Page 1 of 6 1 2 3 4 5 6 LastLast
Results 1 to 10 of 56

Thread: 18 Channel 10W Flood System

  1. #1
    Rudolph level member sporadic's Avatar
    Join Date
    Mar 2013
    Location
    Charleston, SC
    Posts
    279

    18 Channel 10W Flood System

    I've become a little obsessed with floods so am in the process of throwing this together... Main goal here is to drive many floods cheaply and be Renard protocol compatible with plans for DMX in the future. To keep cost and size down, everything except for headers and jacks are SMD with the smallest components being 0603. This may kill it for some DIYers, but I urge you to pickup the skill as its not really that hard and you can work much quicker. The main components are the FloodBrain (controller), FloodStick (driver inside each flood), and the retrofitted floods themselves.

    FloodBrain:
    Not being a PIC guy, the controller is designed around an AVR Xmega8E5. Initial firmware will support the Renard protocol with plans for DMX and wireless in the future. There is a header for NRF24L01 modules as well for future wireless support. Each FloodBrain can control 6 RGB floods, utilizing 18 channels. Power is supplied via a 12V 7A brick and distributed along with the control signals out RJ45 jacks. The +12V lines ride on pins 3&5 so you shouldn't fry any Renard-type SSRs if you accidentally plug one in. The FloodBrain is designed to fit inside a TA-200 case.

    Click image for larger version. 

Name:	FloodBrain-2013.09.06.jpg 
Views:	70 
Size:	189.6 KB 
ID:	1450 Click image for larger version. 

Name:	FloodBrain_Top-2013.09.06.jpg 
Views:	64 
Size:	116.6 KB 
ID:	1451 Click image for larger version. 

Name:	FloodBrain_Bottom-2013.09.06.jpg 
Views:	56 
Size:	87.8 KB 
ID:	1452

    FloodStick:
    Driver designed to fit inside a typical 10W LED floodlight housing. Utilizes the OnSemi NSI50350AS 350ma CC regulators switched via mosfet.

    Click image for larger version. 

Name:	FloodStick-2013.09.05.jpg 
Views:	70 
Size:	57.4 KB 
ID:	1453 Click image for larger version. 

Name:	FloodStick_Top-2013.09.05.jpg 
Views:	61 
Size:	48.7 KB 
ID:	1454 Click image for larger version. 

Name:	FloodStick_Bottom-2013.09.05.jpg 
Views:	51 
Size:	46.5 KB 
ID:	1455

    Flood lights:
    To keep costs down, we use 10W white floods and replace the LED with a RGB part. All other electronics are stripped and replaced with a FloodStick. Total cost is a few bucks less than a RGB flood and you end up with an extra 10W white LED to play with. I personally went with high temp white so I can use the extra LEDs for strobes later.

    Here are a few of the non-typical Renard parts and higher cost items so you can get an idea for BOM costs. Single quantities unless noted:
    Atmel Xmega8E5 TQFP - $1.83 - http://www.mouser.com/ProductDetail/...rS5%252b1a4%3d
    12V 7A Power Supply - $9.99 - http://www.ebay.com/itm/NEW-12V-7A-P...-/290810012160
    NSI50350AS - $.66 - http://www.mouser.com/ProductDetail/...qSldko3Q%3d%3d
    10W White LED Flood lights - $7.99 per in lots of 10 - http://www.ebay.com/itm/171118435011
    10W RGB LED - $2.28 per in lots of 5 - http://www.ebay.com/itm/181206449977
    Estimating around $10 for 1 FloodBrain board and 6 FloodStick boards

    All in all, I think it's a pretty cheap solution for a 10W x 6 RGB flood system. My only concern so far is that the regulators I'm using are spec'd at 350ma and the 10W LEDs are spec'd at 300ma. Every other RGB LED I've seen in 3W-30W range has been rated at 350ma with varying voltage so I'm hoping there won't be issues with over driving it. I can always cap the PWM in firmware as well if its an issue. Thoughts, comments? I'm planning to send the boards off for prototyping this weekend.

    Thanks!
    -shelby

  2. #2
    Rudolph level member sporadic's Avatar
    Join Date
    Mar 2013
    Location
    Charleston, SC
    Posts
    279
    Just noticed an issue with the FloodBrain board. You can't reach the RJ45 release tabs once the board is mounted in a TA-200. Not sure there is clearance to rotate them either as the connections would be too close to the edge and the board barely fits width-wise as it is. Will play with it a little more..

