ena NE a Sat hl ern SOE Fluxgate magnetometery ational magnetometers use induc ive loops and long wire lines to measure magnetic flux by induced P, Such instruments are bulky. Hall effect semiconductors will measure low flux levels although their response tends to bbe non-linear and temperature dependent The fluxgate magnetometer, which depends for its action on the detection of saturation ‘occuring in magnetic material, ean be made booth small and highly accurate. In explain ing its operation, itis easier 19 describe & simpler arrangement than would actually be ‘sed in practice. In its most basic form, it comprises asin tle straight nickel-iron alloy core earying "0 windings. One Winding functions a8 an excitation colin srhich current Hosving cre- ‘es afield to magnetis the core in alternate directions, The other acts as pickup coil producing a voltage proportional to the rate ‘of change of magnetic fix linking it [At low levels of excitation this structure, cutlined in Fig. 1, will obviously behave as iT were an ineliient transformer. To con vert itto # fuxgate transducer, the excitation is increased so as to force the core into sate ration on alternate peaks, This would be undesirable behaviour in a transformer but is the essence of fluxgate operation ‘A square voltage waveform is applied 10 Measuring the direction of the Earth’s magnetic field provides the basis of an electronic compass while the strength of the field says much about solar activity, a factor important to HF radio propagation. Richard Noble describes the heart of an instrument which can be used for serious scientific study. the excitation coil, of sufficient magnitude ' wea ae ‘osaturate the core. During the periods when the core is not saturated, the inductance of the col causes the current passing through it {o change linearly from one saturation state (o the other. As the core reaches saturation the inductance falls to @ low value and the current rises rapidly toa limit set by the DC resistance of the coil forcing the core well imo saturation, The waveform of this exerent is shown in Fig. 2 which also serves as an illairation of the magnetic fleld produced Dy the col, since this must be directly pro- portional tothe current. "The induction in the core is magnified by the high permeability, except in the sarura- sion regions where the result can be seen 10 be a trapezoidal induction waveform (Fig. 3). The voltage induced in the pickup coil by this waveform is directly proporional to the rate of change of magnetic induction. During the periods of saturation, there is no change and no signal voltage, but during the linaely changing regions thee isa con stant volage output of appropriate polarity, as shown in Fig. 4 ‘All of the above sxsumes that the only influence on the core is the magnetic Feld produced by the excitation col. If this is nat {he ease and the core is affected by an add tional external field, a small change takes place inthe output. The component of exter CET ELECTRONICS WORLD + WIRELESS WORLD September 1991Tr lonis one ofthe important aspects of aio ee eee eee related a solar activity. Amagretomele, detecting changes inthe fel aig ho charged arcs dled ound ath, measures Me iar and non-corusculst Pee retin cn vancas "Non-corpuseul radiation isin the for of high enetgy Xeays which, assuming thatthe earth is inthe path of the rays, wil teach dhe earth ‘ton minutes, This radiation may incase the depth ofthe "layer diet ionisation and prodice the ll too wellknown Dellinger fode ‘ur when longistance HF communication cestes abruptly, ‘Corpusculat of parculate radiation appea's inthe form of protons And neitons whieh tke longer each the cats uppes aimosphers than the haher eheiay adiaions. These nuclear paricles ave at he FF and f2 layers approximalely 48 hours following a solar event and ‘produce ionisation by cal ding wih gaseous molecules and cosmic [articles This explains why 5 few protons are detectable ate earth's Suflace excep afer very msjorevente 5 “Te lncreated ionisation alg the eats ines of magretc force make them mare conductive and leads 1 an Increased cutent flow bee DESIGN Init iho aston, vain etal ices he ionosphere infuenees el 1 welecnce of adlo eee: Iebien Bands will ba more productive. tis interesting to note hata Dalligaee ae 04 & ofan fllowed bya magnetic fed change round, Abou ite ‘Many ales have apeared in the amater ado press ard ‘elsewhere desing thi many aspects of propagaion ard giving more ‘desalied information onthe thegry of and tees of slar radiation on “comimnjcaions systems, This bee and snplifid inoducton has ld for cornoetenes, and hapeully to show that the ‘oarination ofthe cars mayncosprere av elaon to propagation is ota dificul project io undertae, the equipment is est to build and fas The advantage of beg a both teresting and useful ation to your capability, Dasid Lomax GWOPKA ‘wih consequent rie i rmayetic Hux lets. ts unde these conditions thatthe magnetometers uslul in detecting te fluctuations Tnthe magnetic feds. ral field which isin ine with the core axis cither aids or opposes the exitation field in its alternating polarity phases. When it aids the excitation field, the entry into saturation ‘occurs slightly earlier and the departure from saturation occurs slightly later than would be the case. The opposite effect arses Wwhen the external field opposes the excita tion Field. For clarity, this is illustrated in very exaggerated form in Fig. §. The Amount of advance and delay in the wave form corner