Low noise fundamental mode orthogonal fluxgate (FM-OFG) magnetometer built with an amorphous ribbon core.

Room temperature operating magnetometers able to detect a few pT magnetic field would play a crucial role in such areas as bio-magnetic field detection and very tiny ferromagnetic contaminats detection in battery production lines. We have already reported fundamental mode orthogonal fluxgate having a noise density $\sim 3$ pT$/ \surd $ Hz at 1 Hz using a Co-based amorphous wire (made by Unitika) core sensor head [1]. However, the supply of the amorphous wire is limited. In this paper, we present that the narrow amorphous ribbon core is an alternative to the wire core if the dc bias current for the fundamental mode operation [2] is properly adjusted. In a particular case of as-cast Metglas 2714A ribbon of 1 mm width, the dc bias current elevated to 200 mA, which is to be 40 mA for the wire core sensor head, reduces the noise down to a few pT$/ \surd $ Hz. In the experiment, we made several sensor heads listed in Table I using the above mentioned ribbon. A 1 mm width ribbon which is made by slitting a wider sample is bent in a U-shape or more likely in a hair-pin shape and inserted into a cylindrical pickup coil of length 30 mm, bore diameter of 2 mm and having 500 turn copper winding. The driving and signal processing electronics used is the one developed for the wire core sensor head. The only difference is that an adopter circuit is inserted between the ribbon core sensor head and the driving electronics to supply a large dc bias current from the battery and to moderately $( \approx 2\mathrm {x})$ amplify the ac excitation current; the ac excitation current for ribbon core sensor heads is 24 mArms while that for wire core one is 12 mArms. The driving frequency is 100 kHz and the magnetometer is configured to operate in a closed loop (feedback) using $\mathrm {a}10 \mathrm {k}\Omega $ feedback resistor. The sensitivity increases slightly with increase in length of the core, because the demagnetization effect becomes smaller for longer cores. The sensitivity of the magnetometer with the wire core sensor head is about one half of that of the magnetometer with the ribbon core sensor head because the number of turns of the pickup coil used for the wire core is 1000. It should be noted that the sensitivity of the closed loop system is $R/ n[\mathrm {V} /\text{A}$/m] where $R$ is the feedback resistance and $n$ is the winding density per meter of the pickup coil. The noise spectral density of the magnetometer was measured by placing the sensor head in a five shell cylindrical shield and by using an FFT analyzer (Stanford Research Systems SR780). Results are listed in Table I from which we can conclude that the noise level obtained with a ribbon core sensor head is comparable to that obtained with a wire core sensor head. Fig. 1 shows the noise spectral density of the ribbon core sensor head R19 for three different dc bias currents. It is interesting to see that the noise level is dramatically reduced by increasing the dc bias current. This is probably related to the nature of the as-cast Metglas 2714A ribbon such that weak magnetic anisotropies are distributed randomly in the ribbon plane. It is also interesting to compare the magnetic field produced by the dc bias current at surfaces of the ribbon core and of the wire core. For the wire core, the radius is about $60 \mu \mathrm {m}$ and the bias current is 40 mA, hence the magnetic field is about 106 A/m. For the ribbon core, because the thickness $( < 20 \mu \mathrm {m})$ is negligible compared to the width (1 mm), the magnetic field at the surface of the ribbon is quite uniform along the width direction, hence the magnetic field can be calculated by dividing the current (220 mA) by the peripheral length which is twice the width, yielding 110 A/m which is similar to the one for the wire. One may expect that necessary dc bias current can be reduced by making the width of the ribbon narrower. A very weak magnetic field of rectangular waveform of 15 pT in peak-to-peak is well detected with a short averaging. Details on the adapter circuit will be explained.