> Any input would be appreciated, both about the actual current
> sensing as well as the delivery mechanism from the garage (no
> cat5 drop; wireless?) to my room (my computer or otherwise)
I don't think you'd need to isolate any particular wire to use a Hall
Effect clamp. A washer/drier would be drawing a fairly high current
that could be picked up pretty easily. You can get some pretty cheap
(NZ$10) 433MHz transceivers for comms, or make an FM "bug"
to transmit a beep in the broadcast band that you could pick up
on an ordinary radio
Jinx wrote:
>>Any input would be appreciated, both about the actual current
>>sensing as well as the delivery mechanism from the garage (no
>>cat5 drop; wireless?) to my room (my computer or otherwise)
>
>
> I don't think you'd need to isolate any particular wire to use a Hall
> Effect clamp. A washer/drier would be drawing a fairly high current
> that could be picked up pretty easily.
Hmm, as I understood you could only apply a current sensor on one
"strand" (either ingoing or outgoing) so that the fields don't cancel
each other out?
You can get some pretty cheap
> (NZ$10) 433MHz transceivers for comms, or make an FM "bug"
> to transmit a beep in the broadcast band that you could pick up
> on an ordinary radio
>
> Jinx wrote:
>
>>> Any input would be appreciated, both about the actual current
>>> sensing as well as the delivery mechanism from the garage (no
>>> cat5 drop; wireless?) to my room (my computer or otherwise)
>>
>>
>>
>> I don't think you'd need to isolate any particular wire to use a Hall
>> Effect clamp. A washer/drier would be drawing a fairly high current
>> that could be picked up pretty easily.
>
>
> Hmm, as I understood you could only apply a current sensor on one
> "strand" (either ingoing or outgoing) so that the fields don't cancel
> each other out?
Have to measure one just conductor- if going across both, they cancel
each other out.
I might add that this is how those bathroom fault detectors work, i.e.
if one leg doesn't match the other PRECISELY, there is a fault and the
current on both legs is immediately broken.
>
> You can get some pretty cheap
>
>> (NZ$10) 433MHz transceivers for comms, or make an FM "bug"
>> to transmit a beep in the broadcast band that you could pick up
>> on an ordinary radio
>>
>
> Hmm, interesting idea!
>
> - Marcel
> Hmm, as I understood you could only apply a current sensor on
> one "strand" (either ingoing or outgoing) so that the fields don't
> cancel each other out ?
The sensor is going to be closer to one wire than the other. You'd
be pretty unlucky to get exact cancellation
At 01:28 PM 8/21/2005 +1200, you wrote:
> > Hmm, as I understood you could only apply a current sensor on
> > one "strand" (either ingoing or outgoing) so that the fields don't
> > cancel each other out ?
>
>The sensor is going to be closer to one wire than the other. You'd
>be pretty unlucky to get exact cancellation
If a high-permeability core goes around both conductors, then the fields
cancel out pretty well.
> If a high-permeability core goes around both conductors, then
> the fields cancel out pretty well
Is that likely with an appliance lead ?
I just tried a UGN3503 on our washing machine lead and had
no problem picking up a distinctive 30mV 50Hz signal when the
sensor is over the active and the machine is turned on. There is
negligible signal when the machine is turned off. Enough of a
difference to detect, even unamplified, with a PIC comparator
or ADC
At 02:25 PM 8/21/2005 +1200, you wrote:
> > If a high-permeability core goes around both conductors, then
> > the fields cancel out pretty well
>
>Is that likely with an appliance lead ?
>
>I just tried a UGN3503 on our washing machine lead and had
>no problem picking up a distinctive 30mV 50Hz signal when the
>sensor is over the active and the machine is turned on. There is
>negligible signal when the machine is turned off. Enough of a
>difference to detect, even unamplified, with a PIC comparator
>or ADC
Sure, without something like a toroidal core around the wire(s).
With a clamp-on ammeter I test 0.0A with 15A AC 2-conductors flowing,
regardless of position of the wire inside the core, so the
cancellation seems to be >99.6%.
