Re: Looking for Temp and humidity sensors

["Robert Green" <ROBERT_GREEN1963@xxxxxxxxx>]

"Marc F Hult" <MFHult@xxxxxxxxxxxxxxxxxxxxx> wrote in message
news:ttv4s1lpdlv29h6e5ms70c9hq4353u1o2p@xxxxxxxxxx
> On Sun, 8 Jan 2006 17:49:45 -0500, "Robert Green"
> <ROBERT_GREEN1963@xxxxxxxxx> wrote in message
> <cZednZkgjNcqBVzeRVn-gQ@xxxxxxx>:
>
> >"Marc F Hult" <MFHult@xxxxxxxxxxxxxxxxxxxxx> wrote in
> >
> ><stuff snipped>
> >> Some very accurate temperature sensors (a silicon diode for example)
cost
> >> 1/1000th of what I'd speculate that a  HAI temperature module costs
And
> >> calibrated thermistors, Dallas 1-wire devices, and LM34/35's cost on
the
> >> order of US$1 and so enjoy about a 40-fold cost advantage compared  to
> HAI
> >> (and Adicon and ELk and Aprilaire ) temperature _modules_.
> >
> >Why *IS* that?  I was dismayed by the price and the restricted function
of
> >the ADI "Bobcat" line of temperature and humidity sensors.  I see that
> >Worthington has some very cheap sensors listed with their ADI, HAI and
ELK
> >offerings.  I would ideally line to monitor temperature and humidity info
> >for perhaps 20 different points in the house, with a wide array of
> >temperatures, too (chimney temps, hot water heater, fridge, etc).
> >
> >Monitoring temperature and humidity (as well as minute barometric
pressure
> >differences between rooms) from a lot of different locations *easily and
> >cheaply* would be a very good way of making sure that everything in the
> >house was functioning normally.  With Bobcats, that will run me about
> $2500.
> >:-(  We'll see how well Jeff Volp's analog I/O -solution works for me
when
> >I get around to ordering the equipment.  That depends on how well I
> >predicted Christmas and travel related expenses.
>
>
> I suggest that you close your pocketbook for now.
>
> Let's back up a bit.
>
> The Compleat Home Automation Controller needs to cause actions based on
> environmental thresholds (aka "events", "triggers") *AND* be capable of 1)
> measuring, 2) interpreting (performing calculations in order to define an
> event) 3) recording (storing) and 4) reporting continuously variable
> parameters such as temperature, pressure, illumination level and so on.

Certainly.  But in the design phase some would consider it important to
begin taking readings as soon as possible.  I'm sure you're aware one of the
biggest problems in weather prediction is the lack of accurate and fairly
granular data for all but the most recent human epoch.  When I get around to
considering the CPU of a home automation system, I already want to have a
good feel for some of the raw data that will be feeding it.

I've got 6 different places in the house monitored with RatShack remote
thermo/hygros, but that data just evaporates with no suitable way to collect
it.  I'm willing to devote an old PC or laptop to the task of data logging
since I can see that doing with with the Elk, Ocelot or HAI boxes is going
to be a hack, and probably an unsatisfactory one at that!

Since the system Jeff described would be inexpensive and fairly easy to
implement, it doesn't seem like a big risk to me.  After all, this is a
hobby.  If it wasn't this it would be titanium golf clubs or tropical fish.
:-)  It also had the attraction that the SECU unit was something I intended
to acquire for the Ocelot anyway.

> Most of the current crop of HA controllers and programs to a credible job
at
> event-based control, and some are adequate for many purposes at
measurement
> and using those measurements to define event thresholds.
>
> But despite whatever comp.home.automation's well-intentioned "brand
> advocates" might lead folks to believe, the devices known in these parts
as
> Home Automation Controllers (Adicon Ocelot, HAI, Elks etc) are mediocre or
> useless for general-purpose monitoring purposes for a variety of reasons.

Well, that's certainly an intriguing analysis.  I would imagine you'd get a
lot of static with a comment like that were it not buried deep in reply like
this!  <g>

> The answer to your question of why these sensors cost so much lies in larg
e
> part in that fact. If one needs only a few sensors, the add-on modules may
> be cost effective and have adequate performance for many purposes.
However,
> as you have found, this approach becomes prohibitively expensive as you
add
> more modules.

It's apparent one would use the Bobcats for something discrete and
important.  It's also pretty obvious that Ocelots aren't the best devices to
use for critical temperature or humidity readings for a large number of
locations.  The fact that you can even log and act on temperature/humidity
data with an Ocelot (or Elk or HAI box) at all speaks volumes of the
"plastic" nature of modern microprocessors.

 > If one compares the capabilities of today's HA controllers with even
> 20-year-old environmental data loggers one is immediately struck by:
>
> 1) Inadequate signal resolution (and in the case of some, precision) of
the
> analog-to-digital converter (Adicon/Ocelot, Elk)  or no analog input (HAI)

What do you consider "inadequate?"  Is it quantifiable in terms of errors
and types of errors?  What I am trying to get at is that even if an Ocelot
turns on a fan at 78F rather than 80F, does that lack of precision really
matter?  It may not matter to most end users whether they can measure with +
or - 1% or less accuracy.

> 2) Inadequate data storage/retention.

Again, most HA enthusiasts don't seem to want to retain data, and you
correctly point those that do to:

> There are excellent, inexpensive solutions.
> Do check out www.phanderson.com for starters.

> The key to good analog measurement is proper signal conditioning. I will
> post the design for the conditioned four-sensor pcb where I posted the
> 16-channel analog input multiplexor a couple of weeks ago.
>
> http://www.econtrol.org/analogsensors.htm
>
> I use a previous version inside commercial motion detectors to provide
> temperature, light level, humidity and motion detection in a single
device.

I'm not sure what you mean.  Did you take a commercial motion detector and
retrofit it to transmit information about three other "states" as well as
motion?  That sounds fascinating and like an emulation of nature.  Take the
ear.  Great for sound processing AND also provides information about air
pressure, gravity, acceleration and body orientation.

> With the addition of a Dallas 1-wire DS2450 Quad A/D converter
> http://www.maxim-ic.com/quick_view2.cfm/qv_pk/2921
> the data can be sent digitally or as analog signals. The DS2450 can
provide
> up to 16-bit resolution which makes use of high-precision, inexpensive
> thermistors practical.

An interesting device, indeed.  The 1-Wire bus seems to be the way to create
an inexpensive but highly accurate sensor net for a house.  Are these specs
they list for the DS2450 really right?

"Very low power: 2.5µW active, 25µW idle"

--
Bobby G.