sign 40x 21

40x40 gray scale (25 8x8 displays) is more reasonable but still no cakewalk. To get any brightness out of it you're going to have to split it up into 5 sets of 8x40 displays.
40 x 60ma = 2.4A
times 5 sets = 12A
A PC power supply can deliver that.
Have fun

 

[jmaia]

synchronise led display

That's exactly this I.m doing. I turn every 8th line at a time.
Regards,
Maia

 

[xorcise]

giant led matrix circuit

Maia,

The sarcasm is well deserved for the people it applies to. What you posted is very pertinent to the subject and is to be commended. By itself those articles should be enough to answer the original questions proposed in this thread.

I would be very interested in your approach to what I did in the previously posted video. Are you multiplexing rows or columns in your design? What is your largest Duty Cycle (oops, just noted 1/8 ), and what is your multiplexing speed?

Added after 5 hours 38 minutes:

Okay... let's add some more concepts and ideas through demonstration. This video actually reads 2 ADC inputs of a PIC and then changes the speed of a moving block of LED's on a matrix. Each ADC voltage/reading changes the speed of an individual and identifiable block. 2 LED's rotate in an uneventful way in the center, but are independent of the other LED events. Also, note that the LED's mix and overlay at some points. This is a fairly easy applicaton to a 40x40 matrix.... it's mono and doesn't involve gray-scaling.

This one simply rotates ASCII 5x7 Characters on a matrix as it moves through the list of characters consecutively. It starts with a fast rotation (faster than my cheapo camera at many points) and then slows enough to identify the character. Neat demo in case you would like to add a xyz sensor, and create an IPhone effect. It shows how a character can be configured in bytes depending on the Rows or the Columns perspective. This is the same difference essentially between how the very common LCD dotmatrix controllers, HD44780 and the KS0108, display characters on their respective screens.

 

[marwan naboulsi]

applyng stm32 interfacing rs485

بهايم
please let'us return to the stricke

 

[blueroomelectronics]

8x8 rgb led

بهايم
please let'us return to the stricke

What's a "stricke"

 

[snarfer]

led rgb display control technique

Well first of all I think this topic deserves a new start. It is a total mess what with all the edited posts. Nevertheless it is quite an interesting subject.

In any case I have been thinking along these lines for some time. I started making a small matrix, 4 by 4 RGB LEDs with 6 AS1109 constant current LED controllers, controlled by ATTiny2313 with a DMX interface via UART to to RS-485 transceiver. I have not finished it yet, but can post pics as soon as it is done. Color changes controlled via Bit Amplitude Modulation as described on website of Artistic License. Actually they make some very large and expensive LED video walls themselves, so it is an interesting site to visit.

My thought was to make lots of small and cheap units that could be grouped into a large array. I had a nightclub interested in covering one wall with this, and could spend about 10 thousand USD. Probably my development has taken too long for them though...

With about 100 or so units one could loop some seconds of basic video. Each matrix would have its own controller that could store the relevant colors for that matrix, then overall control and timing would be via a DMX control board.

In my case I want the LEDs to be lit all the time, so multiplexing wouldn't work. However, if multiplexing is an option then you could do use a different type of controller. I have seen an application note from Altera on how to configure FPGA as constant current LED driver. Not sure if this is cost effective or not.

I had not thought about VGA interface. This seems like too much complexity, and is real time output necessary? Maybe just take the video through an imaging processing program like Shake, extract relevant frames, translate to lower pixel resolution and frame rate, generate individual pixel values, then upload data to the individual controllers. The controllers should support DMX RDM, or better yet ACN, with ability to download some color data.

I am just summarizing here. If it's interesting and/or relevant can discuss further. I have also thought about PFC and load balancing issues. That is a whole different can of worms.

 

[blueroomelectronics]

giant led panel design

If a nightclub wanted to cover a wall with light they'd use a LCD / DLP projector.

 

[xorcise]

full colour led display build

Well first of all I think this topic deserves a new start.

Absolutely NOT!

How badly must we become bored to tears with so-called "experts". All the appropriate information which answers almost any LED Display design question can be found right here, as generously posted by jmaia a few posts before:
http://jsmaia.googlepages.com/Agilent-LEDmatrices.pdf
http://jsmaia.googlepages.com/ConsiderationsoftheDrivingElectronic.pdf


Anyone who has pretended to be smart and know anything on this subject has not even built the simplest unit, and some of these persons have claimed expertise at all levels of the design. It's easy to conceive of meaningless, wrong, and impractical designs, and post them publicly, especially when readers don't have the requisite skills to see through the smoke, and then embarass them for their pretense and nonsense (directly related to the disappearance of posts and pictures in this thread). Starting a new thread will only encourage conceptually and electrically inept people to put in their 2-cents of worthless notions, which helps nobody.

