Re: Current transformer compensation idea

From: John Larkin (jjlarkin_at_highSNIPlandTHIStechPLEASEnology.com)
Date: 07/27/04 

Date: Tue, 27 Jul 2004 14:54:18 -0700

On Tue, 27 Jul 2004 18:25:49 GMT, analog <analog@ieee.org> wrote:

>A typical current transformer may have a one turn primary and
>a hundred or more secondary turns. The secondary is normally
>terminated into a small resistor (possibly through diodes)
>such that the core must support a small ac flux excursion.
>With secondary signals in the volt range, the primary voltage
>burden is minimal, usually a few millivolts.
>
>Although dc drift may be a problem for some configurations,
>a typical current transformer rarely comes anywhere close to
>saturation during normal operation. In spite of this,
>inductive signal droop may be a problem in high fidelity
>applications (magnetizing current is typically very non
>linear).
>
>I have been toying with the idea of using active circuitry to
>minimize magnetizing current. My first idea was to arrange
>the current transformer to drive the summing junction of an
>opamp rather than terminating it into a small resistor. This
>would tend to keep the voltage across the CT's secondary at
>zero, which would be a noticeable improvement over the
>standard arrangement.
>
>However, this would still leave the voltage burden from the
>sense current flowing through the CT's winding resistance.
>Even this could be largely nulled out by actively driving the
>"grounded" end of the current transformer with a feed forward
>signal proportional to current appropriately scaled just to
>equal the drop developed on the internal winding resistance.

That's equivalent to terminating the winding with a negative
resistance. It's tricky in real life, as the winding's resistance will
change with temperature and the thing could get unstable.

>
>Okay, I have never built this circuit and don't have a real
>application for it, but the simulator says all works great.
>What I am wondering is whether anyone has used or seen such
>a technique before or could imagine a situation where such a
>circuit might prove useful. Note that this technique does not
>eliminate the dc saturation problem (although it does make the
>CT's core "look" much bigger). Comments or further ideas?
>
>analog

Active feedback is commonly used with CTs. The usual ways are...

For AC only, there are two "secondary" windings, one sense and one
feedback. The goal is to keep the core at zero flux, namely to keep
zero voltage at the sense winding output. A opamp (or power opamp)
closes the loop, and the burden resistor is in series with the
feedback winding. Very accurate.

For DC, there are two techniques:

1. Use a hall sensor in a tiny gap in the torroid, instead of the
sense winding, with the opamp+feedback winding fighting to keep the
core flux at zero. This works pretty well, but the gap and the hall
drift are issues.

2. Build a real (Danfysik or Bergoz) type DCCT. This is complex: two
saturable torroids (permalloy or metglas), lots of windings. At high
frequencies, it works like the regular AC feedback scheme, but there's
an additional DC servo loop driven by a low-frequency oscillator that
observes DC imbalances as 2nd harmonic signal in the sense coil. I did
this once, didn't sell many, and it's fairly messy. When done right,
you get a DC-wideband CT that's parts-per-million accurate.

John

Relevant Pages

  • Re: 86 560SL no ignition spark???
    ... I will take the coil into work to check the resistance from pos to neg. ... The primary winding has a low ... > When the distributor rotor hits one of the spark plug terminals inside ...
    (alt.auto.mercedes)
  • =?ISO-8859-1?Q?Re=3A_DIY_electromagnet_under_=A3100=3A_Disappointing_res?= =?ISO-8859
    ... it'll have a resistance of 39 milliohms. ... So size the wire to match whatever power supply you have, ... Each winding doesn't know ... I used an old Varian electromagnet with water cooling. ...
    (sci.electronics.design)
  • Re: Coupling coefficient of industrial transformers
    ... the RF transformers are wound specifically for tight ... enough that there is insignificant drop in that resistance due to the ... AC resistance of the windings at the frequency of measurement, ... wits about you and account for parasitic effects like winding ...
    (sci.electronics.design)
  • Re: Transformer color code quandry
    ... color and resistance are consistent with the ... filament winding, another as the 6.3v winding, leaving what I believe is ... another dark color wire. ... the unloaded voltage will be higher than the loaded voltage. ...
    (rec.radio.amateur.boatanchors)
  • Re: Question about magnetic forces beyond saturation
    ... The core material is ideal ... so saturation is not a factor. ... One is to measure resistance from adjacent ... connections are made using the bridge principle so as not to excite some ...
    (sci.physics.electromag)