Jump to content
Light-O-Rama Forums
Scooch

Any ideas??? Pixel colors not correct

Recommended Posts

I have finally gotten my matrix to work. I am having a problem however. I am having bands of color that are not what they should be. If I shift the position of the image, the pattern that the band of bad color changes and moves. It is not in a consistent position of the image. Also, when doing certain effects like the butterfly, the colors are not in correct positions. 

 

Share this post


Link to post
Share on other sites

Could be Power issues or data integrity issues.  Some pixels operate at data rates around 800KHz.  Pulse widths around 1.3us.  

What is the distance between the controller and the first pixel of each string?   Using NULL Pixels?

Are you injecting power?  Measure the voltage on the pixels furthest from where power is fed (for each string) with the LEDs set to white.

Share this post


Link to post
Share on other sites

Thank you for the reply. The distance is about 4 feet. They are 5v pixels and powered from each end of the strand. Rack strand is 84 pixels long. Given the short distance I have not used null pixels. The interesting thing is that when using the pixcon for the test sequence all colors are correct. I can cycle through all the colors on each string while using the hardware utility. The pattern the is off in the butterfly effect would make me possibility thing a config issue but I have no idea where the problem is. 

Here is a link to a video of the butterfly effect. 

 

Edited by Scooch

Share this post


Link to post
Share on other sites

How many strings are connected together ? are you using a zig -zag pattern?

Share this post


Link to post
Share on other sites

Yes on the zing-zag. Each channel is 2 horizontal rows. 42 pixels in each direction. 

Share this post


Link to post
Share on other sites

So 24 rows at 42 pixels is about 60 amps.  How many power supplies are you using.

Start with the easy stuff.  Set the matrix to all white and take voltage measurements.  Nothing should be less than 4.8 volts. 

 

Share this post


Link to post
Share on other sites

Looking at the video, I would make the slightly educated guess that the channel assignments in the simulation does not match the channel assignments in the actual matrix.  You did not specify much of what the hardware is, and it may help us troubleshoot the problem.  The more detail on the configuration, the better.

Let's see if I put the pieces together.  The matrix is 24 rows by 42 columns.  What you refer to as a row in your previous post is shown as up and down in the video.  Because of the channel counts, I am assuming this is not LOR pixels, and is most likely being driven as E1.31 (but of course the Pixcon can use either LOR networking or E1.31).  You are using 12 outputs from the Pixcon controller that you are using.  Each output is driving two rows, with the data going to for example an odd row in one direction, and continuing back on the adjacent even row in the opposite direction (also known as a zig-zag pattern).  You are feeding power at both ends of every row.  You are using a Pixcon controller, but you did not specify it's configuration.   Did I get that much right?  Please fill in the details on controller configuration, universes (if e1.31), or LOR Unit IDs (if using LOR networking).  How did you specify the configuration in PE - i.e. input from a visualization, or created in PE itself?

 

Share this post


Link to post
Share on other sites

OK. Sorry for not including everything. I am running as LOR. The config in PE was set up through the PE not imported. The video was taken from what is the side of the matrix so the dimensions are 42 wide by 24 high. Everything else that you stated is correct. I am powering the matrix and controller using 2x300w 60A power supplies. They feed the controller directly and distribute power to the matrix via a multi post bus bar. The power and data are conducted through 14G wire to the matrix. 

Share this post


Link to post
Share on other sites

I don't see where you are injecting power on the end closest to the camera.  

I bet you get at least 0.5 volt drop from one end of the 42 pixel string to the other with all of the pixels on.

Share this post


Link to post
Share on other sites
15 minutes ago, wbaker4 said:

I don't see where you are injecting power on the end closest to the camera.  

I bet you get at least 0.5 volt drop from one end of the 42 pixel string to the other with all of the pixels on.

All power is fed into the otther end of the matrix. I'm feeding power into the start and end of each channel. 

Share this post


Link to post
Share on other sites
2 minutes ago, Scooch said:

All power is fed into the otther end of the matrix. I'm feeding power into the start and end of each channel. 

Confirming this.  As shown in the video, the power is fed from the beginning and end of each string (the far side in the video), and the near end in the video is just the loop around (if you want to call it that) end.  Therefore, power is fed no further than 42 pixels from any given pixel.  That should be fine.  Your total power supply capacity should be plenty.  The issue you are having, I would not expect to be caused by low voltage anyway.