If you want to start a meaningful thread, start with one where you cannot post anything unless you also post pictures/videos of a working bench-wired prototype. It's up to them to tell how they designed, built, and created the necessary software to drive the display.

This particular subject of LED Displays (small or large) attracts experts and beggers like flies to a pile of fresh cow crap. Most are pretend experts along with a very large group of leeches, and every one of them has been sucking on that pile of crap for way too long. You know what they're full of now! Eating those flies always leaves a strange taste in my mouth...

I think that I have articulated the issue enough.... I say put your money where your mouth is. Then the thread has possibilities to be meaningful and helpful to others. Go ahead... if you have the cajones.

 

[digitalpbk]

led display vertical

hii,
we are trying to do an RGB LED display... (https://rgb.kitiyo.com)

Has anyone done a similar project ? pls let us know ....

 

[blueroomelectronics]

giant led sign project

Well said xorcise, LED matrix displays and SMS phone projects dominate these forums. Seems nobody yourself excluded has built either with any success. I plan on doing a future single color matrix LED kit of a reasonable size 16x16 but not for at least another few months. I'd be happy to see a design that does not exceed the specs of the parts chosen and has plenty of brightness to spare.
RGB displays add triple the complexity of monochrome, and I don't want to use exotic LED ICs like MAXIM for the project.

 

[Alan69]

marwan naboulsi



LOL, almost any LED display design question, that's a laugh. That is one or two methods, at best. A start, hardly everything known to man. Or even just your average sign man.

It's equally easy to dismiss differing designs as just smoke, when it's really the fog of your only really having one or two methods in your own experience. Even when one person's idea isn't exactly correct, it may lead others to useful ideas. While you've built something, I'm not seeing where you're well versed enough to be saying everything else is worthless. I'm not getting the notion of 'well versed in all of the different LED Display driving techniques' from your posts.

"If you want to start a meaningful thread, start with one where you cannot post anything unless you also post pictures/videos of a working bench-wired prototype. It's up to them to tell how they designed, built, and created the necessary software to drive the display."

That's just silly, and obviously geared toward your just being able to dismiss anything else since you've built a little. You'd seem to be of the new generation, everyone owes you everything on a platter, or you can say they're wrong. Hardly. At best they might donate a good idea or two and a start on something, from there get off your own ass and do some work figuring it out. Many designs simply aren't practical unless done on a commercial scale, that hardly means discussion of the techniques aren't worthwhile. Often even when the whole design doesn't make sense in small scale, aspects of it will, and more often will end up applying well to other ideas.

"Go ahead... if you have the cajones."

Don't even enter into it. A board is information exchange, pure and simple. Blustering is simply a cover for those without solid understanding, as soon as someone starts talking in those terms like they're meaningful expect them to be out of their range.



For the most recent poster:

Not outside the specs is a ridiculous consideration. Almost all electronics parts are made with a huge tolerance for manufacturer return reasons. While on a very high end design it can make sense, basically no low end even commercial design doesn't exploit this. You drive things hard, and handle the few actual failures you get yourself, it'd be stupid commercially to not do this, and would make a sign company instantly non-competitive. It makes even less sense in a home/one of environment, you do whatever works, replace it if something fails early on, and realize it will probably stay working for years once you have it working beyond a day or two. You'll be putting in assloads of extra components for very little reason to make it stay in spec, and actually make your designs less likely to stay working from the extra parts and connections. Still makes sense for some power designs etc where heat is a concern, but silly for LEDs etc where failure isn't likely to be a problem anyway.




**broken link removed**

**broken link removed**

 

[xorcise]

New generation.... I doubt it.... I'm very likely older than you old man. I'll ask my 21 and 30 year old kids if they think I'm new generation. Try again.

Blustering is simply a cover for those without solid understanding, as soon as someone starts talking in those terms like they're meaningful expect them to be out of their range.

You doth protesteth too much maybe. I make my point not as bluster for bluster sake.

I wonder how much bluster and cover for my inadequacies I can work up here.... :
https://www.youtube.com/profile_videos?user=circuited
I am open to any questions to test my credentials and aptitude on this subject.

How worthwhile do you think it might be if I was to throw it into this dogpile of ne'er-do-wells?:


I am personally happy to teach and explain, but not on a framework created by others who have intentionally bamboozled and further confused the matter. Of what value is a discussion if it's all about bad advice and incorrect data, and then the discussion proceeds based on wrong and faulty premises.

