greeeg 460 Posted January 18, 2014 Author Share Posted January 18, 2014 @@cubeberg The G2553 has plenty of power to drive these LEDs. a little less backbone if you wanted t do some fancy fading or color transitions. @RobG I've been thinking about this. and I think the ring would look better without having a load of unpopulated part on 4/5 of the board segments. With regard to the LEDs being 5v and MSP430 being <3.6v My launchpad drives the LEDs data line fine. Quote Link to post Share on other sites
RobG 1,892 Posted January 18, 2014 Share Posted January 18, 2014 According to specs, @5V and at full brightness, WS2812 draws 30mA. WS2812B is suppose to draw 60mA, but from my experience, it's <50mA. At 3.6V, your LEDs will be reddish, but they will still work. LP (G2553) can supply up to 250mA, so we are talking ~10 WS2812 or ~6 WS2812B at full brightness (in practice, you can do a lot more, but color is way off.) For the ring of 120 WS2812B, I would suggest 5V/6A, so you can light them all up and have consistent color. t0mpr1c3 1 Quote Link to post Share on other sites
Craig 0 Posted January 19, 2014 Share Posted January 19, 2014 Hi! Great project! I have been fascinated in RGB LED projects like this, although the project with the most RGB LEDs I built was the Aurora 18 mini kit I bought from here May I suggest something to add to the boards - mounting hole! Add a hole to the inner circle of each board, and we have 5 mounting holes, 72 degrees apart. Great for mounting on laser-cut acrylic using M3 copper/nylon screws. I don't have altium though, it would be great if greg could help me to do it. I will send the gerbers for fabrication after Chinese New Year holidays, If I receive 10 perfect e-test pass boards, I could send you a set for free (postage included) =) Oh, and perhaps the driver board would be an arc too, so it can hide behind the clock panel. Just thinking... Quote Link to post Share on other sites
greeeg 460 Posted January 19, 2014 Author Share Posted January 19, 2014 Mounting holes would be a good idea. I think I'll leave the rear of the ring blank. I'll let people decide what they want to use to drive it themselves. I'll mock up some simple boards that could mount directly to the ring. Eg the G2553 could easily drive the ring for a simple clock. But for complex animations a bigger part might work better. t0mpr1c3 1 Quote Link to post Share on other sites
jrychter 14 Posted January 19, 2014 Share Posted January 19, 2014 [i really need to do a writeup of my experiences writing a reliable driver for the WS2812 strips] First, I'd strongly suggest using level-shifting to drive the WS2812 at 5V. It might depend on your particular chip/strip, but I couldn't reliably drive them from 3.3V logic. The datasheet (crappy as it is) states CMOS logic, 0.7*VCC, which is 3.5V. You *might* get away with 3.6V, but I wouldn't count on it. For prototyping I used a 74HCT04 with two gates wired in series, any HCT chip will do. For production devices I'll probably use a M74VHC1GT125DT1G. Second, if you are writing code to drive the WS2812, I'd suggest testing it carefully using various bit patterns. I got best results using alternating 0x55 and 0xaa patterns, with some bytes being continuously and slowly incremented/decremented, in the middle of the string. There were many times when I thought I had correct code, only to see flickering somewhere down the string. Third, I don't really know why, but the WS2812 really does need 5V. That is a big disappointment, because you need complex power management in battery-powered applications, but that's just the way it is. bluehash and RobG 2 Quote Link to post Share on other sites
RobG 1,892 Posted January 19, 2014 Share Posted January 19, 2014 First, I'd strongly suggest using level-shifting to drive the WS2812 at 5V. It might depend on your particular chip/strip, but I couldn't reliably drive them from 3.3V logic. The datasheet (crappy as it is) states CMOS logic, 0.7*VCC, which is 3.5V. You *might* get away with 3.6V, but I wouldn't count on it... Hmmm, I have code that works reliably on LP-G2553, but not so good on LP-F or any 3.3V powered board. Will have to take this into consideration. ...Second, if you are writing code to drive the WS2812, I'd suggest testing it carefully using various bit patterns. I got best results using alternating 0x55 and 0xaa patterns, with some bytes being continuously and slowly incremented/decremented, in the middle of the string. There were many times when I thought I had correct code, only to see flickering somewhere down the string... And don't forget about input/output resistors, they are very important. Without them, any data line longer than few inches will suffer from reflection problems. jrychter 1 Quote Link to post Share on other sites
xxx1 35 Posted January 20, 2014 Share Posted January 20, 2014 @@greeeg Mounting holes is a excellent idea. If would suggest to put the driver on a separate board. I agree with @@RobG. Quote Link to post Share on other sites
t0mpr1c3 91 Posted January 20, 2014 Share Posted January 20, 2014 Mounting holes would be good. I would also prefer the driver on a separate board. Quote Link to post Share on other sites
t0mpr1c3 91 Posted January 21, 2014 Share Posted January 21, 2014 @@greeeg Do you have the final Gerbers? I am going with Elecrow, if that makes any difference. I am putting in an order for 50. There is one set of 5 PCBs left over, enough for 1 clock. Please sign up on the Group Buy thread if you would like them. First come first served! I can also accept last minute orders for up to 250 LEDs (WS2812Bs). Please excuse my cross posting, the group buy top post has got a little convoluted. Quote Link to post Share on other sites
greeeg 460 Posted January 21, 2014 Author Share Posted January 21, 2014 @@greeeg Do you have the final Gerbers? I am going with Elecrow, if that makes any difference. ledRing_2_02_D.zip Here you go. this version has a 3.2mm mounting hole in the middle. unfortunately there isnt enough room to accomidate the top of an M3 sized screw I have also chaanged the de-coupling capacitors to 0603 size from 0402. The space between the LED rings has also being increased by a few mm to accomidate the mounting hole. The overall board size is identical. RobG 1 Quote Link to post Share on other sites
t0mpr1c3 91 Posted January 21, 2014 Share Posted January 21, 2014 0402 is a bit small for me, I don't have a microscope. Is it possible to make the pads compatible with 0603? Quote Link to post Share on other sites
greeeg 460 Posted January 21, 2014 Author Share Posted January 21, 2014 Already done. I have also changed the de-coupling capacitors to 0603 size from 0402. I guess a better wording is "changed from 0402 to 0603" Quote Link to post Share on other sites
t0mpr1c3 91 Posted January 21, 2014 Share Posted January 21, 2014 derp Thanks, I'll order them in the morning when I'm a bit more alert. Quote Link to post Share on other sites
Fred 453 Posted January 21, 2014 Share Posted January 21, 2014 Here you go. this version has a 3.2mm mounting hole in the middle. unfortunately there isnt enough room to accomidate the top of an M3 sized screw You should be able to file the head of a nylon M3 screw to fit snugly in there between the LEDs. Quote Link to post Share on other sites
jrychter 14 Posted January 21, 2014 Share Posted January 21, 2014 I'm not sure if it will help in your design, but I've found that you can save some board space by going with milled cutouts for screws instead of holes. Yes, you still need room for the screw top, but less than you'd need if you had the hole in the middle of the board. See the attached picture for an example (the bottom cutout is marginal, parts are placed slightly too close) spirilis, dubnet and oPossum 3 Quote Link to post Share on other sites
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.