  3. #3
    Rudolph level member sporadic's Avatar
    Join Date
    Mar 2013
    Location
    Charleston, SC
    Posts
    279
    Had my edge clearances a little loose in DRC. Tightened them up and was able to rotate the jacks. Flipped channel colors around for nicer routing with new jack orientation and updated stick pinout as well.

    Click image for larger version. 

Name:	FloodBrain_Top-2013.09.06.jpg 
Views:	31 
Size:	340.4 KB 
ID:	1458

    Click image for larger version. 

Name:	FloodBrain_Bottom-2013.09.06.jpg 
Views:	27 
Size:	223.9 KB 
ID:	1459

  4. #4
    ScoobyDoobyWah budude's Avatar
    Join Date
    Nov 2011
    Location
    San Jose, CA
    Posts
    456
    Have you driven these over longer cable runs (say 30' or so)? Just curious how well the 3.3v drive from the processor will do for that case. Have you considered putting in RS485 driver/receivers or boosting the outputs to higher voltages to allow for more drop?

  5. #5
    Rudolph level member sporadic's Avatar
    Join Date
    Mar 2013
    Location
    Charleston, SC
    Posts
    279
    Quote Originally Posted by budude View Post
    Have you driven these over longer cable runs (say 30' or so)? Just curious how well the 3.3v drive from the processor will do for that case. Have you considered putting in RS485 driver/receivers or boosting the outputs to higher voltages to allow for more drop?
    I checked the voltage drop for 3.3v 20ma on a 100' run of 26ga wire and it came back w/ a 4.85% drop, 3.14V. I'm not an expert on mosfets, but if I'm reading the datasheet right that should be well into spec - http://www.mouser.com/ds/2/302/NX3008NBK-42386.pdf. Should I be ok? I don't have any parts on hand yet for real-world testing.
    thanks,
    -shelby

  6. #6
    ScoobyDoobyWah budude's Avatar
    Join Date
    Nov 2011
    Location
    San Jose, CA
    Posts
    456
    Probably OK - it was a concern back with DynamoBen's Prop controller (3.3v also) but it seemed to work OK for at least 10' lengths OK - never tried longer. You would want to make sure that the signals are across a Cat5 pair to help things out.

  7. #7
    Frosty level member
    Join Date
    Mar 2013
    Location
    Las Vegas, Nevada, U.S.
    Posts
    773
    Quote Originally Posted by sporadic View Post
    Just noticed an issue with the FloodBrain board. You can't reach the RJ45 release tabs once the board is mounted in a TA-200. Not sure there is clearance to rotate them either as the connections would be too close to the edge and the board barely fits width-wise as it is. Will play with it a little more..
    Had the same problem on one of my boards, they have rj 45s going at about every direction. You can get ones that have the plastic tab of the top Instead of the bottom by the pcb.
    http://www.jameco.com/webapp/wcs/sto...001_2150395_-1
    http://www.jameco.com/webapp/wcs/sto...001_2150740_-1

  8. #8
    Rudolph level member sporadic's Avatar
    Join Date
    Mar 2013
    Location
    Charleston, SC
    Posts
    279
    Thanks for the tip. I'll go back and check my pinout tonight. Right now distance is limited to around 20' due to voltage drop on the 12v lines. Anymore and regulators fall out of spec for blue and green. Plenty long for what I'm doing but will see what I can do to make it better

  9. #9
    Rudolph level member sporadic's Avatar
    Join Date
    Mar 2013
    Location
    Charleston, SC
    Posts
    279
    Ran across this http://www.onsemi.com/pub_link/Collateral/AND9008-D.PDF while trying to nail thermals down. Pretty sure I'll have issues on the red channel of the stick due to the large voltage drop in the regulator. Too much dissipation at 2.3W. Could go to the DPAK version of the NSI50350 and a larger board, stuff a big honkin 2W resistor in series on the red channel (nasty) or migrate to constant current bucks and drive everything from a 19V laptop power supply which was my original plan. Wanted to stay with the NSI parts though to keep costs down. Jumping to a 19V supply would also mean switching the brain linear regulator over to a buck smps as well as the linear part would be on the edge of its specs. Plus side is the CAT5 control/power cables could be substantially longer since it'd be carrying a higher voltage at a lower current. Arrrggghh, decisions!!

  10. #10
    Junior Elf level member
    Join Date
    Aug 2013
    Posts
    10
    I see that you can parallel time to increase allowed power and increase the max from the 350ma. I would be interested to see if we can get 700ma out of each channel instead, what do you think? Has any testing started yet? I am really, really interested in this project you have going. Looking at some changes for the upcoming year and thus would be ideal for it if there is a way to get at least 700ma/channel

Page 1 of 6 1 2 3 4 5 6 LastLast

Posting Permissions

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