Re: Refrigerator monitor ideas?

["Robert Green" <ROBERT_GREEN1963@xxxxxxxxx>]

"Marc_F_Hult" <MFHult@xxxxxxxxxxxxxxxxxxxxx> wrote in message
news:ea37s2dv2rfvuerdu9oetv2daa6mouai86@xxxxxxxxxx
>
>
> On Fri, 2 Feb 2007 10:34:56 -0500, "Robert Green"
> <ROBERT_GREEN1963@xxxxxxxxx> wrote in message
> <yZadnSuqcufcxl7YnZ2dnUVZ_qWvnZ2d@xxxxxxx>:
>
> >We've got a 30 year old White-Westinghouse refrigerator/freezer in the
> >basement that's cosmetically shot, but still humming away, delivering
> >subzero temperatures in the icebox without major incident for years.
> >
> >Several times in the past 10 years or so, the drain tube from the pan
under
> >the freezer coils (it's an upright with a top freezer) has clogged,
leading
> >the coils to become caked in ice.  When this happens, the first clue is
that
> >the freezer temperatures begin to rise.  Other than placing a sight glass
in
> >the freezer floor pan, what are some simple options to detect a coil
> >freeze-up condition.  I had previously had this on a yearly inspection
> >schedule but the addition of a number of shedding dogs to the house in
> >recent months has made that too infrequent a cycle.
> >
> >What I would really like to do is design something simple that can catch
the
> >formation of serious ice long before the coil became caked.  Simple
> >temperature sensing is inadequate because the self-defrosting feature
causes
> >a sharp upward swing every night.
> >
> >The conditions that occur during a "freeze-up" (in addition to an overall
> >increase of average freezer temperature each day it's building up) are
that
> >very little water reaches the evaporation pan at the bottom of the
fridge.
> >But checking that pan is about as inconvenient as unscrewing the drain
tube
> >connector that runs from the center top of the refrigerator to the
backwall
> >where it meets a drain tube that empties into the floor pan for
evaporation.
> >
> >One thought that had occurred was to modify the drain tube to include an
> >aspirin sensor between two spring loaded contacts at the top portion of
the
> >tube.  But I am not sure that would work because the typical mode of
failure
> >is an ice dam that occurs at the drain hole.  I would suspect that before
> >that happens, the entire drain tube in filled with water, but I can't say
> >for sure.
> >
> >Probably the most convenient solution would be to stick a $12 CMOS board
cam
> >in a baggie with some white LEDs to use as a video inspection port for
the
> >evaporator tray on the floor.  I'd only have to power it up during
> >inspection times, so it wouldn't require a device that consumed power
24/7.
> >Any monitoring of the temperature changes will likely involve a lot more
> >power consumption than a video inspection "port" would.  On the other
hand,
> >the video evaporator inspection method probably won't tell me the coils
have
> >frozen until it's too late.
> >
> >What I really want to know is when 5 on the freezer control no longer
means
> >an average of X temperature in the freezer compartment.
>
> FWIW, a major effort nationwide by utilities is to encourage folks to
> properly dispose of inefficient 'garage' refrigerators.

Oh yeah, my utility is my friend and *always* wants the best for me.  Arf
arf!  You and Bill are always so quick to spend MY money.  (-:

> The power consumption for a modern monitoring system could trivially made
to
> be 1/1000 or less of what the fridge consumes so your concern for the
power
> consumption of the monitoring system is puzzling (to me).

Monitoring system?  You've completely missed the main points of the original
specs.  I want to know when the starts freezing up.  112W divided by 1000 is
about 1/10 of a watt?  What kind of temperature montoring AND evaluation
system are you proposing that consumes less that 1/10 of a watt? Or are you
leaving out what I need to do with the temperature information to make it
useful in the context of the original question?

Further, assuming the box only draws 112 watts when the compressor is
running, you probably need to divide your 1/1000 number by another 10 to
give 1/100 of a watt.  How does this 1/100 watt device render useful
information about impending coil-freezeups?  Keep in mind, Marc, my original
spec called for something *simple*!

I want to know when the coil has frozen and/or water is not draining into
the pan beneath the unit.

> <ROBERT_GREEN1963@xxxxxxxxx> BobbyG summarized/ended his post/question
with:
>
> "What I really want to know is when 5 on the freezer control
> no longer means an average of X temperature in the freezer
>   compartment.
>
> Why don't you do just that by leaving the dial set at 5 and measuring the
> temperature with any of a bajillion different ways discussed in this
> newsgroup?

And what mechanism correlates temperature to dial to tell me if my wife has
turned up the control as opposed to the temperature rising as a result of a
coil freezeup? Simple temperature readout can be achieved with a $10 LCD
thermometer pasted to the side but it requires continuous inspection AND
interpretation.  Those sytems probably consume 1/100 of watt but they don't
tell me if my wife's changed the dial, if the dog's opened the door, if the
machine's in a defrost cycle, the compressor has failed or if the coil froze
up.  Remember my original specs.  I want to know when the automatic defrost
system has been compromised by dirt in the drain tube.  I'd like to do it
automatically, but it seems that the best I can do it to inspect and clean
it with a greater frequency.  That consumes zero watts.  Well, maybe just a
few if I keep the door open too long when I disconnect the tube.

> A Dallas/maxim 1-wire-wire interface from www.PHAnderson.com might be a
good
> place to start. Maxim will send you samples free.
>
> http://para.maxim-ic.com/index.mvp?tree=1wire

I'm well aware of that option.  What that *doesn't* provide is anything more
than a $10 LCD thermometer with a remote probe would.  It's not necessarily
useful to solving the problem I outlined in my OP.

> Or add a LM34/35/other analog sensor to your Ocelot (or whatever).
> http://catalog.national.com/Cat/jsp/national.jsp?i=71
>
> Or add a HA7E to your www.Homeseer.com setup and have the computer email
you
> the temperature or call you on your cell phone.

No, not what I want.  I only want to know when ice has formed on the coil.
As I mentioned in the OP, that occurs when the drain tube gets plugged.
Your suggestions would lead to an enormous number of false alarms and would
require an awful lot of intelligent processing to meaningfully indicate a
simple coil freezeup or stuck drain tube.

> There was a post a while back in comp.home.automation by an Australian
feller
> who wanted to hook up his beer still to his refrigerator and then (based
on
> the tenor of his posts) directly to a gastric feeding tube -- or something
> like that ;-)  IIRC, there was some useful discussion in those threads
about
> 1-wire temperature measurement.

Crikey, mate, you're not poking a mullock at our Bazzaland brethren, are
you?  Careful, seppo!  (-:

You've settled on temperature measurment as the cure but it's neither simple
nor cheap and it's eventually not very useful in detecting coil freezeups.
The temperature of a frost-free freezer compartment can vary wildly and yet
still be operating without any problems like coil freeze-up.  That makes
temperature monitoring not nearly as useful as some of the other options I
have.

The $20 microcam with some white LED's located so that it can reveal whether
there's water in the floor level evaporator pan or not would take an hour or
so to fabricate and install and could be plugged into a spare MUX input.
The camera solution, while cheap and easy is still not what I really want.

The best solution would be something that tells me either the drain line is
plugged or the coil is frozen by sounding a sonalert or closing alarm
contacts when either condition is detected.  There's probably a commericial
solution somewhere because we had plenum AC at work that overflowed in the
same way and ruined a ton of IT equipment.  The replacement unit had both an
extra "catch tray" and a system of failure alarms including a dialout
warning system.

--
Bobby G.