I've been working on a circuit design that will drive LEDs using an tlc5940. To do this i'm going to use a 2n0355 npn transistor and have the tlc5040 drive it at the base. The tricky part is making sure that the amount of current and voltage drops across the leds fall within a safe range.

For the red luxeon 111 star, the typical ranges for V and I are 2.95 and 1.4. By ohms law, the resistance of this led should have 2.1 ohms. By this analysis you could put four resistors(leds) in series connected to a 12V source and get the appropriate drop and voltage across all four resistors(leds). The current coming of a pin on the tlc5940 can be set anywhere around 60mA, this current is perfect for driving the transistor as it's Beta is 20 and can push up to 15A.

In another thought, i was playing with the idea of using old processor heat sinks and fans for heat dissipation from the diodes. More come on that one.

-- BezBurn - 20 Jul 2007

Xander sez

This idea was something I considered doing with the tlc5940, using it to drive a transistor, but ultimately realized that to do so would be a waste.

The biggest key to happy high-power LED drive is constant current, as they tend to heat up considerably and, when they do so, they change their Vf and current parameters significantly. Note the specs for the Red Luxeon III at 1400mA with, in particular, forward voltage drop. The 'typical' drop is 2.95 volts, but the minimum drop is 2.31 and the max is 3.51. That is a massive variance and most of it is temperature driven (note the column "Temperature Coefficient of Forward Voltage" in which is specified that the forward voltage drop decreases by 2 mV for every degree C ). Notice that when the LED decreases its voltage drop, that makes more voltage available to the other elements in series with it, which means more current is able to flow, which means the LED heats up even faster, causing it to drop even less voltage, etc etc etc.

If you don't use constant current drive, you have to plan your design using the most conservative numbers.

However, the cool thing is that the TLC5940 has constant current drive, you just need to be careful not to ask it to drop so much voltage (as heat) that it burns out. Also, the TLC5940 can actually sink 120mA of current, not 60 mA, provided that you set Vcc to > 3.6 V (which it will be.. our logic power supply is 5V. Note that this is different from what you drive the LED's at. It's entirely feasable to have the TLC5940 sink a string of LED's which are powered by 12V while the TLC5940 is only powered by 5V.

So how do we be careful to not have the TLC5940 drop more volts than it can chew and burst into flames? Simple, we pre-regulate the voltage entering into a LED string to be as close to what the LED's will drop as possible. Thus if we are powering a string of 3 LED's which each drop 3 volts, for a total of 9 volts across the string, we put a regulator at the top of the string which regulates the 12V to, say, 10V, leaving only 1V for the TLC5940 to drop and it can manage the thermal variance of the LED's from there.

However, personally I'm not going to be able to have the TLC5940 drive ready to go by burning man this year. Instead I've been planning on going the simple route of controlling all the LED's of a particular color off of one beefy MOSFET, with the LED's broken out into chains (3 per chain), each driven with a linear regulator like the LT350 in constant current mode. With a massive freakin heatsink on the voltage regulators. I'm expecting regulators driving the red strings are going to need to dissapate 5 Watts (since the Reds have the smallest voltage drop).

--x

-- BezBurn - 20 Jul 2007

Met with Xander last night and discussed the current out look of what we might be able to finish for burning man 2007. Which is, the caustics generator, and led light dish. Both of these would have to be run off of programmable chips and will not be interactive at all. By taking down any components that we can, even in an unfinished state, it'll give us a chance to see and evaluate what we have in the environment of the Black Rock Desert. I'm now going to be heading up the light dish so that Xander can concentrate on the caustics generators.

A portion of zoe made it down to critical massive, now dubbed Protozoe, and made a visual impact even though the final version will look and behave very different. Protozoe used a simple design of balloons filled with water, rope, and pullys to allow participants to move water in a dish elevated at about 11ft. The light source was a 100W green and red flood lamp. The main problem with using a flood lamp is that the reflector makes the light source look bigger, thus the patterns on the ground were less defined than what i had hoped for. The only way to offset this and get more definition of the wave patterns was to raise the lamp higher, which i did until the throw pattern became too small.

