TYPE 6220-1A AUDIO ISOLATION TRANSFORMER for conducted audio frequency susceptibility testing Another secondary winding is connected to a ADDITIONAL MODELS pair of binding posts suitable for connecting to Type 6220-2 100 amperes high current a.c. voltmeter as directed by the applicable transformer. EMI specifications. This winding serves to isolate the voltmeter from power ground. Neither the Type 6220-4 50 amperes 4 KV transient primary nor the secondary windings are voltage HV transformer. connected to the end bells of the core. The Type 9707-1 10 amperes low current transformer may be used as a 4-ohm primary and transformer. 1-ohm secondary or 2.4-ohm primary and 0.6-ohm secondary or 2-ohm primary and 0.5-ohm secondary. SPECIFICATIONS Primary: Less than 5 ohms. FEATURES Secondary: One-fourth the primary impedance. � Provides a convenient bench model unit APPLICATION Frequency Response: 30 Hz to 250 KHz. with three-way binding posts on primary The Type 6220-1A Audio Isolation Transformer and output voltmeter leads. Standard 0.75" Audio Power: 200 watts. was especially designed for screen room use spacing of binding posts allows use of in making conducted audio frequency suscepti- Dielectric Test: 600 volts d.c. primary to standard plugs. High current secondary uses bility tests as required by MIL-STD-461/462 and secondaries and each winding to end bells. 1 /4-20 threaded studs. other EMI speciŢcations. Secondary Saturation: 50 amperes a.c. or d.c. � Capable of handling the audio power required maximum. The transformer may also be used as a pickup by EMI speciŢcations and up to 50 amperes of Turns Ratio: Two-to-one step down. device to measure low frequency EMI currents at a.c. or d.c. through the secondary in series with lower levels than conventional current probes. the test sample. Secondary Inductance: Approximately 1.0 mH (unloaded). � May be used as a pickup device or an isolating In addition, its secondary may be used as an inductor in other tests. Weight: 18 pounds. isolating inductor in the power line during � Suitable for fastening to the bench top in Size: 4.5" wide, 5.25 " high, 6.25" deep plus transient susceptibility tests. (See Application permanent test setups. terminals. (114 mm x 133 mm x 159 mm.) Note AN622001.) DESCRIPTION The transformer is capable of handing up to 200 watts of audio power into its primary over the fre- quency range 30 Hz to 250 KHz. The turns ratio provides a two-to-one step down to the special secondary winding. The secondary will handle up to fifty amperes of a.c. or d.c. without saturating the transformer. 57 TYPE 6220-1A AUDIO ISOLATION TRANSFORMER SOLAR 8850-1 or 6550-1 POWER SWEEP GENERATOR AUDIO SUSCEPTIBILITY TEST SETUP FOR D.C. LINES SOLAR 8850-1 or 6550-1 POWER SWEEP GENERATOR TEST SETUP FOR MEASURING LOW FREQUENCY, LOW AMPLITUDE EMI CURRENT See Application Note 622001 58 APPLICATION NOTE AN622001 USING TYPE 6220-1A TRANSFORMER FOR THE MEASUREMENT OF LOW FREQUENCY EMI CURRENTS At any rate, this impedance suddenly began radiation, particularly at the lower frequencies, appearing in specifications which demanded since the coupling between power leads at low its use in each ungrounded power line for frequencies is inductive, not capacitive. determining the conducted EMI (then known as As it turned out, Stoddart was successful in RFI) voltage generated by any kind of a gadget. developing a current probe based on Alan The resulting test data, it was argued, allowed the Watton’s suggestions regarding the torodial government to directly compare measured transformer approach which is still the primary RFI/EMI voltages from different test samples and basis used today. However, the development of different test laboratories. No one was concerned the voltage measurement probe suffered for lack about the fact that Ţltering devised for suppress- of sensitivity. Watton’s hope had been to provide ing the test sample was based on this artificial a high impedance voltage probe with better impedance in order to pass the requirements, but sensitivity than was then available for measure- that the same filter might have no relation to ment receivers designed for rod antennas and reality when used with the test sample in its INTRODUCTION 50 ohm inputs. Since this effort failed and normal power line connection. “There is more than one way to skin a cat” your Watton’s funds (and probably his interest in the great grandfather and my father used to say. The Not until 1947, that is. At that time, this same subject) faded out of the picture, the program evolution of methods of measuring conducted Alan Watton, a propulsion engineer having no came to a halt. interference illustrates this homely expression in a connection with the RFI/EMI business, decided This meant that the RFI/EMI engineer could either distorted kind of way. to rectify the