I've never worked with a PixCon, but if you can post a screenshot of the configuration, it might help. 

I'm still going to go with a mismatch between the channel configuration in the actual matrix vs the channel configuration in PE as my suspect...

 

Share this post


Link to post
Share on other sites

" can cycle through all the colors on each string while using the hardware utility "

I assume that pattern is static - certain colors on, followed by another color, etc.  No animation like your butterfly pattern?  And it works?

I would consider cutting the distance your data signal is fed to the matrix in half to see how that changes the butterfly pattern.

Your pixel array is about 14 feet long?    

 

Share this post


Link to post
Share on other sites
4 minutes ago, wbaker4 said:

I would consider cutting the distance your data signal is fed to the matrix in half to see how that changes the butterfly pattern.

He said it's only four feet.  That should not be a problem even with the most picky of pixel types.  BTW, what type of pixels are they?

 

Share this post


Link to post
Share on other sites

DevMike put up a Pixcon checklist, you may want to go through it and check your config:

  • 1.  Some pixels have a front and a back. Ensure you know which side data goes IN to.  Pixels from LOR can accept data from either end, are plainly marked, or are pre-wired.
  • 2.  Pixels have a voltage. Ensure you have the proper power supply.  LOR Pixels are 5v.  If your pixel package from LOR included a power supply, you do not need to worry about this step.
  • 3.  Pixels have current draw (amperage). Ensure you are not exceeding the amps per channel/power supply/etc.  Pixels from LOR will draw 3A per 50 pixels at 5V (15 Watts).  If your pixel package from LOR included a power supply, you do not need to worry about this step.
  • 4.  Wire power supply to PixCon16, if not already wired.
  • 5.  Remember – never connect or disconnect anything from PixCon16 with power applied.
  • 6.  Power up board. Follow instructions in DMX and E1.31 for pixel control document if needed.  Open Network Preferences/Find Configure PixCon16/Click on PixCon16 in list, and open the PixCon16 configuration window.  
    Only work with a SINGLE Pixcon16 connected to your network at one time until properly configured -- IP address conflicts can occur if you attempt to configure 2 or more.  
    Do not continue until you are able to reliably talk to the board and configure it.
  • 7.  Ensure board firmware is AT LEAST 1.4.8. Do NOT downgrade if higher. Do not continue until firmware is updated to at least 1.4.8. Instructions are found in the manual.  If you have the LOR Control Panel loaded, UNLOAD IT before attempting to update firmware.  Updating the firmware while the Control Panel/Comm listener are running can cause the board to FAIL.
  • 8.  Properly set up Pixel Type and Pixel Speed in configuration. Ignore all other parameters.
  • 9.  Properly wire 1 pixel string to 4 wire connector. Follow pin-outs in the manual.   Note -- Some pixels do not have/require the clock signal.
  • 10.  Connect pixels to a single port, power up. Pixels may briefly flash. Pixels that stay ON indicate incorrect pixel type or speed selected on #8
  • 11.  Run hardware test from manual using buttons on the PixCon16. If pixels do not work, problem is most likely #8. Go back there. 
    Do not continue until the Hardware Test works properly.
  • 12.  Wire the rest of your pixels to the 4 wire connectors. Connect them to the PixCon16. Run the hardware test again. Ensure all pixels are working correctly. Pixels that are not working indicate a problem with the pixel itself, or an under-rated power supply.
    Do not continue until the Hardware Test works properly for ALL pixels.
  • 13.  On the first tab of PixCon16 configuration, ensure you have the correct option specified for J3/J4. If you will use the board in ELOR mode, turn this option ON. For E1.31 ensure this option is OFF.
  • 14.  On the Second tab set up each port of the PixCon16 properly. You can use simple mode or advanced mode. All options are documented in the help file for the software or in the manual for the PixCon16. Remember, for E1.31 universes must also be set up in Net Prefs (the software will also offer to do it for you), for ELOR, 500K speed and Enhanced Light-O-Rama network (done in step 16). It is best to write down the settings.
  • 15.  Ensure you have the LOR Control Panel loaded (which will load the LOR COMM Listener).  The COMM Listener is ALWAYS required to control pixels.
  • 16.  Properly configure your network preferences.
    • E1.31: Ensure Universe matches universe on the board & Set IP address correctly (multi or unicast- specify).  It is easiest to use multicast.
    • ELOR: Ensure the COM port is set to at least 500K and ELOR mode is turned ON.
  • 17.  E1.31: Do not disconnect anything. 
    • Both LEDs should be SOLID ON. 
      Do not continue until they are.
    • ELOR: Disconnect Network cable, ensure jumpers on PixCon16 for J3/J4 are set to LOR. Connect HS USB adapter to computer, CAT5 from USB adapter to J4 of Pixcon16. 
      Both LEDs should be SOLID ON. 
      Do not continue until they are.
  • 18.  Start Hardware Utility. If asked if the HWU should take control of a port say NO. Open the pixel console. Set Pixel Console params correctly and test pixels with sliders.  
    Do not continue until you can successfully control pixels.
  • 19.  Your board is now ready to use.

