DiyLightAnimation
Hardware => Lynx Smart String => Topic started by: tconley on November 10, 2010,
-
have you seen these rgb lights avalible at lowes and home depot. I wonder if we could adapt a controller to address them.
http://www.youtube.com/watch?v=7al1vIinhuo (http://www.youtube.com/watch?v=7al1vIinhuo)
-
I am Thinking the same way, I did pick up one set at Costco for $59.00
-
So, who's going to determine what chip they use? (non-destructively, of course).
The colors look very nice in the video. I'll probably be in Lowes this week and will look for myself. Perhaps a string might fall on the floor, cracking open a pixel. That would be a shame.
-
I purchased a set of the 50-light version at Lowes and taken the last node apart. The controller chip (in each node) is under a blob of epoxy so I can't tell what it is. The strings use 3-wire interconnects (+dc, gnd, data). I have scoped the data line but nothing conclusive yet. The strings run on 5vdc and are controlled by a 6051 micro in the box near the power block. I can't imagine these being too diffcult to reverse engineer for direct control from a different controller.
I will say they look very well constructed. The "globe" pops right off and does not affect the weathertightness (it has weep holes already).
-
If it's a three wire then it's likely a Manchester Encoding protocol. A popular RGB controller IC that uses Manchester is the CYT3005. Just FYI.
On the other hand, the same company that produces these lights also produces synchronized color changing plastic icicle lights. A set of 36 from Sears cost $99.99. I say this because I made my own by recycling plastic icicles from an old set and attaching 8mm RGB LEDs with WS2801 control chips to them. I then developed a controller using a PIC that translates DMX into the WS2801 protocol so I could control the exact color of each indivdual icicle. I've attached a picture of my icicles hanging in my cubical.
My cost for 56 homemade, DMX controlled icicles comes to about $40. Yeah DIY!
-
You're probably right about the encoding. I was probing at the end of the string and found it shooting out 27 pulses regardless of the operation being done (fade, color, etc). I wish I had more direct control of the operation of the strings but that is one of those chicken and egg things I guess. It was suggested in another forum that I try probing elsewhere since the 27 bits could be leftovers as opposed to the entire stream contents so I'll put in some pin probes to see if there's anything more obvious. I've included a pic of the top side of the node out of it's covers/silicone caulk.
This would not be a bad deal for right around $1/node (at least from Costco).
-
Interesting, I'm going to have to grab a set of these and do some playing. A have a set of 3-wire strings that use the CYT3005 chip. The intention was to figure out the protocol and incorporate it into my pixel controller, but I haven't made much headway in reverse engineering the protocol. Are you looking at this with a logic analyzer? It could be some other chip and there's no reason that 3-wire has to use Manchester encoding, itr could just as easily be a simple asynchronous data format. Looking at the data going into the last pixel may be more helpful. You should see some amount of data followed by the data that you are seeing now. Whatever data bits precede the last 27 pulses are the bits that are being grabbed by the last pixel.
The bit sequence you see coming out of the last pixel might be what tells each pixel that the data frame has ended and they should output their data bits to their leds. Do the bit values change or are they always the same sequence?
-
I've done probing at the end and at various points in the string and the data appears to be exactly the same (used both channels on my O-scope) for a given sequence and it's always 27 pulses long. The data itself differs for a given sequence. For what it's worth, the middle set of bits seem more active when performing a fade whereas the others seem more active with color changes. The pulses are one of two widths as well so at least on the surface it does fit into a possible Manchester type encoding anyway. I don't have a logic analyzer unfortunately. I know the PICkit2 can be used as one but does it require the PIC connection to work? I've never used it for that - something to look at I guess... I wish I had more control over the sequences though - that's where the benefit of the DMX front-ended widgets like you mentioned can really help since you control exactly what you want.
A question for RJ - - if you feel it's not appropriate to discuss reverse engineering these on your forum, please let me know and I'll refrain from doing so. I just wanted to be sure you were OK with this before discussing further.
-
I'm not reverse engineering anything. I likely can help you with this without doing anything wrong.
The chips on it could be anything. But likely they are dedicate to the system and do not send large amounts of data telling each what to do a step at a time.
They send a code of what effect to do and start and timing info.
The chips know what to do from there and just do it.
This is why the data you see does not change. Try a different pattern and see that it likely changes just a little.
This is why only 27 pulses.
RJ
-
Sorry - what I meant is that if you were OK with us discussing these here.
