The Electromagnetic Interference (EMI) Project – Part 1 Introduction

Part 1: Introduction by Pat Brown

This article addresses this question, “What are the effects of electromagnetic interference EMI on various types of audio wire?”

Hum and buzz pickup is a common problem for audio systems. SynAudCon Hum and Buzz workshops are always a sell-out, and the subject comes up often in our email discussion group.

A recent thread has spawned a group project to determine the effects of electromagnetic interference (EMI) on various types of audio wire. The idea is to generate a strong magnetic field under controlled conditions and test the susceptibility of various wire types. Ray Rayburn provided the initial design for the magnetic loop that would be used to generate the interfering magnetic field (the source loop) as well as many other insights on the procedure. Figure 1 shows the 10ft square loop which consists of a solid 12AWG conductor stapled to a wooden frame. The idea is to simulate an electrical circuit that has line and neutral separated due to miswiring. Normally line and neutral are within the same conduit or jacket, and their close proximity cancels much of the radiated magnetic field caused by current flowing through the wire. When separated, the field doesn’t cancel and it can be picked up by sensitive circuits or cables.

Electromagnetic Interference EMI Test setup

The loop is driven with an amplifier configured as a constant current source – nine amperes of current flow at each test frequency. Thirty log-spaced frequencies between 40Hz and 10kHz were used. The frequency sweep and subsequent analysis are handled by an Audio Precision ATS-2 analyzer. The noise floor of this analyzer is -116dBV, which is lower than the noise floor of just about any piece of audio gear. Various wire types will be tested, including wires with good and bad field immunity.

In a weak moment I volunteered to build the test rig and collect the data. Dale Shirk has agreed to write-up the test results in a series of articles. This initial installment is just to familiarize you with the project and objectives.

EMI field as picked up by a zip cordPhoto of the original wooden loop

The test rig and procedure were evaluated using zip cord as the pickup loop – a worst case scenario since it is not twisted. A ten foot length with far end shorted was connected to the analyzer input, and placed parallel to the source loop with the zip cord nearly touching the loop conductor. The test was repeated at 2:1 distance ratios (beginning at 0.75in) to a final distance of 16ft. Figure 2 shows the signal as picked up by the zip cord. Figure 3 shows the wooden loop. One surprise regarding the results of this test was that everyone expected an inverse fall-off of the field signal (-3dB per distance doubling) up close to the loop, and inverse-square fall-off (-6dB per doubling) at the greater distances. Dr. Patronis points out that the field from a magnetic dipole source (which this is) will attenuate at the rate of -18dB per distance doubling at large distances measured to the center of the dipole. We did not get far enough away to realize this. Ray Rayburn and Bill Whitlock explain:

At all distances we have the inverse square law. At near distances the far side of the ZIP cord was exposed to significantly lower magnetic field than the near side, and since the output of the “pickup loop” was dependent on the difference in magnetic fields (the field cutting the far side subtracts from the output level) you got a lot more output relative to the average magnetic field in the center of the ZIP cord. At near distances the relative distance to the far side of the source loop is at such a huge ratio compared to the near side that the reversed magnetic field from the far side of the source loop has close to zero effect.

At moderate distances the field strength at each side of the ZIP cord receive loop is much closer to equal, and the ratio if the relative distances from the two sides of the source loop is going way down.

At really long distances (beyond what you were doing in your test setup) you have three effects knocking down the level:

  • 1) Inverse-square law
  • 2) the drop in sensitivity of the receive loop since the magnetic levels are for all purposes equal at both wires
  • 3) the drop in magnetic output of the source loop (at least in the plane of the loop) because the reverse polarity field of the far side of the source loop is canceling out the field of the near side

Note that effects 2 and 3 are the same effect, only the relative spacing between the wires is very small in 2 and much larger in 3.

While this setup worked well for testing the susceptibility of zip cord, it did not fare as well for other cable types. Twisted pair cables reduced the magnetic field pickup to near the noise floor of the analyzer. “End effects” were also encountered due to the finite length (10ft) of the exposed side of the source loop, making the received signal very sensitive to the exact lateral relationship of the source and receive loops. And finally, at greater distances the field strength fell quickly as the victim wire became more equidistant to the near and far side of the loop. While there are good lessons in all of these behaviors, we have decided to modify the test setup for the remainder of the project. True to the scientific method, what we learned from this initial investigation was how not to do it.

The new test configurations

As we proceed, two scenarios will be evaluated. Figure 4 shows the layout:

  • 1. The susceptibility of each wire to a point source radiator (a Han-D-Mag tape demagnetizer).
  • 2. The susceptibility of each wire type to a straight extension cord carrying 9 amps of current. This is to emulate field pickup that occurs when audio snakes and other cables are run along with power cables. pb