is proportional (othe size ofthe external field. Tn this way the desired objective has been achieved in the form of a signal variation ‘whieh isa function ofthe extemal magnetic field strength, It remains to exploit this. Since a perfectly symmetric waveform does ‘not contain even harmonics and an asym: rmettic one does, a possible technique would be to isolate the even harmonics as the usable signal However the presence of very large Funda- ‘menial and odd harmonic components inthe ‘utpat makes the isolation of the small even Iharmonics a very daunting task. A Better Technique Fortunately there is very effective simple solution in the form of two parallel cores ‘with opposing excitation windings and asin- file overwound pickup coil (Fig. 6). The Indvetions in the two cores cancel out pre= cisely when the cores are placed in a zer0 extemal field an external field causes asym. descebed in this article. metric pulses of amplitude and position ‘dependent onthe polarity and strength ofthe ‘extemal field, The resultant wavetorms are shown in Fig. 7. In this way, the trick of using two cores hus performed the seemingly impossible task of isolating the tiny wanted signal from the comparatively hoge unwanted one, Ifthe pultes marked with arrows in Fig. 7 are s0- lated from the others and applied to an appropriate iow pass fer, the output is a DC or slowly varying voltage whose mas tude and polarity model the external field. It ‘would appear that isolating the se of pulses rt marked with arrows would work equally Welland, in Fact, combining both in sui able way could double the sensitivity. For Tow gain systems such as fluxgate compass- 5 or short range metal detectors this is true, but for high sensitivity application such as the magnetometer described here, there are ‘good reasons to avoid this approach, ‘A Practical Solution All of the foregoing assumes perfectly matching cores and windings and in practice ‘snot easily achieved without indivigual and fiddling adjustment. The problems associat ced with the need for perfect maiching ean be avoided by one further design change. The long straight cores are abandoned “They are instead hent into semicircles and joined together a thei ends to form s solid Circular toroid. The two opposing excitation coils merge into a simple, single, full The Srace on the elt and page shows the disturbances caused to the the earths 2 sola toe occurring between Jane 12. ‘and 14, 1991, The measurements were made withthe sytem toroidal winding, The pickup coil remains a single winding over the toroid (Fig. 8). By ‘hese modifications the sought-after simpli lty is achieved, coupled with the virtual climination ofall high-level signals. As an additional benefit the closed magnete cir cuit of a toroidal core greatly reduces the drive requirements to produce saturation levels, simplifying the circuitry needed by the sytem, Te may be felt chat replacing two straight cores with a ring structure is too drastic « change to gloss over, But the mechanism is not (00 diferent in the two cases. Figure 9 represents the way in which the lines of force are concentrated from the immediate vicinity and makes plausible the idea that an fverwound coil would not ce much differ tence in the fax changes it experiences from either system. "An additional advantage ofthe ring core is that it wil aceept mote than one overwound pickup coil and they can have ifferent of fentations tothe extemal magnetic field. For texample while the coil orientation shown in Fig. 10a produces the largest ouput, that 10b Links none of the changing flux and has @ null output. Angles between these produce fan output which varies as the cosine of the angle, leading to the familiar igure-of- ritudes of these two readings should be the same, ifthe zero-ffset has been set correct Iy. This is not as difficult to do as it sounds, simply because of the very broad maxima ‘nd minima If the intention is 10 make serious mea- surements calibration with a Helmholtz coil ‘will be required, The outlined design pro duces a field of 395 gauss/ampere. Strictly {his implies a precision greater than is rea sonable and should probably be read as 4.0 ‘gaussfampere or 400 gamma/mA (one gauss 100 000 ganna). ‘The sensitivity sought atthe test point is 20 000 gammaWolr: #2.5V coresponds to a ange of 100 000 gamma so the initial Helmholt col eurent setting should cover this. At 400 gammalimA this ealls fora cur rent of 125mA variable. This should include 4 reversing switch to enable the current through the coils to be reversed easily Place the sensor in the centre of the Helmholtz coils with one pickup winding aligned to pick up maximum flux. Monitor the appropriate amplifier test point with a voltmeter or calibrated oscilloscope and rotate the Helmholtz coils and sensor to give 2er0 volts. Switch on the col current (Set {0 125mA) and the test point voltage should change by about two vols of 0, positively or negatively. Using the amplifier gain tin pat, adjust so that ee output changes by SV when the coil eurent is reversed. ( 42.5V range). Since the final stage gains are set by 1 tolerance fixed resistors, this completes the calibration, the output sensitivities being either 1000 gamma/volt or 100 gamma/vol Repeat the exereise to calibrate the second pickup winding and the instrument is ready for use . The core used inthis design isa Teleon Metals HCR ally core type 7a and ts ‘available from Telcon Metals, phone 0203 528800 ELECTRONICS WORLD + WIRELESS WORLD September 1991