Shop OBEX P1 Docs P2 Docs Learn Events
Household KWH usage monitor — Parallax Forums

Household KWH usage monitor

Erik FriesenErik Friesen Posts: 1,071
edited 2010-10-20 21:13 in General Discussion
I have been thinking about getting one of these, or making my own.

http://www.theenergydetective.com/

http://www.envirogadget.com/energy-saving/emonitor-total-home-energy-management-system/

One of the problems with making my own is, I don't know where to find inexpensive clamp on current transucers(transformers?)

I have considered making an open source project using the prop, however lack of usb and adc could be a drawback.
«1

Comments

  • FranklinFranklin Posts: 4,747
    edited 2010-10-02 18:37
    You can buy a pair from TED for $70 (I think that's the price)
    http://www.theenergydetective.com/additonal-mtu-cts
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2010-10-02 18:54
    Yes, but obviously if the other monitor comes with 20 or so rings, they are not paying $35 a piece for them as the complete unit sells for around $650.
  • FranklinFranklin Posts: 4,747
    edited 2010-10-02 20:21
    So buy theirs instead.
  • ercoerco Posts: 20,260
    edited 2010-10-02 21:51
    You may have missed my post a few days ago, when these were just $20: http://forums.parallax.com/showthread.php?t=125996
  • kwinnkwinn Posts: 8,697
    edited 2010-10-03 09:30
    It is possible to make and calibrate your own current transformers for this purpose.

    Use a toroidal ferrite core large enough for AWG 12 wire to pass through the center after the small gauge wire has been wound on the core.
    Wind several turns (one layer) of small gauge wire around the core.
    Use an op amp for each transformer to amplify the signal to a level suitable for input to an ADC.
    Use a calibrated clamp on ammeter to calibrate the transformer/op amp/ADC over the desired operating range.

    The scale is reasonably linear so a 3 point calibration is fairly accurate.
  • kwinnkwinn Posts: 8,697
    edited 2010-10-03 09:36
    PS - For north american 110V circuits and european 220V circuits the transformer can be placed on the neutral line. For 220V north american circuits it needs to be on one line, or both lines if both 220V and 110V are available (utility outlet on stoves).
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2010-10-03 15:23
    I don't suppose using this style of transformer should be that hard, I just haven't done it before.

    http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=582-1004-ND

    http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=811-1109-ND

    What is the correct way to rectify this? Post burden resistor with bridge rectifier and cap for smoothing? Whats the proper way to choose cap size for this?

    Did some checking on the e-monitor, they do pretty much everything online. I do not care for that type of equipment, as you are left in the lurch if they go out of business, or you go offline.
  • FranklinFranklin Posts: 4,747
    edited 2010-10-03 15:26
    http://media.digikey.com/pdf/Data%20Sheets/CR%20Magnetics%20Inc%20PDFs/CR3100%20Series.pdf
  • Brian_BBrian_B Posts: 842
    edited 2010-10-06 06:45
    Cool topic,
    I bought a AmWatt meter ($24) at the local home improvement store .I took the current transformer out ,installed a shunt resistor across the 2 wires. I'm having a hard time getting any kind of voltage reading off of it to be read by a ADC .The voltage seems to go up and down even with a steady load on it. any suggestions?

    Brian

    Ps I was using Some florescent lights for the load if that matters

    WARNING: Do not mess with live AC power if you don't what your doing ,I left the transformer in the case and sealed up all bare wires.
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2010-10-06 07:11
    Its ac coming off those wires.
  • Brian_BBrian_B Posts: 842
    edited 2010-10-06 08:02
    Eric,
    Yes it's very low voltage ac, I tried a small diode bridge to rectify it and still jumps up and down.

