Date: Thu, 26 Aug 1999 16:12:07 +0100 From: Michael Searle <searle@longacre.demon.co.uk> To: "Louis A. Mamakos" <louie@TransSys.COM> Cc: freebsd-hardware@FreeBSD.ORG Subject: Re: performance of home automation hardware Message-ID: <19990826161207.52569@longacre.demon.co.uk> In-Reply-To: <199908231551.LAA59899@whizzo.transsys.com>; from Louis A. Mamakos on Mon, Aug 23, 1999 at 11:51:14AM -0400 References: <19990823153444.29882@longacre.demon.co.uk> <199908231551.LAA59899@whizzo.transsys.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Aug 23, 1999 at 11:51:14AM -0400, Louis A. Mamakos wrote:
>
> > What home automation hardware should I use/avoid
> > if I want it to be fast and reliable?
>
> You need to decide what it you want to control. This sounds sort
> of silly on it's face, but lighting control is probably the
> least interesting application of the home automation stuff I've
> done. So in my case, it's served "well enough" using X10 power
> line control stuff. I just make it a point to only have non-critical
> functions on X10. Please don't take this the wrong way; less than 5%
> of my X10 traffic doesn't work correctly, and some of that is poor
> implementation rather than unreliablity of the medium. Some of it
> is bugs in the computer interface implementation with back-to-back
> traffic on the powerline.
5% sounds bad, but X10 doesn't do collisions - is some of it
caused by this? (This can be worked around in a silly way by sending
everything through the computer = 1 transmitter only.)
You could also send commands more than once - for simple X10 at least,
where two ON's is the same as one. It wouldn't seem any slower except
for complex strings of commands, as normally the first one would work.
X10 devices whose state can be read could be read after writing to confirm,
although this may just introduce more errors.
The HCS2-X10 interface deals with this by keeping a copy of the state of
all X10 devices and refreshing them occasionally, or if the controller
is reset.
>
> > According to the web page though, HCS 2 modules
> > are slooow ('several seconds on large networks')
> > and unreliable, reading between the lines this is
> > because they continuously poll all devices on a
> > 9600bps net using a Z80. (Anything connected
> > directly to the controller is OK, but most modules
> > don't.)
>
> If you use their modules, but not their controller, and you
> have multiple RS232->RS422 (I think) interface, you could
> poll them in parallel to speed this up.
>
This would be great, and I've found a RS232->RS485 interface.
I wouldn't even need parallel interfaces - the slowness
is from the Z80 (which will now be a 486 PC) and the software
which continuously polls all devices. If I only check devices
when I want the results, the 9600bps line shouldn't be a
bottleneck (unless the half duplex RS485 is really slow to turn
around.)
--
searle@longacre.demon.co.uk
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hardware" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990826161207.52569>