Thanks for posting picutres/videos... now we have something to discuss that can prove educational and worthwhile. On the smaller whirling display (2nd video), can you tell me the rpm of the spinning? Also, how are you tranferring power to the LED's through the armature? What microcontroller are you using; or by what method are you transmitting display data?

 

[Alan69]

Thanks for posting picutres/videos... now we have something to discuss that can prove educational and worthwhile. On the smaller whirling display (2nd video), can you tell me the rpm of the spinning? Also, how are you tranferring power to the LED's through the armature? What microcontroller are you using; or by what method are you transmitting display data?

Regardless of age, you still seem to be a bit victim to thinking everyone owes it to you to spell out every detail or what they said doesn't count. Right ideas are right ideas, period.

Nice enough, and bothered to upload some video. But still relatively standard fare. I was doing in that range 25 years ago out of the Commodore 64 user port with assembly.

**broken link removed**

Sort of unlikely anyone else here has worked for a message center manufacturer, and worked with signs like this since the mid 80s. I was troubleshooting message center controllers and driver boards 25 years ago when I was 15 or 16, and had already been working with them in other aspects since I was 12. It's all old hat, almost all of this stuff was already done in the 80s. It's only cheaper/easier/brighter now. Heck even full color was available then, but it was wedge based lamps. or 35W bulbs.

That was part of the point, some of the other ideas that you'll be likely to dismiss can have plenty of bearing. There are many ways to skin this cat, whether or not they can take them all the way through. Dismissing them for silly reasons like they haven't built it can be just as bad as offering complete crap..


**broken link removed**

**broken link removed**

The little one was on a 16f84, the larger one is an 877. One bearing gets the ground there, second bearing gets isolated and takes the power from stationary to rotating. Programmed data and IR so you can type with a normal IR remote. The big one in the dark was a resolution test, the outside blips are 'Joy ' at minimum cycle time, something like 1200 or 1300 dots per rev. LED die is wider than that, but that's the edge resolution.

Need to find some pics of my little sign, 7 or 8 x 64. Also have a 12' $5k red/green message center in storage..

Personally I have far more interest in building one of these right now, but I'll try and dig out some of my sign stuff shortly anyway..

Alan

 

[xorcise]

That's a very bad assessment of my world view, and you must terribly underestimate my experience in this regard. I'm full of social graces.... but getting pissed is within my realm of emotions, and not necessarily in a bad way. It actually helped clean up and redirect this thread rather quickly.

Asserting myself seems to have the most offensive part of my verbal antics. I could only wish people would listen and understand words rather than just note that his lips are moving.

I see that you also have played with the Rockwell 6502 as I did in the late '70's. My other love was the Motorola 6800, which was the twin brother of the 6502, from a different mother. I developed my first industrial programmable controller around the 6502. Or maybe you would rather hear about my projects with Blue LED's when they were brand new and $50.00 each.... donated, of course, to the university for research purposes.

 

[Alan69]

That's a very bad assessment of my world view, and you must terribly underestimate my experience in this regard. I'm full of social graces.... but getting pissed is within my realm of emotions, and not necessarily in a bad way. It actually helped clean up and redirect this thread rather quickly.

...

I see that you also have played with the Rockwell 6502 as I did in the late '70's. My other love was the Motorola 6800, which was the twin brother of the 6502, from a different mother. I developed my first industrial programmable controller around the 6502. Or maybe you would rather hear about my projects with Blue LED's when they were brand new and $50.00 each.... donated, of course, to the university for research purposes.

Hey you said the only thing that counts is etc etc.. While I agree with some of the useless posts comments etc, going to an equal extreme on the opposite end of the scale isn't really any more valid..

Yep on the 6502/10 but never got into the 6800. Also worked test and repair at a burglar alarm manufacturer troubleshooting 8751 based systems, so went straight to microcontrollers. Definitely a lot easier when someone else is footing that $100 each bill when one blows..

Yes had the blue LED in my C64 power LED early on, about as soon as they came out. Being at an LED sign manufacturer and about 30 miles from Cree, about that same time we were getting test panels with hundreds of blue LEDs per. Subsidized of course, since they knew no one will use them at high cost like a sign manufacturer. Seriously thought about melting the epoxy on one and selling off the LEDs, but was difficult to do without damaging the LED. And they would have gotten mad anyway and figured where they came from. Funny thing was with the real super bright reds recently out, there were two more expensive greens and two blues for each pixel in a larger X, so a whopping 256 blues on the 8x16 test panels. Done to get the ball rolling for signs, and no doubt not the prime LEDs like they'd sell individually. Also of course the first 2 or 3 years you only saw a sign like that at a major football stadium, no one else would have revenue to justify the expense. Signs were pretty much too expensive to justify until they hit $.50 each in quantity, so the original panels were more to generate buzz and get interest going so things would be ready..