Marcel, our washing machine has a series of LEDs, one for
each part of the program. When the cycle is started, all the
LEDs are on. The last to go off is the red Final Spin. Looking
at this LED might be a better way, on this machine anyway,
than detecting power in the cord, because the amperage is up
and down with each part of the cycle. At times it's very low,
eg when just a valve is open to re-fill, and that may not be
enough to give a big enough comparison to actually finished
and off. A drier won't have those lower power states - it'll
be on (motor at least, with or without heat) or off
Jinx wrote:
> Marcel, our washing machine has a series of LEDs, one for
> each part of the program. When the cycle is started, all the
> LEDs are on. The last to go off is the red Final Spin. Looking
> at this LED might be a better way, on this machine anyway,
> than detecting power in the cord, because the amperage is up
> and down with each part of the cycle. At times it's very low,
> eg when just a valve is open to re-fill, and that may not be
> enough to give a big enough comparison to actually finished
> and off. A drier won't have those lower power states - it'll
> be on (motor at least, with or without heat) or off
>
Ours unfortunately has no visual indicators. I can only work with the
environment around the machine. This leaves me with a few parameters to
analyze:
- sound
- current draw
- vibration
So for sound, the baby monitor solution would work. But it's not
something I consider a very elegant solution, really... I'm not just
trying to monitor laundry, I'm trying to learn things about engineering
and, if applicable, electronics as well. I guess a small microphone
connected to an ADC could wait until the machine is entirely quiet and
then signal. That would be subject to environmental noise though. If
both machines are running, it won't work.
For current sensing: the actual current doesn't have to be constant. I
can always slow down the sample rate to the point where any gaps are
skipped or averaged out. The main concern I have here is that I'd have
to isolate one of strands of the wire, and I'm not sure that's such a
good idea.
Vibration testing hadn't actually occured to me until now, but I'm not
sure I can use that. I guess a very precise accelerometer would be able
to detect the shaking from the machine, so maybe that would be another
approach.
Of course, I could just train a monkey to wait for the buzz and then
race to my room to notify me.
Marcel Birthelmer wrote:
> This means that I can never hear when the washer or dryer are
> done, and so my laundry scheduling is always suboptimal.
>
> ...
>
> Now, I was thinking that it should be
> possible to pick up the current draw of those appliances and then, via a
> small embedded system or maybe an old computer, forward this information
> to the local network or something.
Picking off the current shouldn't be that hard. My first knee jerk reaction
is to make a small adaptor thingy that has a plug on one end and a socket on
the other. In other words it can be plugged in line with a standard
appliance. Now take a ferrite rod and wind a few turns of the hot side on
one end and the return wire on the other. Make sure the turn directions are
such that the magnetism in the rod adds when current flows thru the
appliance normally. Now wind maybe 100 turns of magnet wire around the
center of the rod. Leave 5mm of space between the center windings and
either end winding, and of course keep the insulation on the end windings.
This will give you a nicely isolated 60Hz voltage signal proportional to the
appliance current draw. For simplicity power the circuit from a small wall
wart which can be part of the overall adapter.
Hopefully you can figure out the rest. You won't need much analog
electronics to take the current signal and have it drive a PIC A/D input
(just three resistors and a cap if the signal is strong enough). The PIC
samples the current signal often and the rest is software to decide how low
for how long is considered off.
*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com
On 8/21/05, Marcel Birthelmer <.....marcelKILLspam.....carrietech.com> wrote:
> Ours unfortunately has no visual indicators. I can only work with the
> environment around the machine. This leaves me with a few parameters to
> analyze:
> - sound
> - current draw
> - vibration
>
There's one you left out -- time.
You could calibrate one of these:
www.acehardware.com/sm-sunbeam-60-minute-timer--pi-1277749.html
to the cycle time. It's also a portable, wireless solution.
Regards,
Mark
markrages@gmail
--
You think that it is a secret, but it never has been one.
- fortune cookie
Marcel Birthelmer wrote:
>
> I guess a small microphone
> connected to an ADC could wait until the machine is entirely quiet and
> then signal. That would be subject to environmental noise though. If
> both machines are running, it won't work.
> Of course, I could just train a monkey to wait for the buzz and then
> race to my room to notify me.
Well, if there's a buzz, what about recognizing that?
