Theyâre so tidy that wayâŚa nice contrast to my desk at the moment.
Anyway, I hate the idea of cutting themâŚI will eventually, since I have a project in mind that wonât work without cutting, but for the moment, Iâm keeping them in one strip (per reelâŚI got two reels).
The code was written for NETMF 4.1, so I had to tweak it a little for 4.2, since it appears that some of the SPI stuff got moved around and a few of the constructors changed, but it only took me a few minutes to get it working, [em]once I realized that plugging everything into the output wonât work[/em] no matter how good the code is.
But did not get as far different chip Texas Instruments TLC5947 but I think the same sort of functions. I hope you do a nice write up on this as I gave up on this more confused than before I started.
I understand the theory of SPI but I am going back the the 101 stuff
Donât assume that I understand it yet. Thankfully, Dave Durant posted code that works on 4.2 with only minor tweaks, so I didnât really have to dive into the bowels of SPI to get the blinkies working. At some point, Iâll do a deeper dive on his driver class, and try to understand how it works. But for now what really matters is having Blinkies for MADExpo.
@ devhammer - glad you like them. Have fun next week.
Iâve been busy prepping for a local digital media festival (http://www.woodstockdigital.com) that kicked off today and goes most of the weekend. Weâre turning a lot of local artists, makers, educators, and tinkerers into microcontroller maniacs
For those patiently waiting for their blinkies - thank you. Once I finish spreading the local micro mojo, Iâll go back to being a postman.
For those who wished they had some - weâll be doing another order at some point.
BTW - Adafruit has a great tutorial on how to properly cut these.
@ ransomhall - If you do another order, please drop me an email in case I miss the postâŚI might well want to get some more if you can get the same deal.
Dave Durantâs demo program allows you to choose the number of LEDs to control while the program is running. I think it defaulted to 64, so thatâs why only part of the strip is lighting up. At full white, the entire 5m strip could draw as much as 10A, so given that I was powering off the mainboard power supply, I didnât want to risk frying anything.
This morningâs project is converting a PC PSU over for use as a bench supply (adding posts for 12v, 5v, and 3.3v, and a switch), which should give me up to 22A to play with (hopefully enough to power the 2 5m strips I have).
There are some minor differences between them. I opted to put in a switch and a power LED (though for some reason my LED blinks when the unit is powered upâŚnot sure whatâs up with that). I ended up with only 5v and 3.3v, because I cut the 12v lines back to the PCB, before I realized what I was doing.
A couple of tips:
[ol]Leave more slack in the wire than you think youâll need. I cut my ground wires too short and ended up having to pull them tighter than optimal to make the connection.
Make sure that you leave plenty of space around each of the binding posts inside the case, as youâll be hooking up 2-3 ring terminals to each, and it gets crowded fast.[/ol]
Just tested the FrankenPSU, and it successfully powered two 5m reels, for a total of 320 RGB LEDs. Sweet!
[quote=âdevhammerâ]âŚso I didnât really have to dive into the bowels of SPI to get the blinkies working. At some point, Iâll do a deeper dive on his driver class, and try to understand how it works.
[/quote]
Trust me, once you look through the code and get it, youâll be all âthatâs it? really??â
The biggest help for these particular strips might be to get more details on the interface. All I could find for documentation was that you basically send the bytes for all LEDs, all the time, plus a few protocol bytes. For effects that donât require all LEDs in the strip to change, especially on longer strips, it seems like itâd be a lot more efficient to have a way to just address the ones you want to change. Then again, SPI may be fast enough that itâs not really worth doing that. Dunno. Maybe ransomhall or GHI knows somebody that knows somebody who could provide more insight hereâŚ
Trust me, once you look through the code and get it, youâll be all âthatâs it? really??â
The biggest help for these particular strips might be to get more details on the interface. All I could find for documentation was that you basically send the bytes for all LEDs, all the time, plus a few protocol bytes. For effects that donât require all LEDs in the strip to change, especially on longer strips, it seems like itâd be a lot more efficient to have a way to just address the ones you want to change. Then again, SPI may be fast enough that itâs not really worth doing that. Dunno. Maybe ransomhall or GHI knows somebody that knows somebody who could provide more insight hereâŚ
[/quote]
Would be cool to get a better understanding of how it works.
For example, as noted, I got enough current from the FrankenPSU to drive 2 full reels, but it was pretty apparent that using 320 LEDs all at once had a significant negative impact on performance. Things are noticeably slower than when using just one reel, or part of a reel.
I also think I might add a button or some other means of switching the number of LEDs directly from the mainboard, as changing the value and redeploying is kind of inconvenient (not complaining about your code, just thinking about how I want to use it, demo-wise).
If anyone with more knowledge of the RGB LED strips would care to comment on ideas for improving performance when driving lots of LEDs, Iâm all ears.