Lutron RS485 between devices?

[vc1234]

As part of my zwave issues investigation, I looked at the elk rs485 bus with a borrowed cheap logic analyzer to discover the serial signal properties.  It turns out that elk uses 38400 transmission rate with 9 bit no parity(it was pretty trivial to see with the analyzer).  The 9 bit part is tricky for a PC UART that cannot do 9 bits.  There is a trick of enabling parity and just substituting an opposite bit in the case of a parity error.  It looks as if what I see using the cable you have matches the LA bits.  But elk is quite chatty, and I am not sure what those bits may mean so far.

 

[wkearney99]

Which analyzer did you use?  I'm debating the purchase of a PC-based oscilloscope, like a Picotech.  
 
9-bits, ugh, that's annoying.  It's apparently possible to do that with a PC UART but not from most 'easily used' programming environments.
 
Still, it's interesting to know why:
http://electronicdesign.com/embedded/use-pcs-uart-9-bit-protocols
 
Likewise good to know fellow cocooner's have slogged through this jungle before:
http://cocoontech.com/forums/topic/18196-rs-485-connection-question/
 
Just when you think you're getting closer to making things easier, you discover another wrinkle in the mess.
 
From STFW it would seem the Arduino community has played around in this realm.  I picked up one yesterday, as it came in a kit with a bunch of other stuff I planned on hooking up to a Raspberry Pi.  Ugh, not ANOTHER branch of tinkering!  

 

[vc1234]

Which analyzer did you use?  I'm debating the purchase of a PC-based oscilloscope, like a Picotech.  

I used a Saleae analyzer.  I wanted to attach a screenshot of the serial capture, but I don't know how to do that.
 
I am going to buy an inexpensive Picoscope myself .  It would be useful to see the signal quality on the bus, not just  0/1's.
 
I also managed to capture a couple of Zwave packets using GNU radio and a cheap RTL DSR usb radio(~$10).  It's still work in progress because the packet slicer is more often than not thwarted by the  noisy zwave signal from the software demodulator although I can decode the resulting packet visually despite the noise
 
What I discovered so far with zwave:
 
1. Zwave packets controlling lights are not encrypted, so theoretically someone can easily discover your home network id and node ids as they are in clear text.  I did not check locks or thermostats (my main goal) as I am trying to achieve more or less reliable reception.
 
2. You can easily see packet spacing in time by just looking at the spectrum analyzer in any sdr software receiver or analyzing sdr capture offline.
 
3. My zwave network uses 9600 baud rate.
 
4. There is always a zwave ack packet from the switch.
 
5. I'd recommend against trying this sort of activity  because it's much more time consuming and less rewarding that say analyzing well formed digital signals.

 

[wkearney99]

5. I'd recommend against trying this sort of activity  because it's much more time consuming and less rewarding that say analyzing well formed digital signals.

 
Yes, but couldn't that be said about ALL of this automation fiddling?  Think of all the other ways we could be squandering our time...

 

[dkulesza]

Ok.  I'm curious in how to support this effort.
 
I'm also frustrated with how Lutron has handled their different product lines and I'm tempted to just rip and replace the whole setup.
 
Out of curiosity:
 
Have you looked at the option to re-flash RadioRA devices with the HomeworkQS equivalent?  I would think its possibility one of two ways:

• JTAG or similar on hardware
• Remote / Over the Air

 
I'm curious to learn more about the RS485 interface as well.  HomeworkQS communicates with the wireless devices using a hybrid/repeater through an RS485 link.  This encourages me as well to explore what options you have interfacing with devices.  
 
I wish I could get my hands on the homework qs software to see how hard it would be to whitelist the radiora2 gear.