But yes - as I mentioned in the previous post, the data does change with different functions as evidenced by the differing pulse widths (indicating either 0's or 1's in some fashion) yet always with 27 pulses and that the same data is seen along the entire length. There was some conjecture that 50 separate pixel streams might be going out and each pixel would eat the next in-line set of data (sort of like Renards) but that does not appear to be the case. Just wish I had more control over the string versus the sequences pumped out by the MCU. I'm not planning to use these this year anyway.
-
it looks like that have done a lot of work on these in this thread
http://www.doityourselfchristmas.com/forums/showthread.php?p=134690#post134690
-
Hi pvplacelights,
I'm curious where you found your LEDs. Are they like the ones that have the chip integrated into the led housing?
Thanks
Rob
If it's a three wire then it's likely a Manchester Encoding protocol. A popular RGB controller IC that uses Manchester is the CYT3005. Just FYI.
On the other hand, the same company that produces these lights also produces synchronized color changing plastic icicle lights. A set of 36 from Sears cost $99.99. I say this because I made my own by recycling plastic icicles from an old set and attaching 8mm RGB LEDs with WS2801 control chips to them. I then developed a controller using a PIC that translates DMX into the WS2801 protocol so I could control the exact color of each indivdual icicle. I've attached a picture of my icicles hanging in my cubical.
My cost for 56 homemade, DMX controlled icicles comes to about $40. Yeah DIY!
-
Hi pvplacelights,
I'm curious where you found your LEDs. Are they like the ones that have the chip integrated into the led housing?
Thanks
Rob
The LEDs and the control chip (WS2801) are encased together in a waterproof housing. They are commercial signage lights, very bright, and resistent to harsh weather. I order them along with a DMX512-to-WS2801 controller from a supplier in China.
http://www.aliexpress.com/fm-store/701799/209889132-320389559/12mm-8mm-led-pixel-RGB-led-channel-lettet-waterproof-WS2801IC-256-level-gray-scale-DC5V-input.html (http://www.aliexpress.com/fm-store/701799/209889132-320389559/12mm-8mm-led-pixel-RGB-led-channel-lettet-waterproof-WS2801IC-256-level-gray-scale-DC5V-input.html)
http://www.aliexpress.com/fm-store/701799/209915969-340705002/DD-100-series-mini-DMX-protocol-decoder-support-SPI-IC5-IC8-protocol-control-signal-decoding-128.html (http://www.aliexpress.com/fm-store/701799/209915969-340705002/DD-100-series-mini-DMX-protocol-decoder-support-SPI-IC5-IC8-protocol-control-signal-decoding-128.html)
-
Wow that looks pretty wicked.
Thanks for the links. Something new to mess with...
Hi pvplacelights,
I'm curious where you found your LEDs. Are they like the ones that have the chip integrated into the led housing?
Thanks
Rob
The LEDs and the control chip (WS2801) are encased together in a waterproof housing. They are commercial signage lights, very bright, and resistent to harsh weather. I order them along with a DMX512-to-WS2801 controller from a supplier in China.
http://www.aliexpress.com/fm-store/701799/209889132-320389559/12mm-8mm-led-pixel-RGB-led-channel-lettet-waterproof-WS2801IC-256-level-gray-scale-DC5V-input.html (http://www.aliexpress.com/fm-store/701799/209889132-320389559/12mm-8mm-led-pixel-RGB-led-channel-lettet-waterproof-WS2801IC-256-level-gray-scale-DC5V-input.html)
http://www.aliexpress.com/fm-store/701799/209915969-340705002/DD-100-series-mini-DMX-protocol-decoder-support-SPI-IC5-IC8-protocol-control-signal-decoding-128.html (http://www.aliexpress.com/fm-store/701799/209915969-340705002/DD-100-series-mini-DMX-protocol-decoder-support-SPI-IC5-IC8-protocol-control-signal-decoding-128.html)
-
Follow up to the link with the WS2801 Controller:
Make sure that you get the device model that has programmable DMX start channel. The device model at the link does not allow DMX address setting and remains constantly at address 1. I discovered this the hard way.
Just FYI
-
I came across the following posts which provides more details on the LED lights.
http://hackaday.com/2010/12/01/hacked-led-christmas-lights/
-
Looks like someone alrady did a fair amount of reverse engineering on these (or similar) lights
http://www.deepdarc.com/2010/11/27/hacking-christmas-lights/
ldkennell