# The theory behind a Direct-Current-Comparator Resistance Bridge

#### PaulEE

Joined Dec 23, 2011
423
I will be thoroughly impressed if I get a response to this thread. This topic has been found to be rather obscure as far as google searches and the like.

Here at work, we have a measuring instrument called a current-comparator bridge. The instrument allows the comparison (down to sub- part per million levels) of resistance standards.

This is achieved by comparing the differences in magnetic flux in specially-wound, identical (nearly), high-"mu" magnetic cores. The only difference between the cores is that the turns ratio on one core is adjustable via dials on the front of the instrument. In this way, the operator can run two currents through two resistors (and two cores), and adjust the dials so as to determine the precise ratio between the currents, and thereby, the resistors.

I have a book that describes the operation of these instruments in a fair amount of detail, but I am still unclear about a few things.

"Sensitivity to direct current ampere-turns or flux is achieved by modulating the magnetic core with superimposed alternating current ampere-turns, with the result that, when a direct current magnetizing force is present, even-harmonic components of the modulation are generated."

What on earth is this book excerpt talking about? Also:

"Two identical magnetic cores are normally used so that by suitable interconnection of the windings on each core the odd harmonic components tend to cancel one another and only the even harmonics remain"

Again, big question mark.

Since this is a magnetic device/magnetic amplifier of sorts, there is a modulation source that is constantly flip-flopping the orientation of the magnetic field to prevent permanent flux offset in the cores themselves. The oscillator frequency is usually just under 1,000 Hz; 700-800 Hz, typical.

My questions are as follows:

-How does a direct current comparator utilize a modulation source, highly magnetically permeable core, and two identically wound cores (and flux sensing turns), to determine the ratio of currents?
-What is meant by "even harmonic components are generated under the influence of DC magnetizing force being present"?

The circuitry that runs this instrument is equally daunting to look at. There are various sections; the oscillator, toroid diagram, peak detection/demodulator, flux amplifier, master and slave current supplies, nanovolt amplifier, turn balance, and a few other circuit boards.

If anyone can shed some light on magnetic amplifier theory, or anything in the above few paragraphs, I'd really appreciate some guidance.

I am fluent with electromagnetics and electrical engineering, but am not so fluent in specifically magnetic devices. I do understand BH curves and I do know that some are more square, while others have rounded edges.

#### praondevou

Joined Jul 9, 2011
2,942
It sounds in a certain way similar to some devices I used.

Just to be on the right track, the DC current passed through the resistors would create an offset in the magnetic cores if the resistors were different, say current doesn't cancel?

I've seen battery cabinet leakage monitors which I believe to be working on a similar principle. There is a core with a coil which is being driven with a alternating current. + and - wires from the battery cabinet pass through the core. If there is leakage, both currents will be different creating an offset in the core, the core reaches saturation in one direction.

We use a similar principle on a torque sensor where a a fluxgate circuit is used to measure a magnetic field developed by a magnetized shaft. The measuring coils are in this case constantly driven into saturation. The current to the coils is measured and integrated, the circuit gives then an output voltage proportional to the magnetic field strength.

(In your case that would be a static magnetic field developed by the two currents that do not cancel out)

In case this has nothing to do with your device, forget about it.

#### PaulEE

Joined Dec 23, 2011
423
It sounds in a certain way similar to some devices I used.

Just to be on the right track, the DC current passed through the resistors would create an offset in the magnetic cores if the resistors were different, say current doesn't cancel?

I've seen battery cabinet leakage monitors which I believe to be working on a similar principle. There is a core with a coil which is being driven with a alternating current. + and - wires from the battery cabinet pass through the core. If there is leakage, both currents will be different creating an offset in the core, the core reaches saturation in one direction.

We use a similar principle on a torque sensor where a a fluxgate circuit is used to measure a magnetic field developed by a magnetized shaft. The measuring coils are in this case constantly driven into saturation. The current to the coils is measured and integrated, the circuit gives then an output voltage proportional to the magnetic field strength.

(In your case that would be a static magnetic field developed by the two currents that do not cancel out)

In case this has nothing to do with your device, forget about it.
Look at you, being all brave, responding to my post! First person in over 100 views. I thank you!

After no forum luck, I did some pretty deep Google searching. Here's the scheme they use:

Two identical toroidal cores with high-mu magnetic core material carry currents from each of two current loops; each loop goes through one resistor - one is the standard, and the other is the unknown. Each current loop is fed equally (at null) and oppositely through each of the two inner cores. (Think two toroidal inductors sitting on top of each other, with proper shielding between).

A high power oscillator continuously flip-flops the magnetic field within both cores (think of the old chopper op-amps, only magnetic phenomena instead of a reed or electric switch, or varactors).

Here's what I didn't know when I first posted the thread:
The sense coil (which is a very finely wound coil that wraps all around the toroids, in an attempt to capture all flux; think Ampere's law and clamp current meters) outputs into a demodulator circuit. Apparently, as a net DC magnetic flux appears to the sense coil, it causes a shift in the oscillating output signal that is proportional to the difference in current.

This device was called a "second-harmonic magnetic demodulator". After reviewing the patent for it, I've come to the conclusion that the square wave signal is "demodulated" such that only even harmonics result at the output.

THIS IS WHERE I NEED HELP:
I was under the impression that a square wave was a DC offset (could be zero) and odd harmonic components from the fundamental frequency. If this circuit that reads the sense coil is demodulating and looking at EVEN harmonics, could I rush to the conclusion that it is using the appearance of even harmonics to determine how much DC is in the core?

My reasoning behind this is fairly straightforward:
At null and with no DC flux in the core, the square wave has an equal...we'll call it...duty cycle; it is symmetrical.

However, as small DC offset is introduced, the TRANSFER CHARACTERISTIC OF THE CORE, which is a nearly vertical line at zero flux, +/-, causes an asymmetrical wave to result. In other words, think of a differential opamp transfer characteristic. Near zero difference, the gain is zero; as you move to some place VERY near zero but not zero, the characteristic is very sharply up and down. If the input signal has a DC offset (shifted one way or another), you are no longer at the center of the characteristic, and the non-linear edges cause non-linear output signals.

The beauty of the current comparator bridge for resistance measurement is that #1: you do not need ppm current sources; only nulling circuitry. #2. they are very linear and are immune to changes in temperature, humidity, and time. All it is is a huge toroid and op-amps, among other things. #3. They offer sub-ppm comparison between a known standard and an unknown. finally, #4. each resistor in each loop can run at "rated current", while still performing the test. The turns ratio is simply adjusted with dials so that the flux per loop in the cores are nearly the same.

It is an incredibly cool device as far as its inner-workings. I do believe that what I have described is the answer I was seeking, but I wanted to share with you all, in case any one of the hundred or so people stopped back to the thread.

Thank you for your response, praondevou.

To everyone, please critique my logic regarding the even/second harmonic magnetic demodulator if I am mistaken. I think that's how it is functioning, though. Magnetic devices have always had a certain mystery about them, to me.

Thanks!