    Brian
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2010-10-06 08:43
    You'll need to filter it. I was hoping someone would hop on board here and help out on that, as I have never done it.
  • kwinnkwinn Posts: 8,697
    edited 2010-10-06 10:58
    The voltage will depend on the value of the shunt resistor and the current. Normally the voltage would be amplified by an op amp, rectified, and then filtered before going to an ADC.
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2010-10-06 11:01
    What is the reason for the op amp? Does a low value shunt work better?
  • kwinnkwinn Posts: 8,697
    edited 2010-10-06 12:01
    Usually the shunt resistor is a very low value (0.1 - 0.01 ohms) so the AC voltage is low. If you were trying to measure 0-15A with a 0.1 ohm resistor (0-1.5V output) it might be ok, but if the shunt resistor was 0.01 ohms the output (0-0.15V output) would be near the bottom of the ADC range (typically 0-2.56 or 0-5V).
    Using a diode to rectify the voltage makes the problem worse due to the voltage drop across the diode. An op amp can compensate for these problems by amplifying the signal.
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2010-10-06 12:05
    Ok, that's making sense, hadn't thought of the diode voltage drop.
  • kwinnkwinn Posts: 8,697
    edited 2010-10-06 12:29
    Erik, these days the simplest way to do this is to have an op amp that runs from a single supply (3.3 or 5V) biased to the mid point with the ac signal going above and below that point. Using a zero crossing detector and having the micro/adc read the voltage at the positive and negative peaks (4.166mS after 0 crossing, a trivial task for the prop) to get the p-p voltage and calculate the current from that. Slightly more complicated software but much simpler hardware, particularly if you want to read several channels.
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2010-10-06 12:44
    The zero crossing thing is something I hadn't considered. Really, you could about do that with a quick adc, and ditch the op amp, unless you would need that to boost the signal source resistance off the shunt.
  • kwinnkwinn Posts: 8,697
    edited 2010-10-06 13:11
    I am assuming this resistor is on the neutral side for safety, and you use the neutral side as the reference for your ADC. That means the signal goes negative half the time, and most ADC's will not handle bipolar input signals. Other than that I agree that a fast adc alone would work if the voltage from the shunt was high enough.
  • kwinnkwinn Posts: 8,697
    edited 2010-10-06 13:19
    PS - Having a direct connection to the AC power even on the neutral side is not safe. If the neutral wire were to come loose from the shunt you could now have 120V on the circuit via the load. I much prefer some form of isolation such as the current transformer or an optoisolator.
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2010-10-06 13:30
    I still intend to use the current transformer.

    The deal is, I want to have about 30 inputs, and the simpler I can make it the better. I would use one zero crossing detector, and also measure the true voltage from each leg through a voltage divider.

    I'll have to think more about this, not sure if a voltage divider will work for bumping the tranformer output to half voltage.
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2010-10-06 13:42
    I have been looking at these -

    http://www.crmagnetics.com/products/CR8410-P93.aspx

    They have this circuit listed for the lm324 -

    http://www.crmagnetics.com/products/Assets/ProductPDFs/Precision%20Rectifier%20Circuit%20for%20CT%20Signal%20Conditioning.pdf

    Really using the circuit isn't that bad, with sm 324's at $0.57 .

    They don't say much about supply on that circuit, I sort of assume that would work with a single supply?
  • kwinnkwinn Posts: 8,697
    edited 2010-10-06 13:52
    I built a 6/8 (designed for 8 but only 6 used) channel unit in the late 70's using hand wound current transformers, 2 quad op amps, some other circuitry, and a micro (6502 I think). Wish I still had the documentation.
    If the output voltage from your transformer is that high the op amps may not be needed, and one zero crossing detector should be enough, if one is even needed. Taking multiple readings per cycle and letting the software do the calculations would also work.
    Four 8 channel adc's would work and so would 4 8 channel analog muxes and a 4 channel adc. Simplest might be multiple micros with built in adc's.
  • gunstarg10gunstarg10 Posts: 4
    edited 2010-10-10 21:11
    Hi everyone,

    I don't mean to thread jack but I am doing a similar project and want to monitor power consumption as well. I am not that knowledgeable on how to go about it but I am using the propeller and wanted to use the current transformers (http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=582-1004-ND). I just don't know how I would integrate the two. Would I need an adc like the one parallax sells (item code ADC0831) for the propeller to read the input voltage from the CT? Is this even the right way to go about it? I have read the previous post but was kinda lost still. Any help or references is much appreciated. Thanks.
  • kwinnkwinn Posts: 8,697
    edited 2010-10-11 09:34
    At minimum you would need a load resistor for the current transformer and an ADC to digitize the signal. BE CAREFUL. The output voltage from the current transformer can be very high if there is no load on the output.

    Take a look at "http://en.wikipedia.org/wiki/Current_transformer#Design" for a description of current transformers.

    The data sheet for the current transformers you looked at has all the information needed for the design.
  • kwinnkwinn Posts: 8,697
    edited 2010-10-11 10:09
    PS - You may want to take a look at these "http://www.amveco.com/pdf/AC1025.pdf". They are less expensive and simpler to use with an ADC. The only disadvantage is that they are not a split core so the wire to be monitored must be disconnected and fed through the core.
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2010-10-11 18:38
    I kind of think without either monitoring the wave or averaging it, a person could get a bit out of wack if the power factor was too low.
  • kwinnkwinn Posts: 8,697
    edited 2010-10-12 07:01
    Erik, true the power factor does matter if you want to measure true power, however that can be done by measuring the zero crossing point of the voltage relative to that of the current.
  • Erik FriesenErik Friesen Posts: 1,071
    edited 2010-10-12 16:12
    That depends, not? Look at these waveforms - http://jarv.org/pwrmon.shtml

    Judging by those, if you had a sample rate of 1khz or better you could simply add all the values together, not?
  • FranklinFranklin Posts: 4,747
    edited 2010-10-12 17:39
    The Tweet-A-Watt (http://www.ladyada.net/make/tweetawatt/) samples one cycle 17 times at 1ms intervals to get a waveform. and repeats every two seconds.
Sign In or Register to comment.