This is the most complicated yet plausible method I can come up with
using pieces I'm personally conversant with:
Small condenser mic feeds into a PIC's ADC. Find out the fundamental
frequency of the buzzer. Sample at 1-4x Nyquist frequency. Do an FFT
(may require a dsPIC) and look for energy in that spectral band, OR
count zero crossings and match a window of samples around the zero
crossings to find the fundamental (less computationally intensive and
may work well for a single target sound). When you've detected that
fundamental for 90-110% of the normal duration of the buzz (and it has
then stopped), send out a packet over a ZigBee wireless connection
(there are several products that will convert TTL serial I/O to ZigBee).
The other end of that is connected to one of the PICs with USB
support, set up as a HID for simplicity. The USB then plugs into your
PC and talks to an GUI app there.
Now, ideally, your PC is either set up as a web server (if you have a
broadband Internet connection) so you can poll the state of your laundry
from anywhere in the world (and the rest of us could see it too). You
could also use a script to FTP it to an external web server though, even
over dialup.
Great! This will probably one very interesting project for
hobbyist. :)
Who knows what will happen 20 years later? Embedded TCP/IP
seems quite hot some time ago. Everything will has its
own IP address. A person will have lots of IP address
associated with him/her since he/she will carry lots
of gadgets. :)
Seriously the only barrier is now the power source. Wireless
power transfer is not so easy now (only for very low
power).
I use an Allegro Hall-effect device to monitor current. Since
it uses magnetic coupling, there is no problem with 110/220VAC.
If you use a 12F675 you can either use the on-board comparator or
an A/D channel to determine whether the appliance is drawing current or
not. You can convert that information to RS-232 by bit-banging a serial
port.
Finally, a Lantronix XPort does a great job of converting serial to
tcp/ip. You can telnet to it, or it can send an email message or serve
up a web page telling you the state of the appliance.
> Great! This will probably one very interesting project for
> hobbyist. :)
>
> Who knows what will happen 20 years later? Embedded TCP/IP
> seems quite hot some time ago. Everything will has its
> own IP address. A person will have lots of IP address
> associated with him/her since he/she will carry lots
> of gadgets. :)
>
> Seriously the only barrier is now the power source. Wireless
> power transfer is not so easy now (only for very low
> power).
>
> Regards,
> Xiaofan
>
On Sun, 21 Aug 2005 17:08:42 +1200, Jinx wrote:
>...<
> A drier won't have those lower power states - it'll
> be on (motor at least, with or without heat) or off
Not my dryer - it stops and reverses to stop tangling.
The off period is a few seconds each time. (And it's
gas heated, so during those stops there may be just the
control circuits drawing current, or those plus a
solenoid).
My understanding of the project is to make two devices. The two devices
being a remote device and a local device. The remote device will sense if
each of two appliances are either on or off, and send the information to the
local device for display. If there is a failure in the system, some one will
get yelled at for leaving their cloths in the washer.
Would "not being yelled at" be considered "Mission Critical"?
I think that the sensors for the remote device should be cheep, simple to
interface and fun. With "fun" being important. Different uses for common
objects. The computer could integrate or differentiate the input signal
using a simple algorithm.
////////
Is input one on?
No
Has input one been on in last twenty seconds?
No
Then input one is probably not on.
/////////
It is only necessary for the remote unit to transmit data, and the local
unit to receive this data. As only four states are possible, the data
encoding can be simple. A simple way to do this is to use a form of Morse
code or CW transmition. Yes, dits and dahs.
////////////
The following relationships exist between the elements of the code (dits and
dahs), the characters (letters) and the words:
The DIT is the Basic UNIT of Length.
The DAH is equal in length to three DITS.
The space between the DITS and DAHS within a character (letter) is equal to
one DIT.
The space between characters (letters) in a word is equal to three DITS.
The space between words is equal to seven DITS.
(Source: U.S. Army Technical Manual TM-11459/TO 31-3-16 - Sept. 1957)
///////////
So if a dot is a binary zero and a dash a binary one, the start bit could be
a dash, or three time slots. This time could be divided by two on the
receive side and used to determine weather the next two bits are dashes or
dots, 3 or 1. If less than half the start bit, it is a dot. Else it is a
dash. And no signal for a dash time (3) shows end of data.
With the sensors the effect of ambient noise and drift due to age and
temperature can be somewhat mulled out by taking a reading with no signal
present (appliance off) and using this value as a "floor" for the input.
This calibration can be done upon power up of the remote and/or with a push
button on the remote unit.
As a "fail Safe" feature, the local unit could have a preset timer function
that is set via a push button.
engineer.cotse.net .
