LED Cylinder
Last night as we were preparing for the party, Eric broke out his camera and took this cool video of my LED Cylinder project. Thanks Eric!
More details on the project, including schematics and sourcecode, are available at http://dclausen.net/projects/ledcyl/.
If you want to learn how to make projects like this, check out our classes! We have several electronics and LED classes coming up, including Introduction to LEDs next weekend and Building LED Displays the weekend after that. Also you should come to our Party / Art of the Game Gallery Show tonight (Saturday, July 26, 2008) to see this project and a bunch of other cool things on display.
19 Comments so far
Leave a reply
[...] [via NYC Resistor] [...]
Will you also publish the schematics for the microcontroller board?
I am wondering how you made a charge pump out of just a diode and capacitor.
Very neat… thank you for sharing the video.
[...] Clausen of NYCResistor built this very sweet cylinder display using 95 RGB LEDs & ATmega168 microcontroller - The LEDs [...]
[...] Clausen of NYCResistor built this very sweet cylinder display using 95 RGB LEDs & ATmega168 microcontroller - The LEDs [...]
[...] Clausen of NYCResistor built this very sweet cylinder display using 95 RGB LEDs & ATmega168 microcontroller - The LEDs [...]
[...] Clausen of NYCResistor built this very sweet cylinder display using 95 RGB LEDs & ATmega168 microcontroller - The LEDs [...]
Great looking project, something I’ve been looking into myself, albeit in 2D. Congrats.
Its a real shame we seem to be missing half of the schematics, it seems. The charge pump would be useful to know about, and how does it all route back to the microcontroller?
Also, why so many MOSFETs in this design? I’m not too clear on why you need to use 3 x TC4427 drivers, and 5 x IRLU3715 chips? Why two different types? And why wouldn’t the LED driver chip (I have some of the 5940 type but never used them) deal with the power requirements itself without the need for the MOSFETs? My only guess is because you are strobing to save GPIOs on so many LEDs, but if you’re strobing them, how does the TLC5940 manage to achieve its PWM brightness control? Surely there’s not enough time.
Still doesn’t explain why we have two different stages of MOSFETs.
More clarity please! I’m nowhere near the NY area, being in the UK, I can’t attend your classes.
Thanks for a great project posting though!
Stu
Thanks for the questions. I will post a schematic for the microcontroller board soon.
The TC4427As are mosfet _drivers_. They sit between the microcontroller and the IRLU3715s, which are real power mosfets. Both are needed because on their own, neither part has all of the necessary properties. The TC4427As can be interfaced directly to the AVR’s 5V I/O pins, but they can’t handle enough current to drive all the LEDs. The IRLU3715s can switch quite a lot of current, but they require a stiff 10V/0V control signal in this configuration, which the AVR can’t provide. Fortunately, it turns out that the two parts complement each other nicely, and can be combined in such a way that you get the easy interface of the TC4427A along with the high current handling capability of the IRLU3715.
In multiplexed LED matrices like this one, the circuit needs both high side switches and low side switches. In my design, the IRLU3715s are the high side switches, and the TLC5940s are the low side switches. The high side mosfets are driven one at a time, sequentially, to provide power to exactly one plane of the cylinder at a time. The low side switches control which LEDs are illuminated in the plane while it is active. This switching happens faster than the human eye can perceive, and so it looks like all the LEDs are illuminated at once, but if you had super-human vision you would see that the planes were illuminated one after the other.
Brightness control is done by varying the fraction of time that each LED is illuminated during the plane’s “on” window. At full brightness, the LED is on for the entire time that the plane is active. At half brightness, it is on for only half of the time, and so on. (Actually the relationship between perceived brightness and duty cycle is not linear — see the “gamma correction” section of the code — but that’s of secondary importance).
Update: I posted a schematic for the microcontroller board to the project website.
Hi, I wish to build your led cylinder and I am a bit confused about how many components I do need. Can you help me and tell me if I am correct?
4X TLC5940NT (led-drivers)
5X IRLU3715 (mosfets)
3X TC4427A (mosfets drivers)
1X ATMEGA168
2X 100uF electrolytic
1X 16Mhz
1X 10Kohm
2X 15pF ceramic
1X 0.1 uF ceramic
2X 1N4148
lot of rgb leds
Thanks again for your time, we are going to build two of those here in Italy! Federico
You would have to create a BOM in Eagle to get the exact parts count. Off hand, I think your list is missing several caps. On the top board you need four 0.1uF ceramics at each TLC5940. You also need one each of 1000uF and 100uF electrolytics (in addition to the two 100uF caps on the lower board which you’ve already listed).
Also you need ballast resistors for the TLC5940s, 1/10th inch headers (male and female) to connect the two boards, a jack for the power connector, and stiff music wire to build the cage.
Good luck! Let me know how it turns out.
Hi, first of all, cool project and thanks for sharing it !
I have a couple of questions which hope you can help me with!
How did you came up with the color decoding ? I mean, if you want to do white, you have to turn on all three RGB components. But, what about the rest of the colors? Was it something like “try and error” ?
Also, could you please explain how the charge pump works ?
Thanks !!
Hi again!
I have now the partlist but I miss a couple of items. Near each TLC5940 there is a resistor, from the pictures looks like violet,black,red, so a 7Kohm. Am I correct?
Then, on the atmega board there are 3 ceramic resistors listed on the schematic (2x 15pF and 1x 0.1 uF) but from the pictures I can’t see the 0.1 . Perhaps you didn’t used it, in the end?
Can you explain me what is this “music wire”? I’ve never seen it and I don’t know what to look for
Looks like a guitar wire, but solid… what sort of instrument use it?
Last but not least I want to ask if you could provide the board also for the atmega part, because I think I will not be able to draw it by myself. You would be extremely generous.
As soon as possible I will buy the parts i need, I am already waiting for the leds from your same vendor
Hi, Great looking project
I’m gonna try to build one my self.
I downloaded the schematics and source code and orderd all the same parts.
Only i can’t find one thing,
The value of the iref resistor.
Can you tell me the value ?
Thanks D Mensink
I believe the Iref resistors are 2k. The TLC5940 datasheet has a graph showing resistor value versus current. 20ma is a safe value but you should easily be able to go higher if you want it to be brighter.
On the AVR board there really is a 0.1uF capacitor on the Vcc pin of the AVR. It is probably obscured in the photos by wires.
Music wire is sold in hobby/craft stores. As far as I know it’s not meant to be used in a musical instrument. I got mine from a store in New York called Pearl Paint, in the section where they have balsa wood, dowels, and brass tubes of various shapes and sizes.
I don’t have a board layout for the AVR board. As you can see it was hand-wired.
Ok, I have just received yesterday the leds, I am excited about the project
Thanks for the resistor hint, this helps Mensink too. Have to look around for something similar to the music wire here in Milano… Now I have to figure out the bottom board, I am not too good in reading schematics 
Thanks!
Thanks for you’re help.
I downloaded the datasheet and found the graph.
I still have to find a way to get my hands on the
TLC5940NT, there a little bit hard to get here in
the netherlands. earlyest delivery date is feb 2009 :-S
or order 1700 pieces
but somehow i’ll find them
Same here in Italy, for the TLC5940NT :/