Page 2 of 2 FirstFirst 12
Results 11 to 19 of 19

Thread: Test WS2801 pixels

  1. #11
    Join Date
    Dec 2010
    Location
    Murrieta, Ca
    Posts
    554
    Post Thanks / Like

    Default Re: Test WS2801 pixels

    Just send them to me. I'll test them. You might not get them back though ;)
    5 houses using:
    Renard 64, 3 Ren24's, Ren8
    4 E682 pixel controllers
    2,300 WS2801 pixel nodes
    1200 dumb RGB
    Sequenced using LSP 2.5

  2. #12
    Join Date
    Nov 2008
    Location
    Springdale, Arkansas
    Posts
    113
    Post Thanks / Like

    Default Re: Test WS2801 pixels

    Using Zeph's recommendation, I got the pixels successfully tested with an Arduino. The test did produce some wonkiness that leads to more questions.

    1) I've experienced a large amount of flickering at times. At first, it was due to having my USB cable plugged in between the Uno and the computer. Serendipity allowed me to discover that was a bad idea. I've gone through several tests where each pixel was on in sequence. No problem there. But when turning on the whole string or a substantial portion of it at once, the whole string kind of blinks and then each pixel turns white and stops responding to the data. Any ideas there?

    2) I've read in some places where a capacitor is needed across VCC and GND. I tried that and got marginally better results with the flickering. I can push the strand further, but not all the way before the 'white reset' happens. Would a larger cap help? How do I know what size?

    3) Do I need to do anything to the downstream end of the wire?

    I have adjusted the data rate from 0-7 for the FastSPI library. No obvious differences there between tests.

    Thanks for the help.
    Last edited by eroberts; 04-04-2012 at 10:24 PM. Reason: add data rate clarification
    16 channels all Grinch (2008)
    48 channels, 32 Grinch + 16 Renard (2009)
    104 channels of Renard (2010)
    168 channels of Renard (2011)

    somewhere about 22,000 LED's

  3. #13
    Join Date
    Dec 2011
    Location
    Northern California
    Posts
    866
    Post Thanks / Like

    Default Re: Test WS2801 pixels

    Say more about your setup. How many pixels? How are they powered (separately from the Arduino I assume, with only ground connected)? Test the voltage and current being fed into the pixels with full on. Since digital stuff gets weird with flakey power, good power has to be verified before beginning to pull hair chasing gremlins.

    When you say "capacitor across VCC and GND", you mean the power to the pixels (independent of the power to the Arduino), right? A cap might help if the power line is so noisy that it's messing up the pixel chips, so it was worth trying, once you've verified the power is overall good. I think it's more likely the cap is just helping to partially mask another problem, which you'll need to find. The cap, if used at all, should be more like a final tweak to reduce error rates. You seem to be having more massive problems.

    If you have a scope, have a look at power and data lines. (I need to get one again one of these days, they were so handy).

    The only other things which occurs to me, if the power supplies are separated, is some kind of ground loop which is adding noise to the data lines, or an induced voltage drop over some line shared by data and power. For example, suppose a thin wire were used both to carry 3A of power (eg: on the ground lead) as well as serving as the ground for a data connection; and worse, suppose that current draw was rapidly changing, eg: due to PWM. In some cases, that might interfere with the digital communication. I can't see your setup, but these are some ideas of where to look, consistent with your issue of things going wonky as more pixels draw power.

  4. #14
    Join Date
    Dec 2011
    Location
    Northern California
    Posts
    866
    Post Thanks / Like

    Default Re: Test WS2801 pixels

    Quote Originally Posted by eroberts View Post
    Is there a beginners way to test WS2801 pixel strings (or just a single pixel)? It'll be a few weeks before I have time to build the E681 controller.
    Remember folks, Eric apparently has an unassembled e681 (as do I), so he doesn't need another controller - he just wanted to test his pixels before getting around to building the e681.

    If he doesn't get too distracted with the Arduino and forget to build the e681... :-)

    I'm interested because I have an Arduino, but haven't yet placed my pixel order (I'm waiting until I know I can get my e681 built in time, to avoid this very situation). But I'm gonna want to try the pixels on the Arduino anyway, so I hope to learn from Eric's experience. (And here he probably thought I had actual experience with Arduino + pixels; I'm just giving pointers to the results of my own research).

