An explanation for the magic low frequency magnetic field shielding effectiveness of thin conductive foil with a relative permeability of 1

Transcription

1 An explanation for the magic low frequency magnetic field shielding effectiveness of thin conductive foil with a relative permeability of 1 D.A. Weston K McDougall (magicse.r&d.doc) The data and information contained within this report was obtained from an independent R&D project funded by EMC Consulting Inc. The contents may be used and quoted but the source must be referenced in any publication. 1) Introduction A previous investigation entitled Shielding a room using aluminum foil examined the feasibility of using thin foil as an effective barrier to EMI and determined that, counter to conventional theory, thin aluminum walls can provide significant shielding of magnetic fields. This report examines the measurement data and attempts to explain the mechanism of magnetic field shielding that occurs. 2) The enclosure under test For these tests an enclosure was built with the dimensions 1m x 1m x 0.7m high. The frame of the box was constructed of wood and the walls of aluminum foil 0.076mm (0.003 ) thick. Four walls of the enclosure were covered by a continuous sheet of foil, clamped together where the ends meet between two strips of brass screwed into wooden supports. The two remaining sides of the enclosure were covered by individual sheets of foil. At the seams, the foil edges were overlapped and stapled every 3cm to the wooden frame and then covered by packing tape. On one side of the box the foil is covered in a brass plate. A rectangular aperture is placed in the brass and foil, and this is covered by a door. The door has a Spira gasket around all four sides. 3) Magnetic field shielding effectiveness measurement data For the first set of measurements, a multiturn RS101 H-Field loop was used as the transmitting (tx) antenna and its center was located 0.5m from the front face of the foil enclosure. The loop was oriented such that the magnetic field lines incident on the front face of the box would be horizontally oriented, resulting in a current that travels vertically on the enclosure wall. A 13cm H-field loop was used as the receiving (rx) antenna inside the enclosure and mounted at distances of 0.3m, 0.5m and 0.8m from the front wall of the enclosure. See photo 3.1 for a view of the enclosure with the multiturn 1

2 RS101 transmitting loop and photo 3.2 for a view of the 13cm H-field loop receiving antenna inside the enclosure door. Photo 3.1 The enclosure under test with the transmitting antenna 2

3 Photo 3.2 The receiving antenna inside the enclosure under test. From 400Hz to 200kHz the field picked up by the rx antenna inside the enclosure was measured at each of the 3 distances. A current probe was used to monitor the signal going into the tx antenna to ensure that the same field was generated in all of the tests. These measurements were compared to the reference measurements taken at the same distances, but without the shielded enclosure. The difference between the reference measurement and the shielded measurement is the shielding effectiveness (S.E.). Similar shielding effectiveness was also found using another technique, which compares surface currents. For these measurements a 560 turn surface current probe was placed at the same relative positions on both the exterior and interior wall of the enclosure illuminated by the tx antenna. The ratio of the current measured on the outside and inside walls of the enclosure is the shielding effectiveness. Results from both shielding effectiveness tests are shown in figure

4 S.E. (db) Aluminum foil enclosure 1m x 1m x 0.7m magnetic field S.E. External loop at 0.5m Magnetic field S.E. Internal loop at 0.3m from w all Magnetic field S.E. w ith internal loop 0.5m from w all Magnetic field S.E. Internal loop 0.8m from w all" 100 1,000 10, ,000 1,000,000 External to internal current Frequency (Hz) Figure 3.1 Measured magnetic field shielding effectiveness of the aluminum foil enclosure versus surface currents. In addition to the change in S.E. with location of the rx antenna inside the enclosure shown in figure 3.1, a further analysis of the change in S.E. at the spot frequency of 5.1kHz with the enclosure far removed from conductive surfaces on a free space range is shown in Figure 3.2 and from 70kHz to 200kHz in figure 3.3. These tests were to eliminate errors due to field perturbation as a result of proximity to conductive surfaces. 4

