Re: Ethernet-based systems for DIY?

[Marc_F_Hult <MFHult@xxxxxxxxxxxxxxxxxxxxx>]

On 9 Dec 2006 16:56:33 -0800, "Don Stratton" <anguishandfear@xxxxxxxxx> wrote
in message  <1165712193.783352.104350@xxxxxxxxxxxxxxxxxxxxxxxxxxx>:

>I have been carefully considering what I would like to see in total
>home automation for a long time.  it can best be described as Star Trek
>technology; I want to know pretty much everything going on in my home
>and I want a high level of computing assistance (voice responce,
>internet interactivity, intelligent alarm systems that correlate
>multiple sensors to evaluate threat and response, tons more).
>
>Needless to say I have not found it.
>
>I can find pretty much all of it in bits and pieces, but nothing that
>would logically stitch together very well.  Systems like X10 are weak
>in security, alarm systems are weak in home automation, media systems
>ignore both, very little of it embraces a do-it-yourselfer, most
>systems don't scale or intergrate well with each other, and good
>software is absolutely non-existant.  Which brings me to my point.
>
>_IF_ I am right and nothing "good" exists, I may be interested in
>building it myself.  I say this both as an interested homeowner who's
>handy with a soldering iron, and as the owner and senior engineer of a
>company that makes consumer electronic gear.  I did a quick block
>design of a data aquisition module I thought would be a good idea
>(multiple A/D, multiple GPIO and video inputs with ethernet out back to
>the base), checked the net to find nothing quite like it, and am
>intrigued by the idea of developing it into a full-fledged commercial
>product.  My thinking is that systems like this should be more cellular
>in nature, and this DAQ would seem to allow for all home security
>sensing needs, some or all of home monitoring needs, easy as hell for
>DIY installation and integration into existing sensor systems, and
>absolute expandability in the form of ethernet output.  VERY simple
>software would allow you to do a lot, very complex software boggles my
>mind in terms of sheer potential (the Star Trek stuff).
>
>Am I wrong?  Can someone recommend what they think are good systems and
>components?
>
>
>--Don

Excellent hardware is readily available to do everything you want to do and
much more. Complex software is available that does some of what (I think) you
want.

Have you looked into:

AMX www.amx.com and Crestron www.crestron.com ?
HomeSeer www.homeseer.com  and Charmed Quark Systems www.charmedquark.com?
Elk M1G www.elkproducts.com and Omni www.homeauto.com ?

-- to name a few of the more popular competing pairs of products in three HA
product categories (namely, proprietary 'total' HA systems, HA DAQ panels and
PC-centric HA control software).

I suggest you:

1  a) Draw a picture/diagram of actual physical locations of where you want I/)
(environmental sensors, cameras, security devices, input screens, user
controls) and then
   b) explain to yourself _exactly_ how you are going to route the information
from those I/O points to the DAQs. Don't bluff. Details matter. If you choose a
hard-wired (physical wiring) approach, you will most likely have a very robust
system but may spend a lion's share of the effort physically installing wire
especially if a retrofit. If you chose RF, IR and(or) powerline controls (or
some mix) you should be aware of cost/reliability trade-offs that you may or
may not want to make.

2) Diagram the control _software_ that would be required to do everything you
want to do.  Estimate the number of man-years it might take (and or investigate
existing software that can do many/most of the things you mention and can serve
as the basis for your own software extensions). Who is going to do this? Of
what value is your hardware if it isn't available?

+++++++++

My experience is that there are two broad areas in which existing home
automation offerings is in weak: 1) monitoring (not just reacting to 'events')
including subsequent storage, retrieval, presentation, and
interpretation/analysis of data, and 2) systems modeling that provides
decisions based on calculated predictions of system behavior. (not just simple
IF-THEN_ELSE 'events' handling).

HTH ... Marc
Marc_F_Hult
www.ECOntrol.org