Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Fluxgate or zeroflux tecnology


Hi everyone . I've been wanting to understand how the fluxgate or zeroflux technology used in magnetometers or current sensors works for a few weeks. it seems that there is no single method for their realization but there is more than one. This is what leaves me disconcerted. For example, there are those who use a square wave and others instead use an amplifier that reproduces a signal equal to that of the detection coil and then sends it to the compensation coil to reset the flux. there must necessarily be a minimum of flow otherwise how can there be induction ?.
Active current sensors are simple; the way they work is purely magnetic field strength, not "induction".

Typically they have a magnetic core, that has a gap with a hall-effect sensor in it.

The output from the sensor is amplified and used to drive the feedback coil, which has eg. 1000 turns (vs. usually a single "turn" for the cable being sensed).

The current required to balance the net field back to zero at the sensor will in that case, be 1/1000th the input cable current.

By using a load resistor in series with the feedback coil, appropriate for the current range being measured, you can read a voltage across that proportional to the input current.

(Or you can add another opamp to provide gain & current feedback at the same time, etc.)

Some magnetometers may work in a similar way - eg. how much current in a known coil is needed to null the external field.

Such field strengths are just proportional to current x No. of coil turns; Ampere-Turns.

Proton magnetometers are a completely different animal; in those, a relatively strong bias field is used to magnetically align the atoms in the sensing medium - possibly water!

When the "reset" bias field is cut, the atoms will re-align to the extenal field after some time, depending on the strength of that field.
When they switch back, a small pulse is produced in the bias coil & the time that takes is proportional to the external field strength.

(That's all from memory, so possibly a bit fuzzy - I looked at a DIY design in an electronics magazine back in the 70s?).

Edit - found it! There is a copy of the magazine on the World Radio History site:

The output is a frequency rather than a pulse.
Last edited:
The one showing the amp with a detection winding uses exactly the same principle as the one I described, but as it uses a winding for the magnetic field detection, it is an AC-only device, unlike a hall effect based type.

Any non-zero [error signal] AC field in the core is amplified and fed back out of phase, nulling [or nearly so] the error pickup.

The current used to cancel the error is proportional to the input cable current, related by the relative turns of the input and feedback windings.

This extract from a LEM current transducer datasheet shows the internal arrangement, as part of the connection diagram:

The Ip "winding" is the cable being measured, passed through the window in the device, which acts as the primary winding.
The frame of the unit houses a rectangular ferrous core; some larger versions of these use toroidal cores.

The crossed box is a hall effect sensor that detects any field in the magnetic core - but for AC only, it could equally be another winding on the core.

The output of the error amp in the device passes current through the feedback winding on the core, at whatever level is required to cancel out the field produced by the primary, and that current also passes to the external sense or conversion circuit in whatever the devices are used in.

I was not previously familiar with the "fluxgate" part; from a quick search it appears to work by changes in saturation effects:

That's a bit like how a "Magnetic amplifier" works, something I'm rather more familiar with as they have been used for both current sensing and power control in electric motor drive systems.

One type of drive I've worked on many times over the years used magnetic amplifiers as so-called "DC Current transformers" - one winding with the motor armature current passing through it, and another fed with a fixed AC voltage and with a rectifier and load resistor in series with it.

With DC load current, the AC winding impedance is high and little voltage produced across the load resistor.
As the DC current increases, the core inductance falls and the impedance gradually reduces, so higher AC current and more voltage across the load resistor.

Some machines used much larger mag amps to control massive AC motors; one unit in each of the three phases, with a relatively low current control winding in series through all three mag amps.
do you think there may be audio applications for the magnetic amplifier?.
for example , can it be used as a DAC output stage with linear response without hysteresis ?
Magnetic amplifiers act as a variable impedance to a fixed frequency AC signal, such as mains power, and the response speed is limited by the control winding inductance and core etc..
I cannot see any way one could directly be used for audio?
The power attenuation would be proportional to frequency, rather than linear control across the audio spectrum.

The nearest you could get is two devices controlling high frequency power signals feeding to positive and negative rectifiers respectively to drive an output +/-, which could make a very over-complex and inefficient audio amp.

(I know there is an audio amp in existence that the makers say is a "magnetic amplifier", but as far as I know it does not use the same variable saturation as a true mag amp.

A good Class AB amp, or even well-designed Class D is fine for very high quality audio.

Just look at the specifications of such as a Focusrite Scarlett audio interface - rather better than a lot of ludicrously expensive "audiophile" DACs (plus similar spec inputs & ADCs) for around $150, or ~$100 if you search around ebay etc.
Mag amps are great circuits and we should be using many more of them. I believe most mag amps were used for power control and steering. Like the V1 rocket. Mag amps are incredibly reliable and can take loads of abuse. Great for industrial and military functions. The disadvantage is size and weight. And very high frequencies. Those disadvantages can be mitigated to a great extent today. I have seen some beautiful audio mag amps.(not the modern mag amps) And I have played with mag amp radios. There are several PDFs available from WWII circuitry. The downside is winding. There are no off the shelf mag amp transformers. Like there use to be. You have to roll your own.

I think a mag amp.....was the first digital amplifier. Just an opinion.
I think a mag amp.....was the first digital amplifier.
Magnetic amps are analog, proportional devices - not digital!

And, I've never seen any operating at anything other than mains frequency (50 or 60 Hz)?

I would really like to see any of the supposed Audio or radio ones! I suspect the name is being misused.
This will give you some basic ideas. To show concept.

The "Audio" one works exactly as I hypothesised one could in post #7 - two sections, two rectifiers for push-pull.
Still inefficient and overcomplex as an overall system, as you need to generate the carrier waveform (30KHz in the example, rather more for full audio frequency response), at rather higher levels than the required output power.

To control an existing AC signal or power, they are great; I've worked on many machines that use them, and using other core materials for higher frequency makes perfect sense, though I'd never seen those.

Creating a high power AC signal just to be able to use a mag amp seems a bit of a backwards approach, though..

Latest threads

New Articles From Microcontroller Tips