5 Magnetic Field Shielding Effectiveness at 5.1kHz on the OATS of an Aluminum Foil Enclosure, 1m x 1m x 0.7m Multiturn RS101 tx antenna at 0.5m from enclosure wall 30 S.E. (db) Distance from wall to rx antenna inside enclosure (m) Figure 3.2 Effect on S.E. at 5.1kHz of moving rx antenna. Measurements made outside of building Magnetic Field Shielding Effectiveness on the OATS of an Aluminum Foil Enclosure, 1m x 1m x 0.7m 10T RS101 tx antenna at 0.5m from enclosure wall S.E. (db) m inside m inside m inside Frequency (khz) Figure 3.3 Effect on S.E. from 70kHz to 200kHz of moving rx antenna. Measurements made outside of building on a free space range. 5

6 Counter to the theory of reciprocity, we observe that the shielding effectiveness we measure from the aluminum foil enclosure is very slightly affected by the location of the receiving antenna inside the enclosure. The shielding effectiveness seems to increase the farther we move the rx antenna from the tx antenna. Although the field strength from the tx antenna decreases with distance, this fall off should not affect the shielding effectiveness because it is a comparative value. The reduction in H field outside and inside the enclosure is a function of 1/r3, where r is the distance from the source. In the reference field measurement the source is the transmitting loop, typically 0.5m from the outside wall of the enclosure. If we postulate that the source for the field inside the enclosure is the current flow on the wall then we see that the reduction in the resultant field with distance is greater than the field from the transmitting antenna (the reference field). As an example the reduction in the reference field from 0.8m from the source to 1.3m from the source is 20 log (0.8m/1.3m)3 = dB. The field inside the enclosure with the transmitting loop at 0.5m from the enclosure, and assuming the wall is the source, reduces from 0.3m to 0.8m as 20 log (0.3m/0.8m)3 = -25.6dB a difference of 13dB. We do not see this magnitude of change in our measurements and we discuss why this hypothesis is not correct in section 4. We have proved by measurements that the distance of the transmitting antenna from the wall of the enclosure does not affect the S.E. as predicted by the theory of reciprocity. Thus attenuation due to reflection, which depends on wave impedance, does not explain our test results. This independence with source location is very different from the results using the classic approach to magnetic field shielding, which uses a transmission line analogy involving the difference between the incident wave impedance and the metal impedance as well as absorption through the metal to achieve the S.E. Using this method of analysis the predicted S.E for our aluminum foil is shown in table 3.1, which is much lower than, measured and the prediction technique described in section 4. However this classical approach is for an infinite planar sheet and not an enclosure. Table 3.1 Transmission line analogy theoretical magnetic field S.E. of 0.075mm aluminum foil with the source at different distances from the material Distance (m) H field S.E. (db) With the enclosure and tx loop inside a shielded room the variation in measured S.E. With rx antenna location was much greater. With the enclosure just outside of the 6

7 shielded room and outside of the building, far removed from conductive surfaces, this affect was very slight. We, and other investigators, have seen this apparent contradiction to the theory of reciprocity in other measurements and a number of explanations have been proposed for the dependence in source location, which we do not see. The explanations are based on the real wave impedance of the incident field. Reference 1 explains to some extent this effect. We also found that the measurements of S.E outside of the lab were different above 10kHz from the measurements inside the lab and this is shown in figure 3.4. Aluminum foil enclosure 1m x 1m x 0.7m magnetic field S.E. External loop at 0.5m, internal loop at 0.5m Performed inside lab Performed outside of lab S.E. (db) Fre que ncy (k Hz) Figure 3.4 Comparison between S.E. measured within the lab to outside the lab. To understand the field incident on the enclosure and the eddy currents induced into the enclosure a thorough investigation of the currents on the exterior of the enclosure was carried out. The 560 turn surface current probe was used to measure the pick-up on all sides of the enclosure and also to measure the H field pick-up in air at the same relative positions without the enclosure in place. The probe was oriented in both vertical and horizontal orientations to obtain the highest pickup. The highest measurement observed with the probe in air was then given the nominal value of 1 and all other measurements were compared to 1. A sketch of the relative pick-up in air at 5kHz is shown in figure 3.5 and a sketch of the pick-up on the enclosure at 5kHz is shown in figure

8 Top ar Re t on Fr Far side m 0.5m Near (door) side Bottom m 1m Figure 3.5 Relative pick-up in air from the surface current probe at 5kHz. Top ar Re t on Fr Far side m m Near (door) side Bottom m 1m Figure 3.6 Relative pick-up on the enclosure from the surface current probe at 5kHz. 8

9 Examining the pick-up in air, one might question the validity of the measurements, since some don t seem to reduce as 1/r3. However, we must remember that as we move away from the plane of the loop, we must not only take into account Hθ, but also Hr. By calculating the composite field at these locations, we find the measurements extremely accurate on all sides except for the top. This calculation is provided in section 4. In accordance with classic field theory the total magnetic field on the outside of the enclosure facing the transmitting antenna is twice that for the incident field The measurements on the wall facing the transmitting loop show a field higher than the incident, but not exactly twice the magnitude. The difference between this increase and a doubling in the field is 19dB, exactly the level of S.E. at 5kHz. The same measurement was made on a very large conducting wall versus the field in air with similar results. The fields on all other surfaces of the enclosure are incredibly close to the field without the enclosure as shown in figures 3.2 and ) Magnetic fields generated by the transmitting loop and induced into the enclosure and the fields generated by the enclosure currents The H fields generated by an electrically small loop antenna are shown in figure 4.1 H = 2, Hθ = 0, Hr = 2 H = 1.978, Hθ = 0.17, Hr = 1.97 H = 1.91, Hθ = 0.34, Hr = H = 1.8, Hθ = 0.5, Hr = H = 1.66, Hθ = 0.64, Hr = Z H = 1.49, Hθ = 0.76, Hr = H = 1.32, Hθ = 0.86, Hr = θ m Y H = 1.163, Hθ = 0.94, Hr = 0.68 H = 1.04, Hθ = 0.985, Hr = H = 1, Hθ = 1, Hr = 0 Figure 4.1 H fields generated by a loop antenna. 9

10 At low frequency and close to the loop these fields are referred to as quasi static and are in the near field. The magnitude of Hr, when theta is 0, is twice the magnitude of H θ, when theta is 90. Figure 4.1 shows the relative magnitude of these fields with the angle theta varied. Both fields reduce with distance from the loop, in the near field, as a function of 1/r3. As the circumference of the loop is much smaller than a wavelength, the antenna is electrically small, the current around the loop is virtually constant and the same magnitude of Hθ and Hr generated around the circumference of the loop is constant, as long as the distance and z and y locations are kept constant One mechanism for magnetic field shielding is absorption, and this is effective when the metal thickness is greater than the skin depth. The skin depth for aluminum is shown in table 4.1. Table 4.1 skin depth of aluminum. ƒ(hz) Skin depth (mm) It can be seen from table 4.1 that even at 100kHz the skin depth is much greater than 0.076mm, the thickness of the aluminum foil used in the enclosure under test. This means that virtually the same current flows on the internal wall of the enclosure as on the outside and intuitively it would seem that no magnetic field S.E for the foil enclosure is achieved, certainly not by absorption. The mechanism is that the incident field sets up an eddy current in the aluminum which generates in turn a magnetic field. The field lines cutting the enclosure are shown in figure 4.3. The field lines inside the enclosure, generated by the eddy current is in the opposite phase to the incident field as shown in figure 4.3 and so bucks the incident field. 10

11 Multiturn RS01 Transmitting Loop 0.5m Direction of Current Flow Direction of H-Field Figure 4.2 Magnetic field lines from a loop tx antenna cutting an enclosure. 11

12 Multiturn RS01 Transmitting Loop 0.5m Direction of Current Flow Direction of imposed H-Field Direction of current-induced H-field Figure 4.3 Magnetic field lines incident on an enclosure and the magnetic field set up by the current flow on the enclosure. As the current flow on both sides of the enclosure is in the same phase, the magnetic field from this current on the two sides repel and so the field does not enter the enclosure, as shown in figure 4.4. This was verified by rotating the receiving antenna 90 as shown in figure 4.5. In this case the field that is generated inside the enclosure (by the current on the face, top and on the bottom of the enclosure) does not cut the loop at right angles but is in the plane of the loop and no voltage should be induced. As no field due to the current flow on the sides penetrates the enclosure the pick up on the rx loop antenna should be negligible and this was confirmed by measurement. 12

13 Top Side Internal Field Side I I Rx Front Tx Figure 4.4 Field generated by current flow down the sides of the enclosure does not enter. Internal field due to flow on face top back and bottom. 13

14 Top Side I Internal Field Side Rx I Front Tx Figure 4.5 Internal field cuts the Rx antenna at 0. Field due to current on the wall does not penetrate the enclosure. As the field generated on the inside of the enclosure opposes the incident field the two will cancel and theoretically the S.E. will be infinitely high. This explains why the incident field and the field inside the enclosure should not reduce at different rates based on distance from the source. However as discussed in section 3, a small difference does exist in the measurements. The current flows in: the inductance of the loop, the resistance of the foil, and where a seam exists, across the contact impedance of the seam and the inductance of the seam. As the skin depth of aluminum is much greater than the thickness of the foil then the current flows through the thickness of the foil and no correction for skin depth is needed. The equation for the total dc resistance of the foil enclosure in the current path generating the internal field is: RS = 2 L2 + 2 D GL1t

15 Where L2 = height of enclosure D = depth of enclosure L1 = width of enclosure G is conductivity in mhos/m t = thickness of material Equation 6.28 for the shielding effectiveness of the enclosure, from reference 2, can be simplified to 2π fµ 0 A L1 RS + LN R + Rc f + 2π fl j ( ) 4.2 Where A = area of enclosure = L2 x D L = length of joint (joint or slot in the current path) R = DC resistance of joint RS = resistance of an enclosure wall (normally as a function of skin depth) RC = contact resistance L = equivalent series inductance of slot N = number of joints in the current path The shielding effectiveness is the ratio between the inductive reactance of the enclosure given by the upper term in equation 4.2, and the combined resistances and impedances. As the inductive reactance increases as a direct function of frequency and the inductive reactance of the seam is typically low, unless large gaps appear in the seam, the shielding effectiveness increases with frequency, as seen in all measurements. In a perfect seam free enclosure the S.E will increase directly proportional to frequency. The calculated resistance of the aluminum foil enclosure, using the conductivity of aluminum as 3.54e+7mhos/m, is 1.2mΩ. At low frequency the impedance of the seam is lower than the dc resistance of the foil. Only above 10kHz do we see that the shielding effectiveness no longer increases monotonically. By using the dc resistance of 1.2mΩ, and a contact impedance of 2.2µΩ, the predicted S.E. from equation 4.2 is remarkably close to the measurements made on the free space range, as shown in figure 4.6. By setting the contact resistance to a very low value we obtain the theoretical maximum S.E. achievable with a seam free enclosure, as shown in figure

16 Aluminum foil enclosure 1m x 1m x 0.7m magnetic field S.E. External loop at 0.5m Magnetic field S.E. with internal loop 0.5m from wall Measurements made outside Theoretical S.E. with Rc = 2.2uOhm Theoretical for perfect seam free enclosure Figure 4.6 Measured magnetic field versus theoretical for the 1m x 1m x 0.7m enclosure At these frequencies where the skin depth is greater than the thickness of the material the type of seam is not so important. For example, a seam covered with a strip of metal in a 1m x 1m x 1m enclosure described in a companion report entitled Shielding a room using aluminum foil has similar magnetic field S.E when compared to the seam clamped between brass plates described here. However the type of seam does greatly change the plane wave S.E. as described in the same report. 16

17 References 1) Shielding Effectiveness and Wave Impedance. Torsten Sjoegren, Telub Teknik AB, Vaxjo, Sweden. 2) Electromagnetic Compatibility: Principles and Applications. D. A. Weston. Published by Marcel Dekker,