  5. #15
    Join Date
    Nov 2008
    Location
    Springdale, Arkansas
    Posts
    113
    Post Thanks / Like

    Default Re: Test WS2801 pixels

    Let me see if I can describe this concisely (I can diagram this if that will help):

    50 WS2801 pixels from DIYLEDEXPRESS pre-sale
    5V power supply from same sale
    Arduino Uno R3

    input side of pixels (the side with the IC for these strings) +5V (red) and GND (blue) are connected to +V and -V respectively on the PS
    ground is connected from PS to GND on Arduino
    input side of pixels clock is connected to pin 13 on Arduino
    input side of pixels data is connected to pin 11 on Arduino
    ceramic cap marked 68 across +V and -V of PS
    Arduino is powered separately with a 9V 500ma wall wart
    output side of pixels are bare wire (though not touching)

    PS measures 5.21 V under no load, 5.20 V under 50 pixel load
    downstream voltage is 3.14 V and last ~20 pixels show diminished color brightness

    I have no scope

    Hope that helps someone, though.
    16 channels all Grinch (2008)
    48 channels, 32 Grinch + 16 Renard (2009)
    104 channels of Renard (2010)
    168 channels of Renard (2011)

    somewhere about 22,000 LED's

  6. #16
    Join Date
    Dec 2010
    Location
    Murrieta, Ca
    Posts
    554
    Post Thanks / Like

    Default Re: Test WS2801 pixels

    I wonder if the voltage regulator on the arduino could be screwing things up? Why not get rid of the 9v and power the arduino from the same 5v supply that the pixels are on? See if that makes a difference.
    5 houses using:
    Renard 64, 3 Ren24's, Ren8
    4 E682 pixel controllers
    2,300 WS2801 pixel nodes
    1200 dumb RGB
    Sequenced using LSP 2.5

  7. #17
    Join Date
    Dec 2011
    Location
    Worksop UK
    Posts
    474
    Post Thanks / Like

    Default Re: Test WS2801 pixels

    I think I have read of a similar problem where the pixels went white and stopped responding. If I rember correctly it was due to volt drop in the pixel cable.
    You could try the following it will help faultfind anyway.
    Set all the pixels in the string (50) to a solid single colour at full brightness (eg red) this will only draw 1/3 of the total current. If all work ok add a second colour (eg red and green) this will draw 2/3 full load. This may or may not work if your problem is due to volt drop. If it works ok go for all 3 colours, if it is volt drop you should experiance the problem.
    If the problem occurred during the above tests you could then try.
    Put only the first pixel on white (all 3 colours full brightness) then with the second pixel put it on red then green then blue then white. If that was ok leave it on white (you now have the first and second pixels on white) and run through the colour test on the 3rd pixel and so on until the string fails to respond correctly.
    If the string failed at say 40 pixels it is possible that volt drop is the cause. Normaly 50 pixels strings don't suffer this effect. Firstly if the cable from the power supply to the pixel string thin you could try increasing it, secondly you could try bearing the 5v and 0v cables part way along the string and adding extra power at that point (don't let the wires short and so long as more than 25 pixels worked ok in the last test it could be after the 49th or 50th pixel if it has a cable).
    This should test for and cure volt drop related problems, normally a 50 string shouldn't suffer from volt drop but I have read reports of the connecting cable becomming very thin.
    But the best thing is to build the E681, my first one took about 5 hours as I read everything twice and reshaped the tip on my soldering iron about 10 times, the second one only about 2 hours, then you can do a real test.

    IMPORTANT If you add a second connection from the power supply to the pixels - check the polarity twice and use a fuse - some of the power supplies can supply 50 amps @ 5 volts, this is easily enough to melt the wires connecting the pixels if you get it wrong.
    Last edited by Barnabybear; 04-05-2012 at 10:20 AM. Reason: Note on checking polarity and fuse
    2 X E681 running 8 universes of 3001s as a 32s x 42p mega tree or a 96s x 14p matrix hanging from my guttering.

  8. #18
    Join Date
    Dec 2010
    Location
    Murrieta, Ca
    Posts
    554
    Post Thanks / Like

    Default Re: Test WS2801 pixels

    What kind of wire are you using between the Arduino and the pixels? I used SPT2 and had issues. Switched it out to a thinner communications cable and the problem went away.
    5 houses using:
    Renard 64, 3 Ren24's, Ren8
    4 E682 pixel controllers
    2,300 WS2801 pixel nodes
    1200 dumb RGB
    Sequenced using LSP 2.5

  9. #19
    Join Date
    Jul 2012
    Posts
    1
    Post Thanks / Like

    Default Re: Test WS2801 pixels

    I purchased 100 of the square modules with 4 5050s in each, plus a WS2801 chip in each. I'll put 45 of the modules in each of two Coro Flakes and be able to control each module separately thought the E681 controller.

Page 2 of 2 FirstFirst 12

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •