Skip navigation
TechXchange
Digi-Key TechXchange Communities > Component Discussion and Reviews > Discussions
36935 Views 30 Replies Latest reply: May 20, 2014 7:21 PM by JerryS RSS
Currently Being Moderated

Aug 5, 2013 10:36 PM

Split core current transformers

Hi, all.  I'm looking for split core current transformers at a reasonable price.  This will be for

monitoring circuits in my home.

 

The Magnelab SCT-400 and SCT-750 series look good because they come in multiple

ratings , and all have the same output (.333V) at max rated amperage.  This would greatly

simplify circuit design - changes for different current ratings could be done in software.

 

But at $45-55 ea., these get pretty expensive pretty quickly.  I'd have well over $1K just in the

transformers to monitor each circuit in the house.

 

I've seen cheaper ones, but their outputs vary significantly.  Each circuit would have to be

different; some could get by with just changing the ballast resistor, but others would

require additional changes in the amplifier gain (or loss, as the case may be).

 

So, does anyone know of a less expensive source of current transformers like Magnelabs?

I'm especially interested in standard home circuit breaker ratings, i.e. 15A, 20A, 30A, 40A,

100A, 150A and 200A.

 

Or would you just design a circuit for each one and use current transformers with ballast

resisters?

 

TIA.

  • Currently Being Moderated
    Aug 8, 2013 10:01 AM (in response to JerryS)
    Split core current transformers

    Seems like the easiest thing to do is go with the cheaper ones and use pots (variable potentiometers) to adjust the output to your requirements.

  • JimHarrison Apprentice 83 posts since
    Sep 9, 2011
    Currently Being Moderated
    Aug 12, 2013 8:46 AM (in response to JerryS)
    Split core current transformers
  • Freemarket Novice 4 posts since
    Aug 28, 2013
    Currently Being Moderated
    Aug 28, 2013 10:57 AM (in response to JerryS)
    Split core current transformers
  • Freemarket Novice 4 posts since
    Aug 28, 2013
    Currently Being Moderated
    Aug 28, 2013 10:59 AM (in response to JerryS)
    Split core current transformers

    How about Wifi networkable Current Sensing?

     

    http://www.functionaldevices.com/pdf/datasheets/RIBTWXV2401B-WI.pdf

      • Freemarket Novice 4 posts since
        Aug 28, 2013
        Currently Being Moderated
        Aug 29, 2013 7:37 AM (in response to JerryS)
        Split core current transformers

        Hi Jerry,

         

        No burden resistor required, all that is taken care of and the device is prepackaged.  I havent looked at the Magnelabs yet.

         

        Freemarket

          • Freemarket Novice 4 posts since
            Aug 28, 2013
            Currently Being Moderated
            Aug 29, 2013 4:22 PM (in response to JerryS)
            Split core current transformers

            Jerrys

             

            That statement is just to protect from touching the wire the current sensor is connected around.  The current sensor is completely safe to touch, the output terminal or wires, depending on the model you get, are low voltage and safe.  Of course, you still have to be careful with the Line Voltage wire these are wrapped around (your house wire).

             

            Also, if you use the RIBXGTV5-20, 50, or 100, no power supply is required.  The power supply is only required for the 4-20mA output devices (RIBXGT420 for example).

             

            Freemarket

  • Bill_Thomson Novice 3 posts since
    Dec 29, 2013
    Currently Being Moderated
    Dec 29, 2013 2:17 AM (in response to JerryS)
    Re: Split core current transformers

    Jerry,

     

    This looks like what you're looking for. At 18 bucks each, the total damage to your wallet would be a lot less too!

    http://www.byramlabs.com/index.php?cPath=78&filter_id=71

     

    Depending on what you're going to do the monitoring with (Arduino, PIC, etc) I'd say you'll want the model that

    puts out 1 volt. If you try to use CTs with a 333mV output, you're severly limiting the range and resolution without

    resorting to some sort of amplifier to increase the CT output. (I speak from experience here, I tried using the 333mV Magnelab

    units, and though they DID work, accuracy and range suffered.)

     

    Check out the OpenEnergyMonitor project. There's a LOT of good info regarding CTs and energy measurement on the site.

    I've built their wireless transmitter module (emontx) and graphic LC display module (emonGLCD). I use a Raspberry Pi to collect the data.

     

    When used with the right CT, and calibrated with reasonable care, accuracy is within 2-3%.

     

    http://openenergymonitor.org

     

    HTH

     

    regards,

     

    Bill

  • Jason058 Novice 1 posts since
    Mar 30, 2014
    Currently Being Moderated
    Mar 31, 2014 4:01 AM (in response to JerryS)
    Re: Split core current transformers

    Hi Jerry, 

    How about this split core current transformer. CT303, Max current 120A, 33.3mA, 1.0 Accurately class, 16mm ID. 4.90USD/pcs.

    http://www.zntar.com/split-core-current-transformer/39-ct303-split-core-current-transformer-with-35mm13mm-dc-plug.html

     

    Best Regard,

    Jason

    • davecuthbert Apprentice 87 posts since
      Feb 18, 2014
      Currently Being Moderated
      Mar 31, 2014 10:28 AM (in response to Jason058)
      Re: Split core current transformers

      Jerry, you can can take your measurements with one current transformer, no hardware changes and no firmware changes. For example, configuring your circuit to use the full range of a 16 bit A/D at 100 amps provides a resolution of 3 mA. Even a 12 bit A/D provides resolution of 24 mA.

       

      The LEM USA model TT 100-SD split core transformer is quite linear from 0 to 100 amps and is available at Digikey for $25.20 each in quantities of 10. The Digikey part number is 398-1079-ND.

       

      http://www.lem.com/docs/products/tt100sd_e.pdf

       

       

      And the CR Magnetics CR3110-3000, 75 amp split core transformer is only $11.64 in quantities of 10. The Digikey part number is 582-1004-ND.

       

      http://www.crmagnetics.com/Products/Assets/ProductPDFs/CR3100.pdf

  • Bill_Thomson Novice 3 posts since
    Dec 29, 2013
    Currently Being Moderated
    May 6, 2014 4:09 PM (in response to JerryS)
    Re: Split core current transformers

    Hi Jerry,

     

    That same model CT (Wattcore WC1) is also available with a 2 volt RMS output. Perhaps a better fit for your application?

    http://www.byramlabs.com/product_info.php/products_id/20486/product/WattcoreWC1%20SERIES%20WC1-025-RV002%20Split-Core%20Current%20Transformer%20%2825:2000mV%29

     

    Regards,

     

    Bill

    • davecuthbert Apprentice 87 posts since
      Feb 18, 2014
      Currently Being Moderated
      May 6, 2014 5:05 PM (in response to Bill_Thomson)
      Re: Split core current transformers

      Diode linearity should not matter with an active precision rectifier. In fact the ADC can be driven with a DC level shifted AC waveform and with some math it can calculate the RMS voltage with no rectification needed.

       

      Amplifying the signal a potential 0.2% or 2% gain error is introduced by the 0.1% or 1% Op Amp resistors. And the Op Amp introduces an offset. These issues can can be compensated for in firmware if each unit is calibrated.

      • Bill_Thomson Novice 3 posts since
        Dec 29, 2013
        Currently Being Moderated
        May 18, 2014 10:38 PM (in response to JerryS)
        Re: Split core current transformers

        Jerry,

         

        That's 2V RMS, ie. 5.65Vp-p!

         

        Just thought if you had a CT that could give you the 3Vp-p you're looking for it might simplify things a bit...

          • davecuthbert Apprentice 87 posts since
            Feb 18, 2014
            Currently Being Moderated
            May 19, 2014 8:48 AM (in response to JerryS)
            Re: Split core current transformers

            The accuracy of a precision half wave rectifier does not depend on the diodes. The two dominant error sources are Op Amp input offset voltage and the resistor tolerance. The circuit can be configured for a single power supply. A dual power supply precision rectifier is shown below.

             

            If the A/D is followed by a microprocessor capable of doing some math, no rectification is needed. The microprocessor can simply measure the waveform P-P value and calculate the RMS value. If significant waveform distortion is anticipated the microprocessor can perform a true RMS measurement rather than multiplying the P-P value by 1/[2(2)^0.5].