-- BezBurn - 06 Jul 2007

It is now 2:55 in the morning. One particular thing that happens with me and sleep dep is i start to see little black cats. This can be startling as i don't have a little black cat. The model is finished and it's looking good and seems to be fairly strong even though it's made out of poster board with a sheet of paper glued to it. The stand offs will be a little smaller comparatively, but with the washers used in attaching them to the structure will make up for it i thinks. The one area of concern is that the legs have a tendency to bend at certain spots. It is not very much and due to the material almost expected. The bend is also due to imprecision of the s. o. slightly twisting the poster board. This problem should be greatly reduce with precision manufacturing of the pieces. The shape of the top of the model makes if very strong and should hold up nicely. The model is built to a 0.058 scale, or is ruffly 6% of the intended size. For more pictures, see the attachments on the main page. Using sketchup i've completely redrawn the entire structure to spec. with each piece being componetized so that it can easily be taken apart and measured and/or copied for manufacturing. The only thing left to do is to hole patterns on the side which will have to wait until sunday or monday. By then i will have talked to rusty again about the thickness of the steel based on findings from the model.

-- BezBurn - 12 May 2007

I realized the other day when drawing up the final drawing of zoe that i'm not sure what thickness of metal that we should use for zoe. So i've contacted Rusty from hazard factory to see if he can give us an idea. When he writes back i can finish all the little details. As you would imagine, the thickness is going to through off where the holes are going to be for through hole bolting. In finishing the drawing, i'm attempting to incorporate more slight curves to the legs to make them have a more natural feel and look.

-- BezBurn - 07 May 2007

-- AimeCeleste - 28 Mar 2007

I've attached a file which we are distributing to get a quote on the cost of forming the steel. See Quote_list below.

-- BezBurn - 10 Mar 2007

So, i have made some sketch ups of some possibly cheaper models, please see attachments. I've been playing with sketch up for so long now that i'm trying to rotate this screen. Tell me what you think. The plusses of doing something like these sketches is that it uses less pipe and has significantly less welds. I went looking at premade skylights and found one that is 50" x 50". I'm waiting to hear back from them about how high the dome is, but i think this may be an option because they only run $260. here is the link PDF of a skylight, and here is a place that has prices.prices.

The project will go forward. We can still get some money from the nw ignition, as well as a fund raiser. I'm going to bed now.

-- BezBurn - 28 Feb 2007

* Notes From Bez*

I have contacted several suppliers of acrylic domes to see if we can get ourselves a better price. So far i have found one supplier who can make a 37" lens for 400. I'm going to go down there on march 2nd, friday, and see if what they can make us is indeed what we need. if it is i'm going to go ahead and put in an order for it. On the phone, Louie the guy i talked to, said that they had some preset sizes of 37" and 49". These of course cost less than what we want cause they have to retool everything for a special order. Depending on how much is costs to get a special order i will get what we want. Although it might be cool to get the 49" dome... In other news, I'm looking into getting us another grant from the NW burning man group. If we got this grant we would have to take zoe to the regional burn on june 25th, which i was planning on attending anyway. I am from this moment forth referencing psyphozoa as zoe. It's shorter / takes less time to type and it's better than chuck or barry. Don't get me wrong, there is nothing wrong with chuck or barry, i just don't think they make very good names for jellyfish.

Topic attachments
I Attachment Action Size Date Who Comment
docdoc Quote_List.doc manage 305.0 K 28 Mar 2007 - 03:00 AimeCeleste This is the list of requirements for what we're doing - created so we can get a quote on the price
elseskp Zoe_with_3-_2inch_pipes_on_each_leg.skp manage 876.8 K 10 Mar 2007 - 10:38 BezBurn Newer vesion of zoe with 2" pipes and only three per leg
elseskp triangle_pyramid_with_top_cut_off.skp manage 182.7 K 10 Mar 2007 - 10:37 BezBurn Triange pyramid with out peak
Topic revision: r8 - 31 Jul 2007 - 05:30:55 - XanderRRednaX
 
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback