Announcement

Collapse
No announcement yet.

Multi-frequency EM Induction Sensor

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Multi-frequency EM Induction Sensor

    Hi everybody,

    As most of you have probably seen on this forum, we are now out of the design and prototyping stages of our Mark II project (i.e. a proton magnetometer) and in production phase.
    I plan my next challenging project to be the feasibility study followed eventually by the design and build of a "Multi-frequency EM Induction Sensor".
    I already have accumulated a pile of useful data on that subject and I have defined a number of possible design tracks.
    My simple question is: Is there anybody out there who is willing to make this trip with me?
    In that case, I am ready to uncover what I have already learned about the subject and what is the current status of my project.
    Thanks,

    Willy

    For those who are not yet documented about this, this is short functional description:


    What can an EMI sensor do?

    A magnetometer is used in a geophysical survey to measure magnetic susceptibility variations in earth. It is a passive sensor because it uses the ambient earth magnetic field as the source of excitation.

    An electromagnetic (EM) sensor operating below the radio frequency (RF), or in an EM induction (EMI) mode, is commonly used to measure electrical conductivity variations in earth. (An EM sensor operating above the RF, commonly called ground-probing radar or GPR, can measure variations in dielectric permitivity, which we do not consider here.)

    For this reason, an EMI sensor is often called a conductivity meter. It is an active sensor because it carries its own source of excitation. It has been common to consider that a magnetometer is the principal sensor for measuring magnetic susceptibility, and that an EMI sensor is the principal sensor for measuring electricalconductivity.

    An EMI sensor responds to both electrical conductivity and magnetic susceptibility. In fact, an EMI sensor operating at sufficiently low frequencies acts more as a magnetometer than as a conductivity meter. At the so-called “resistive limit” where the conductivity-frequency product is small, an EMI sensor responds only to magnetic susceptibility and ignores electrical conductivity. It would be a serious misnomer in this case to call an EMI sensor a conductivity meter. A correct designation would be an active magnetometer or, as we propose here, an electro-magnetometer.

    A properly designed electro-magnetometer can serve two simultaneous functions: that of a magnetometer and of a conductivitymeter.

    EMI surveys are used to:
    · Locate buried tanks and pipes
    · Locate pits and trenches containing metallic and/or nonmetallic debris
    · Delineate landfill boundaries
    · Delineate oil production sumps and mud pits
    · Map conductive soil and groundwater contamination
    · Characterize subsurface hydrogeology
    · Map buried channel deposits
    · Map geologic structure
    · Conduct groundwater exploration
    · Locate conductive fault and fracture zones

    See GEM-2 Data examples for a few typical multi-depth, apparent conductivity and apparent susceptibility profiling results.


    Old EMI Technology

    · The EMI instruments were originally made of two separate and quite distant coils due to the influence of the powerful transmitter magnetic field on the receiving coil. This made those instruments difficult to deploy and operate; two operators were necessary. The modern instruments use a bucking coil to cancel the transmitted field allowing assembling the coils in a more compact packaging that can be carried by a single operator.
    · Geometrical sounding versus frequency sounding.
    By technical necessity, most old frequency-domain (FD) EMI sensors have employed the principle of “geometrical sounding” where the coil separation is the only variable since all coils operate at a factory-set frequency. Because these coils are tuned to a particular frequency by their inductance and external capacitance, one cannot change the operating frequency without replacing all coils or associated electronics.
    A tuned coil derives its signal strength from its Q (called the figure of merit—the sharper the resonance, the higher the Q), a voltage amplification factor at the tuned frequency.
    Therefore, the only means left to the user of such a system is changing the coil separation, which often requires multiple operators tending separate coils connected by cables to a measuring console. Furthermore, the system must maintain a considerable coil separation to avoid RX saturation from the primary TX field. It is obvious that such a sensor cannot be made into a small, handheld package.
    In contrast, depth sounding by changing frequency (or “frequency sounding”) measures the earth response at multiple frequencies at a fixed TX-RX geometry. In frequency sounding, there is no exclusive relationship between the coil separation and the depth of exploration.
    · Multi-depth Profiling
    Measurements are taken at successive intervals along a profile with multiple or stepping frequencies. Data are presented as profiles or contour maps and interpreted qualitatively. The depth of the profile essentially depends on the operational frequency used to plot it (lower frequencies show deeper ground layers, higher frequencies show shallower ground layers).

    Objectives

    Advantages of a broadband, multi-frequency EMI sensor are obvious. The idea of using multiple frequencies stems from the so-called “skin-depth”, also known as the depth of exploration, which is inversely proportional to frequency: a low-frequency signal travels far through a conductive earth and, thus, "sees" deep structures, while a high-frequency signal can travel only a short distance and thus, "sees" only shallow structures. Therefore, scanning through a frequency window is equivalent to depth sounding. However, it is also possible to use the I and Q values coming from one particular frequency (i.e. one depth) to produce the 2D apparent Susceptibility and/or apparent Conductivity profile of the survey area at that depth.

    Willy

  • #2
    Hi Willy,
    This EM probe definately interests me as I'm an exploration geologist who likes to build things and find that commercial products are either too expensive, or quite cumbersome to use.
    I may not be of much assistance, but I would like to help out.
    cheers

    Comment


    • #3
      Hi,

      As a geologist, I am sure that you will happen to be of some help later on.
      My latest prototype tests showed that I am on the good track.
      The results show a very good SNR even when executed in the noisy environment of my lab.
      I already made a first real survey in the field with it and that was rather positive.
      I'll come back to you later.

      Willy

      Comment


      • #4
        Hi Willy, how about using it for hunting large meteorites out in fields or the desert?

        John Tomlinson,CET
        John's Detectors

        Comment


        • #5
          Hey Willy,
          That sounds great.
          I was debating whether to build an gamma ray spectrometer, or an EM probe after I finish this magnetometer that I am currently working on (waiting for PCB).
          I was trending toward one where the receiver and transmitter coils were separated for use with two people, as I have no experience with one of these small ones mounted all within one probe. But, it definitely looks cool to get it all into one pole.
          So what kind of SNR are you are getting with your prototype?
          cheers

          Comment


          • #6
            Originally posted by Wirechief View Post
            Hi Willy, how about using it for hunting large meteorites out in fields or the desert?

            John Tomlinson,CET
            John's Detectors
            That would be ideal to look for the metallic meteorites but a PPM would also work well as most of the metallic meteorites are also made of a large percentage of magnetic material.

            Willy

            Comment


            • #7
              Originally posted by willemite View Post
              Hey Willy,
              That sounds great.
              I was debating whether to build an gamma ray spectrometer, or an EM probe after I finish this magnetometer that I am currently working on (waiting for PCB).
              I was trending toward one where the receiver and transmitter coils were separated for use with two people, as I have no experience with one of these small ones mounted all within one probe. But, it definitely looks cool to get it all into one pole.
              So what kind of SNR are you are getting with your prototype?
              cheers
              Hi,

              As in all professional work, these is a specific tool for each specific task.
              - If you want to make geological sounding type of surveys to locate water beds or pollution plumes, your best bet is still an FDEM system with variable separation of coils to measure the changes of ground conductivity in various depth layers.
              - If, like myself, you want to make profiling surveys at more reasonable depths of maximum 2 meters, then the double coil, multi-frequency FDEM has proven to be effective enough if carefully designed and built.

              My aim at starting this EMI project was to complememt my own PPM tool to detect various underground archaeological structures like wall foundations, wells, fire pits and bronze/ceramic ovens. It happens that the foundations stones used by roman armies in my region have a magnetic susceptibility very close to the one of the surrounding soil and thus, are difficult to detect by a PPM alone. My hope is that the EMI system will be able to measure both the ground apparent conductivity and magnetic susceptibility at multiple depths with a single survey trip.

              Note that there is no rocket science in designing and building such a system.
              a. Transmitter module
              It must generate a constant multi-frequency signal inducing a powerful induction on the ground. Thus, we need a signal processor using a DAC or PWM to mix several sine wave signals into a common digitally-coded stream. I have limited the number of frequencies ot 3 and I set the individual sine wave magnitude to share the maximum power to be proportional to p=1/f.
              The mixed stream then goes through a power audio class-D 4-channel amplifier giving a max of 28W per channel. I have connected the four channels in parallel to get a maximum power on a coil of 1 ohm.
              This power amplifier is I2C-controlled to set various control states and parameters including its gain.
              b. Receiver module
              The receiver must provide two separate broadband receive channels connected one to the bucking coils alone and serving as reference channel and the other connected to the receive + bucking coils in series (called the 'bucked coil').
              The 'bucked coil' channel goes through two stages of LNA (low-noise amp) followed by an I2C-controlled PGA stage (Programmable Gain Amp) and ended by a fast 24-bit ADC while the reference channel only goes through PGA+ADC stages.
              The receiver module is also the main controller, dtat storage (SD card) and human interface of the system. It also controls the transmitter module through an I2C link.

              The really critical part of the whole game is the signal processing of the two recieve channels.
              In principle, as people experienced in MD technology know well, the variations of magnitude of the signal measured at the 'bucked coil' give a measure of the variations of ground apparent conductivity. This measure is used in MD's to define the all-metal detection mode and beep in presence of a large variation (gradient) of it. In the same way, the positive or negative phase shift of the receive signal on the bucked coil measures the variations of magnetic susceptibility. This measure is used in MD's to define the metal discrimination between magnetic and non-magnetic targets.

              Thus, the problem is to calculate those two values for each transmitted frequency at a high enough rate to be compatible with a walking survey operator.
              This requires a fast digital processor and optimized DFT algorithms. With the 32-bit 60MHz ARM processor I am using, I can calculate these values and store them on SD card with a rate of four full cycles per second. Note that all programming is written in C, no assembly code!!

              Hope this will give some hintsight on this project.

              For the SNR, the best info I can give you is the noise spectrums as calculated by the system itself over the full audio range of frequencies. Attached are three files showing frequency spectrums of noise levels measured in my lab, in my backyard and in my garden. You can see that the general level of noise is quite low except for specific frequencies which are induced by environmental conditions. I guess some of them (around 15KHz) are signals coming from the horizontal oscillator of TV sets.

              Willy
              Attached Files

              Comment


              • #8
                Amplifier needed

                Willie, do you know where I can buy a amplifier to connect to my function generator, sine wave, 1-watt, 12V, I will use 16 guage wire laid out in a square for the antenna. I am needing this built or bought for my EMFAD receiver. It was manufactured to receive VLF in the 44-142 khz range, frequency being supplied by some old Naval transmitters, which it will not find any of these signals now. The newer model has it's own tx and it is built similar to what I am describing. I wish to put something together and experiment with it. I will hook it to an old BK 3031 and see what happens. I have advertised for someone to build me one here but no takers, surely there is an amp out there that will suit my needs. I am not a tech, and anything you will tell me will be over my head, but as I said, maybe there is a simple solution. Thanks. MB

                Comment


                • #9
                  Originally posted by Black Cloud View Post
                  Willie, do you know where I can buy a amplifier to connect to my function generator, sine wave, 1-watt, 12V, I will use 16 guage wire laid out in a square for the antenna. I am needing this built or bought for my EMFAD receiver. It was manufactured to receive VLF in the 44-142 khz range, frequency being supplied by some old Naval transmitters, which it will not find any of these signals now. The newer model has it's own tx and it is built similar to what I am describing. I wish to put something together and experiment with it. I will hook it to an old BK 3031 and see what happens. I have advertised for someone to build me one here but no takers, surely there is an amp out there that will suit my needs. I am not a tech, and anything you will tell me will be over my head, but as I said, maybe there is a simple solution. Thanks. MB
                  Hi,

                  You do not need to go so high in frequency for an FDEM system; 0.5KHz to 25KHz is the whole interesting range.
                  I am using the 4-Channel Automotive Digital Amplifier TAS5414A from TI as power amplifier for the transmitter.

                  Willy

                  Comment


                  • #10
                    Thanks for the timely reply, Willy. The EMFAD receiver will not receive anything below about 44 khz. Shouldn't the frequency generator and the amplifier be the same? The instrument (EMFAD) is capable of 44-142 khz reception, but the lower the freq, the deeper it goes ( At the lower #, I will lose a little resolution in the imagery, but since I am looking for caves, etc. so the choice is as deep as possible. Were you familar with the little BK 3031 generator? I had it from another project years ago, but the whole idea is to make this transmitter as lightweight as possible for backpacking reasons, so I'm hoping it is acceptable. Mike

                    Comment


                    • #11
                      Hey Willy,
                      Any progress on your EM thing?

                      Comment


                      • #12
                        Not much yet.
                        I was still kept quite busy on our PPM project but I'll go back to this project soon.
                        Willy

                        Comment


                        • #13
                          Hello, Willy. It's good to see experiments in this geophysical area.
                          But I must notice that GEM-2 won't fully work as planed and claimed: http://www.geophex.com/GEM-2/How%20i.../operation.htm
                          There are some moments in GEM-2 construction, which corrupts overall model:
                          1) Fixed coils spacing and multifrequency work (with higher F=24khz)
                          2) Relatievle low working frequencies (large skin depth)
                          3) Frequency domain mode
                          4) etc.

                          In summary device will work, but it's suitable only for metal detection (not for target, depth estimation).

                          Some drawbacks may be fixed, but another will arise. In summary I think following things must be implemented:
                          1) Wider frequency range (up to 1Mhz) with coils spacing significantly higher than 1Mhz skin depth.
                          2) Time domain mode

                          Or use proven but complex and costly methods (earth layering resolve, building apparent resistivity etc.): MT, AMT, CSAMT.

                          Comment


                          • #14
                            Multifrequency FDEM

                            My work involves the use of a multifrequency FDEM instrument (37.5Hz-2400Hz.) Whilst there is no direct correlation between frequency response and depth, the different channels do respond to different targets - the lower frequencies corresponding to deeper targets.

                            There is definitely some value in developing this instrument. As all useful frequencies lie in the audio range - I suggest the use of a 12V auto amplifier in the transmitter (TDA2030), and audio op amps in the receiver signal path (OPA2134.)

                            Good magnetic loop receiver designs can be found on the ELF forums (such as the easyloop.)

                            Comment

                            Working...
                            X