comedy of errors which had measure EMI voltage across an artificial imped- misapplied his original brainchild. He was in a To start with, a clever and versatile propulsion ance which varied with frequency, or he could position to place a small R and D contract with engineer named Alan Watton at Wright Field early measure EMI current ţowing through a circuit of Stoddart for the development of two probes; in WWII created an artiŢcial line impedance which unknown r.f. impedance. Either way, the whole a current measuring probe and a voltage represented what he had measured on story is not known. In spite of the unknown measuring probe. Obviously, he felt that one the d.c. buss in a twin-engined aircraft. Probably a impedance, the military specifications began needed to know at least two parameters for a true DC-3, but memory is dim on this point. Watton’s picking up the idea of measuring EMI current understanding of conducted interference. The work was sponsored by a committee headed by instead of voltage.The test setup was simpler and current probe is not only a measure of EMI Leonard W. Thomas (then of Buships) with active the current probe was not as limited as the LISN current, it is a measure of the magnetic field participation by Dr. Ralph Showers of University in its ability to cope with large power line radiation from the wire or cable under test. This of Pennsylvania and others. currents. And the current probe measurement is a more meaningful measure of magnetic was also a measurement of magnetic field So the Line Impedance Stabilization Network (LISN) was born. It was a pretty good simulation of that particular aircraft and the electrical systems it included. But then someone arbitrarily decided to use this artificial impedance to represent any power line. 59 AN622001 (continued) radiation. The current probe was somewhat better than the LISN for measurements below 150 KHz and above 25 MHz but, even so, the technique was not very sensitive at the lower frequency end of the spectrum. A young Boeing EMI engineer named Frank Beauchamp was the first to apply the current probe to wideband measurements from 30 Hz to 20 KHz. He was smart enough to realize some of the problems in this range so he incorporated the sliding current probe factor into the method of measurement he spelled out in the Minuteman Specification, GM-07-59-2617A. The test method required that the probe factor existing at 20 KHz should be used for obtaining the wideband answer in terms of “per 20 KHz” bandwidth. This injection tests to prevent the transient from being BASIC CONCEPT meant that the speciŢed limit was not a constant short-circuited by the impedance of the power The application described herein has grown out throughout the 20 KHz bandwidth, but was line. In this application all other windings are left of a suggestion by Sam Shankle of Philco Ford in varying as the inverse of the probe factor. A very open. See Figure 1. Secondly, the transformer Palo Alto. He and his capable crew first tried this sensible solution at the time. Regrettably, later can be used for measuring EMI current as scheme using H-P Wave Analyzers as the associ- speciŢcations did not follow this lead. described herein. See Figure 2. At other times, if ated voltmeter. Our work with the idea has When later EMI speciŢcations extended the need it is not needed in the circuit, short cicuiting the concentrated on conventional EMI meters with for measurement of EMI currents down to 30 Hz primary winding will effectively reduce the 50 ohm inputs. without taking into account the sloping probe secondary inductance to a value so low that the Basically, the test method consists of using the factor, the problem of probe sensitivity became transformer acts as if it isn’t there. secondary (S) of the Solar Type 6220-1A Audio critical. Attempts to compensate for the poor Isolation Transformer as the pickup device. current probe response at low frequencies by ACHIEVING MAXIMUM SENSITIVITY The transformer winding normally used as the using active element suffer from dynamic range FOR CONDUCTED EMI CURRENT primary (P) is used as an output winding in this difŢculties and the possibility of overload. MEASUREMENTS case. The method provides a two-to-one step up The basic circuit in Figure 2 provides the most This led to another way of “skinning the cat,” with to further enhance the sensitivity. pickup and transfer of energy over the frequency the aid of the Audio Isolation Transformer already range 30 Hz to 150 KHz. Curve #1 of Figure 3 available and in use for susceptibility testing. USE OF THE TYPE 6220-1A shows the correction factors required to convert The technique described in the following TRANSFORMER IN GENERAL narrowband signals to dB above one microam- paragraphs indicates how to obtain considerably Since the transformer is connected