 

[Alan69]

My thought was to make lots of small and cheap units that could be grouped into a large array. I had a nightclub interested in covering one wall with this, and could spend about 10 thousand USD. Probably my development has taken too long for them though...

In my case I want the LEDs to be lit all the time, so multiplexing wouldn't work. However, if multiplexing is an option then you could do use a different type of controller. I have seen an application note from @ltera on how to configure FPGA as constant current LED driver. Not sure if this is cost effective or not.

I had not thought about VGA interface. This seems like too much complexity, and is real time output necessary? Maybe just take the video through an imaging processing program like Shake, extract relevant frames, translate to lower pixel resolution and frame rate, generate individual pixel values, then upload data to the individual controllers. The controllers should support DMX RDM, or better yet ACN, with ability to download some color data.

I am just summarizing here. If it's interesting and/or relevant can discuss further. I have also thought about PFC and load balancing issues. That is a whole different can of worms.

Don't go too small, costs will eat you up in a hurry if trying to make a large total display. Since pretty much any LED can handle at least 4x current peak, there's no sense in not doing at least that, you'll be adding extra hardware/expense for something that the LEDs can easily handle.

And DVI is the way to go, since it's already digital. Run a display, set it to the lowest res so the timings are lower, and tap the signals from the connector of the running display. Will be trivial compared to doing ADC for VGA.. Yes, you can do all that crap easily enough for each video, but doing the work for every video will quickly outweigh doing work once and getting a direct from DVI.

Load balancing shouldn't be too big a problem with LEDs. Super problem with 50W light bulb signs, plugging in a single extra driver (8x8x50W for 3200W total per driver) on the wrong rail tends to unbalance a sign and trip breakers. Unlikely to get that kind of load with LEDs, but the 1W or 3W LEDs might do it if you can afford enough to make real signs from them..


Also note you could do something like this:

https://www.youtube.com/watch?v=jkqoPkuJaMI&feature=related

Not nearly as many LEDs necessary. Not full video, but could make for a display that would have nearly as good a level of interest. You can always get a projector for full video. People can often tune full video out, something like that may actually be better in the long run.

 

[Alan69]

I ran into a few items and put them up, here are a few pictures..

Several here:

**broken link removed**

37 Matrix sign part from a commercial unit from the early 90s, board up top was for a PIC to run it. I'll find the code and get it going sometime soon.

38 Tristate matrix test. 8 outputs can run 8 seperate 7 segment LED displays. You use Low as the common low to choose display, High lights segments, and tristate for the off segments. Really this was 11 outputs, so can run 11*10 segments, 10 displays and the other did the inputs. Matrix, so you had to have diodes for each switch on the inputs.. Transistors as emitter followers to have enough current for this kind of highly multiplexed system. Complex wiring pattern since there is a rotating common but once done it's easy, code for adjusting for the common pin is easier than one would think off hand from how it works. I'll put up schematics and code sometime on this, very useful for certain things. Likely to burn an element if it ever hangs on something with high current and lots of multiplexing, but worth it if it's easy to fix a burned element.

39 Back of above.

40 Matrix for testing. I have an older one made for testing 120V triac drivers, 1KW plus of 35 or 50 Watt bulbs gets hot in the summer for test and repair, LEDs are much cooler when repairing driver boards. This one was a 541 driver and PIC test board, so from the early erasable PIC era, like 1993 or 1994.. Put LEDs in a board arranged the same, then solder both leads so it stays in, but leave them long. Get all LEDs in this way, laying one lead over to the next to make your row connections. Push a piece of construction paper etc onto the remaining straight leads for insulation, then bend the other lead over and solder to the next to make columns. No problem to make your own matrix with no pattern etc, just use the leads. Not really any more work than just soldering in place..

I've got numerous other interesting sign items in boxes, I'll post some pictures here as I run into them. Just the things I've kept over the years of working off and on doing signs at a manufacturer, lots of items were tossed when I moved last year though. Working on PICs and going in the boxes anyway, so not really any extra work to put up a picture and link..

 

[echo47]

Here's a brief overview of a large LED display controlled by FPGAs. The article talks mostly about the controller, not the LED drivers.

"Sign of the Times - Xilinx makes high-tech outdoor advertising in Times Square possible"
**broken link removed**

 

[drbizzarow]

atlast people comming to my point