             

            When using A/D converters, as with any instrumentation, an error analysis of each stage should be performed. I've seen cases where an analog front end increased A/D measurement error rather than helping it. Running an A/D at a fraction of its range can sometimes be better than adding an amplifier before it to use the entire range. One thing to watch out for that the circuit should be able to source sufficient charge when the A/D sample-and-hold samples. This can be done by placing a resistor between the signal source and the A/D with a capacitor to ground at the A/D input.

            Attachments:
              • davecuthbert Apprentice 87 posts since
                Feb 18, 2014
                Currently Being Moderated
                May 19, 2014 1:56 PM (in response to JerryS)
                Re: Split core current transformers

                Given an ideal Op Amp (infinite GBP) in the circuit shown, the diode non-linear forward voltage drop is completely compensated for. The voltage at the summing junction is zero volts and that is set by the voltage at the diode cathode. The voltage at the diode anode does not enter into the error calculation, given the ideal Op Amp. However, with real world Op Amps a small error voltage needs to exist and be amplified to overcome the diode forward voltage drop. In a practical implementation, for example using the LT1012, the finite gain of the Op Amp creates the need for this error voltage. The LT1012 gain is  about 10,000 at 60 Hz. The diode forward voltage of 500 mV then leads to an error of 500mV/10,000 = 50 uV at the peak of the AC waveform. LTSpice shows an error of 30 uV. The Op Amp/diode gain error at the zero crossing - where the diode forward voltage drop is 0 volts - is zero. The Op Amp/diode FS gain error for a 14 bit A/D having a FS of 1 volt, is about 1 LSB. This error - if it's an issue - can be reduced by using an Op Amp having a higher gain at the frequency of interest.

                 

                The Op Amp input offset voltage is another error source and it can be compensated for in firmware or it can be minimized using an Op Amp having a suitably low Vos. No calibration is needed as the information needed to perform a firmware input offset compensation is present in the Op Amp waveform.

                • davecuthbert Apprentice 87 posts since
                  Feb 18, 2014
                  Currently Being Moderated
                  May 20, 2014 12:51 PM (in response to davecuthbert)
                  Re: Split core current transformers

                  The ADC needs to have considerable headroom to measure non-linear loads having a high crest factor. I see some instruments designed for a 5:1 crest factor. To measure a 20 A RMS, and be sufficient for a 5:1 headroom, the ADC needs to be able to measure 100 A without clipping. The CT can still be rated for 20 A RMS.

                   

                  A simple topology is to drive the CT into a true RMS converter and drive that into the ADC.

                   

                  A more accurate way can be to drive the CT into the ADC then drive that into a uP. The uP then does the RMS computation. Note that in neither topology is rectification needed. To drive the AC waveform into the ADC the CT drives through a capacitor to a DC bias voltage at the ADC input. The ADC measures the waveform periodically (perhaps at a 1200 Hz rate - it depends on the accuracy and resolution needed) and performs an RMS computation.

                   

                  One can get fancy and derive the circuit power from the CT transformer itself. To do this the circuit periodically disconnects the burden resistor and substitutes a shunt regulator which charges a cap. An easier (but more expensive) method is to use a second CT to power the circuit. To go along with this a WiFi link would be great. Clamp the device onto a wire and it can be polled for the RMS current by a PC.

                   

                  In any event the design should conform to the appropriate safety standard. I don't know which one offhand but for my own use I would default to UL60950. If I was designing it as part of my job and having it certified I would locate and purchase the correct safety standard.

                • MikeWeed Novice 62 posts since
                  Oct 21, 2011

                  OPAMP.bmpIf I could add a little light (and not too much heat) to this discussion, a precision ground-referenced rectifier can be made with 2 GOOD rail-to-rail op amps and no diodes, per the attached diagram. When Vin>0, all op amp inputs and outputs go to Vin, so Vout=Vin. When Vin<0, op amp output A goes to its negative rail, which is 0 volts, and this is applied to the + input of op amp B. The output of op amp B does what it has to do to bring its - input to 0 volts; it does this by taking its output to +|Vin|.  It is possible to also get some gain out of this rectifier by using additional resistors.

More Like This

  • Retrieving data ...

Bookmarked By (0)

Legend

  • Correct Answers - 4 points
  • Helpful Answers - 3 points