Share this post


Link to post
Share on other sites

I'm with Jim on this one, is as to be something in the setup in PE.  The pattern not lining up like that usually doesn't have to do with power issues or data length.  I've seen this issue on my matrix and it has always been the way it is set up in PE and/or on the Pixcon16.  I used the custom option when setting up my matrix instead of trying to use the zig-zag on the controller set up. (easier for me) I only told the pixcon16 how many pixels per output and how many null pixels.  In Excel I numbered the pixels the way they are built on my actual matrix and copy and pasted into the custom field in the prop definition.  Make sure on the right hand side you change the max channel to 510.  Here is a screenshot of my set up. Prop setup.jpg

Edited by Little_b

Share this post


Link to post
Share on other sites

Not sure what distance the PIxcon can handle without NULL pixels.

SanDevices E682 can handle about 15 feet without any NULL pixels - or you can lower the pull up resistor values on the drivers.

...also - about a second before the video ends (if you can freeze it at that point) it looks like he has a perfect butterfly pattern, before it turns to mush again.

Edited by wbaker4

Share this post


Link to post
Share on other sites
3 minutes ago, wbaker4 said:

Not sure what distance the PIxcon can handle without NULL pixels.

Depends mostly on the technology of the pixel.  Of the commonly used types, 2801s are the shortest, 2811s are longer, and GE Color Effects are quite long distance.  Since you have not advised what type pixels they are, we are guessing.

 

Share this post


Link to post
Share on other sites

Here is another video showing the problem. Notice that in the blue band the dark areas for the eyes are still in the correct position even though the color of wrong. Oh, I am running 2811 pixels. 

 

 

Share this post


Link to post
Share on other sites

One other thing, are all your power supplies connected together at the V-?  If not, they need to be. 

Edited by Little_b

Share this post


Link to post
Share on other sites

That pretty well eliminated a channel mismatch. As I've never even seen a Pixcon, and don't use Pixel Editor, that pretty well leaves me out of this. I will be watching to see what is determined however.

Good Luck!

Sent from my Droid Turbo via Tapatalk, so blame any typos or spelling errors on Android

Share this post


Link to post
Share on other sites
3 minutes ago, k6ccc said:

That pretty well eliminated a channel mismatch. As I've never even seen a Pixcon, and don't use Pixel Editor, that pretty well leaves me out of this. I will be watching to see what is determined however.

Good Luck!

Sent from my Droid Turbo via Tapatalk, so blame any typos or spelling errors on Android

Thank you. 

Share this post


Link to post
Share on other sites

It may be you camera (phone) but I do notice a difference in intensity (brighter, no corruption) when fewer rows are lit.

At about the 6 second mark, there is one lone pixel that stays on at the far end, center.  After 2 rows go by, it goes off.

Try running other patterns that run from top to bottom (as viewed from the camera).

Edited by wbaker4

Share this post


Link to post
Share on other sites

Now that I watched the videos a little more closely I see similar issues I had this morning when I put mine together.  Most of the "glitches" are caused from improper grounding (V-).  Make sure all your Power Supplies and Controller are connected together at the V-.  Don't connect the V+ together or it will increase the voltage.  

Will you take a pic of your connections to each strand? It looks like your power injection may not be correct.  If you're using 2 power supplies, does each one power half the matrix?  Are you connecting a power supply to each side of the Pixcon 16?  I took the power to the pixels directly instead of going through the controller.  This made it easier to inject power to each strand.  Mine are 126 pixels long (63 down 63 up) and I have power going to each end.

 

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...