Geotech Page 1

Ele

ctro

nics

Now

, Jul

y 19

98 —

Cop

yrig

ht

Ger

nsba

ck P

ublis

hing

, rep

rodu

ce f

or p

erso

nal u

se o

nly

This month, we put our flux-gate sensors to work

and build a practicalmagnetometer and

gradiometer.

JOSEPH J. CARR

Analog Interface to FGM-3. Figure 2 shows a method for providingan analog interface to the FGM-3 andits relatives. The output of the sensor isa 40- to 125-kHz frequency that is pro-portional to the applied magnetic field.As a result, we can use a frequency-to-voltage (F/V) converter such as theLM2917 to render the signal into a pro-portional DC voltage. That voltage, inturn, can be used to drive an analog ordigital voltmeter or milliammeter. TheLM2917 is used here because it iswidely available at low cost from mail-order parts distributors.

The output circuit consists of a bridge

made up of R1-R3 and the outputof the LM2917. A resistive volt-age divider (R2/R3) produces apotential of 1/2(V+) at one end ofthe 22,000-ohm sensitivity control(R4). If the voltage produced by

the LM2917 is the same as the volt-age at R4, then the differential voltage

is zero and no current flows. But if theLM2917 voltage is not equal, then a dif-ference exists and current flows in R4and Ml. That current is proportional tothe applied magnetic field. Meter Mland R4 can be replaced with a digitalvoltmeter, if desired.

The DC power supply uses two regu-lators, one for the FGM-3 and one forthe LM2917. Even better results can beobtained if an intermediate voltage reg-ulator, say a 78L09. is placed betweenthe V+ source and the inputs of IC2 andIC3. That results in double-regulation,and produces better operation.

Digital Heterodyning. A different interface method is shown

in Fig, 3. It results in a more sensitivemeasurement over a small range of thesensors total capability. In other words,that circuit makes it possible to mea-sure small fluctuations in a relativelylarge magnetic field.

In that circuit, a D-type flip-flop isused to “mix” the frequency from theFGM-3 with a reference frequency(FREF). The FGM-3 literature calls thatprocess “digital heterodyning,”although it is quick to point out that it isreally more like the production of aliasfrequencies by undersampling than trueheterodyning.

Two types of frequency sources canbe used for FREF. For relatively crudemeasurements, such as a passing-vehi-cle detector, the CMOS oscillator ofFig. 4A is suitable. This circuit is basedon the 4049 hex inverter connected inan astable multivibrator configuration.

In the June, 1998 issue ofElectronics Now, we intro-duced the FGM-x series of flux-gate magnetometer sensors fromSpeake & Co. Ltd. These rela-tively low-cost devices are madein the United Kingdom, but aredistributed in the United States byFat Quarters Software (24774 Shos-honee Drive, Murrieta, CA 92562; Tel:909-698-7950; Fax: 909-698-7913;Web: www.fatquarterssoftware.com).For those not with us then, lets quicklyreview some of the important featuresof the FGM-3, which is used in theprojects that are described in this article.

The FGM-3 is a three-terminal device(see Fig. 1 A) that will measure mag-netic fields over the range 0.5 oersted(50 µteslas). It can be used in a widevariety of magnetometer and gradiome-ter circuits. The three terminals on theFGM-3 are: RED: POWER (+5 VDC);BLK: GROUND (0V); and WHT:SIGNAL.

The output signal is a train of TTL-compatible pulses with a period thatranges from 8 to 25 mS, or a frequencyrange of 40 to 125 kHz. The magneticfield strength is indicated by the fre-quency produced, and can be read on avariety of display devices including dig-ital frequency counters, digital periodcounters, analog meters, and even com-puters. It is relatively easy to interfacethe FGM-3 device to microcontrollers— such as the Parallax BASIC Stampor Micromint PicStic devices — using aline of interface chips offered bySpeake and Fat Quarters.

The sensitivity pattern of the FGM-3device is shown in Fig. 1B. It is a fig-ure-8 pattern with its maxima along themajor axis of the FGM-3, and minimaorthogonal to the major axis. At otherangles, the sensitivity drops off as thecosine of the angle from the major axis.

Geotech Page 2

Ele

ctro

nics

Now

, Jul

y 19

98 —

Cop

yrig

ht

Ger

nsba

ck P

ublis

hing

, rep

rodu

ce f

or p

erso

nal u

se o

nly

The exact frequency can be adjustedusing R2, a 10,000-ohm, 10-turnpotentiometer.

Where a higher degree of stability isneeded, for example when makingEarth-field variation measurements ortesting magnetic materials, a more sta-ble frequency source is needed. In thatcase, use a circuit such as the one inFig. 4B. That circuit uses a crystal-con-trolled oscillator feeding a binarydivider network. Crystal oscillators canbe built, or if you check the catalogsyou will find that a large number of fre-quencies are available in TTL- andCMOS- compatible formats at lowenough costs to make you wonder whyyou would want to build your own. I’veseen them sold for about the price of acrystal alone in many mail-order cata-logs (Digi-Key, etc.).

The reference frequency is adjusted to

a point about 500 Hz below the meansensor frequency. That frequency ismeasured when the sensor is in the east-west direction. That arrangement willproduce a frequency of 0 to 1,000 Hzover a magnetic field range of 500gamma.

A Magnetometer Project. Figure 5 shows the circuit for a sim-

ple magnetometer based on the FGM-3flux-gate sensor. A PC board, the sen-sor, the ICs, and most other parts (butno power supply, enclosure, or switchS1) can be obtained from Fat QuartersSoftware (contact them directly or seetheir Web page for more information).That circuit takes the output frequencyof the FGM-3, passes it through a spe-cial interface chip (IC1), and then to adigital-to-analog converter to produce avoltage output.

The heart of the circuit, other than theFGM-3 device, is the special interfacechip, Speake’s SCL006 device. It pro-vides the circuitry needed to performmagnetometry, including Earth-fieldmagnetometry. It integrates field fluctu-ations in one-second intervals,producing very sensitive output varia-tions in response to small fieldvariations. It is of keen interest to peo-ple doing radio-propagation studies, andthose who need to monitor for solarflares. It also works as a laboratorymagnetometer for various purposes. TheSCL006A is housed in an 18-pin DIPIC package.

The D/A converter (IC2) is an Ana-log Devices type AD557. It replaces anolder Ferranti device seen in the Speakeliterature because that older device is nolonger available. Indeed, being a Euro-pean device, it was a bit hard to find inunit quantities required by hobbyists onthis side of the Atlantic. The kit fromFat Quarters Software contains all thecomponents needed, plus a printed cir-cuit board. The FGM-3 device is boughtseparately.

The circuit is designed so that it couldbe run from 9-volt batteries for use inthe field. A sensitivity switch, S1, pro-vides four positions, each with adifferent overall sensitivity range. Theoutput signal is a DC voltage that canbe monitored by a stripchart or X-Ypaper recorder, voltmeter, or fed into acomputer using an A/D converter.

If you intend to use a computer toreceive the data, then it might be worth-while to eliminate the D/A converterand feed the digital lines (D0-D7) fromthe SCL006A directly to an eight-bitparallel port. Not all computers havethat type of port, but there are plug-inboards available for PCs, as well as atleast one product that makes an eight-bit I/O port out of the parallel printerport.

Geotech Page 3

Ele

ctro

nics

Now

, Jul

y 19

98 —

Cop

yrig

ht

Ger

nsba

ck P

ublis

hing

, rep

rodu

ce f

or p

erso

nal u

se o

nly

Gradiometers. One of the problems with magnetom-

eters is that small fluctuations occur inotherwise very large magnetic fields,and those fluctuations can sometimes beimportant. A further problem with sin-gle-sensor systems is that they are verysensitive to orientation. Even a smallamount of rotation can cause unaccept-ably large, but spurious, output changes.The changes are real, but are not the

fluctuations that you are seeking.A gradiometer is a magnetic instru-

ment that uses two identical sensors thatare aligned with each other so as to pro-duce a zero output in the presence of auniform magnetic field. If one of thesensors comes into contact with somesort of small magnetic anomaly, then itwill upset the balance between sensors,producing an output. The gradiometergets its name from the fact that it mea-

sures the gradient of the magnetic fieldover a small distance (typically 1 to 5feet).

These instruments can be used forfinding very small magnetic anomalies.For example, the metallic firing pin ofplastic land mines buried a few inchesbelow the surface, or a shipwreck bur-ied deep in the ocean silt. Archeologistsuse gradiometers to find artifacts andidentify sites. Also, people who exploreCivil War battlefields, western miningcamps, and other sites often use gradi-ometers to facilitate their work.

Figure 6 shows the constructiondetails for a simple gradiometer basedon the FGM-3 device. It is built using alength of PVC pipe. One sensor is per-manently mounted at one end of thepipe, using any sort of appropriate non-magnetic packing material. In oneexperiment, I used standard 0.5-inchadhesive-backed window-sealing tape,which is used in colder areas of thecountry to keep the howling winds outof the house in wintertime. It workednicely to hold the permanent sensor inplace.

The other sensor is mounted in theopposite end of the tube using an O-ringthat fits snugly into the tube. Four posi-tioning screws made of non-magneticmaterials are used to align the sensor.

Geotech Page 4

Ele

ctro

nics

Now

, Jul

y 19

98 —

Cop

yrig

ht

Ger

nsba

ck P

ublis

hing

, rep

rodu

ce f

or p

erso

nal u

se o

nly

The position of the sensor is adjustedexperimentally. The idea is to positionthe sensor such that the gradiometer canbe rotated freely in space without caus-ing an output variation.

The gradiometer sensor is usuallyheld vertically such that the end withthe wires coming out of the FGM-3devices is pointed downwards. Thatalignment allows you to find buriedmagnetic objects even if they are quitesmall.

A practical gradiometer can be builtusing a special interface chip bySpeake. the SCL007 device. It is an 18-pin device that accepts the inputs fromthe two sensors, and produces an eight-bit digital output. The circuit is shownin Fig. 7. It can receive the signals fromthe sensors in Fig. 6, and produce a dig-ital output proportional to the fieldgradient. If you want a DC output, thesame sort of D/A converter used in themagnetometer of Fig. 5 can also be usedfor the gradiometer.

Figure 8 shows a method of usingdigital heterodyning to make a very sen-sitive gradiometer at low cost. Theoutputs of the two FGM-3 sensors arefed to the D input and clock (CLK)input of a D-type flip- flop. The outputof the flip-flop is fed to an F/V con-verter such as the LM2917 devicediscussed earlier.

Conclusion. As we’ve seen, magnetic sensor

projects are relatively easy to buildwhen the FGM-3 sensor is used. Thedevice is well behaved, and will servenicely for both hobbyist and profes-sional instruments, as well as forclassroom demonstrations.