in series with pere. Since the sign of the factor is negative for greater measurement sensitivity for conducted each ungrounded power input lead (sequentially) most of the range, the sensitivity is considerably narrowband EMI currents and a means for obtain- for performing the audio susceptibility tests, it can better than that of conventional current probes. ing a ţat frequency characteristic without the use be used for two additional purposes while still in The sensitivity achieved by this technique is of active elements for broadband or “wideband” the circuit. First, the secondary winding can act better than .05 microamperes at frequencies EMI current measurements. as the series inductor suggested for transient 60 AN622001 (continued) above 5 KHz when using an EMI meter capable of measuring 1.0 microvolt into 50 ohms. For EMI dB TO BE ADDED TO EMI METER READING (IN dB) TO meters such as the NM-7A and the EMC-10E, OBTAIN dB / �A. the meter sensitivity is a decade better and +20 it is possible to measure EMIcurrents of .005 microamperes at 5 KHz and above. CURVE #4 +10 R�0.5 OHM FLATTENING THE RESPONSE At a sacrifice of sensitivity, the upper portion of CURVE #3 0 R�2 OHMS the frequency vs. correction factor curve can be flattened to provide a constant correction factor from about 1 KHz up to 150 KHz. This is depicted –10 CURVE #2 in curve #2 of Figure 3, where a -20 dB correction R�10 OHMS is suitable over this part of the frequency range. The ţattening is obtained by loading the primary –20 with a suitable value or resistance. The resistance CURVE #1 value used in this example is 10 ohms. The R=∞ –30 flattening still allows the measurement of a 100K 1M 10 100 1K 10K FREQUENCY HERTZ 0.1 microampere signal when using an EMI meter FIGURE 3 – TYPICAL CORRECTION DATA VS. FREQUENCY with 0.1 microvolt sensitivity. An advantage of this response curve is the sloping correction at frequencies below 1KHz which acts like a high LIMITATIONS OF THE METHOD THINGS TO BE WARY OF pass filter to remove some of the power line When measuring EMI current on d.c. lines, there The 10 �F feed-thru required by present harmonics from wideband measurements. are no problems, but on a.c. lines there are limita- day specs has appreciable reactance at 30 Hz If you are only interested in frequencies above tions. The a.c. voltage drop across the winding (S) (�54 ohms) and acts to reduce the actual EMI 150 Hz, a 2 ohm resistor is all that is needed. See due to power current flowing to the test sample current flowing in the circuit. This means less curve #3. is the principal problem. This voltage induces trouble in meeting the spec, but when calibrating twice as much voltage in the output winding (P) the test method described herein, it is wise to STILL MORE FLATTENING at the power frequency. Since we prefer to limit short circuit the capacitor. Like the girdle ads say, you can Ţrmer and ţatter, the power dissipation in the 50 ohm input to the In the case where the input circuit to the with a loss in sensitivity , by further reducing the EMI meter so that it will not exceed 0.5 watts, the EMI meter is reactive, such as the EMC-10E, it is value of the shunt resistor. This is illustrated induced voltage must be kept below a safe limit. necessary to use a minimum loss ‘T’ pad at in curve #4 of Figure 3 where a 0.5 ohm shunt For 400 Hz lines, the power frequency current the input to the meter. The Eaton NM-7A and resistor (Solar Type 6920-0.5) is connected must not exceed 16 amperes to avoid too much NM-12/27A units do not require this pad and across the transformer primary winding used as 400 Hz power dissipation in the input to the EMI its loss. an output winding to the EMI meter. The overall meter. Also, the resistance ‘R’ used across the ţatness is achieved at the sacriŢce of considerable output winding (P) must be at least a 50 watt DETERMINING THE NARROWBAND sensitivity, but the sensitivity is well under the rating on 400 Hz lines. This resistor should be CORRECTION FACTOR requirements of existing specifications and the noninductive to avoid errors due to inductive The test setup of Figure 4 describes the simple correction network utilizes no active elements. reactance. method of determining either the transfer 61 AN622001 (continued) impedance or the correction curve, whichever is sary to convert the meter reading to dB above desired. Actually, there is no need to plot the one microampere for narrowband measurements. answer as transfer impedance, since the desired In most cases, the correction will have a negative end product is the correction factor to be applied sign. For example, at 100 Hz the EMI meter to the meter reading to obtain decibels above may read 88 dB above one microvolt. Since the one microampere. The correction must be reference is 80 dB above one microampere, the obtained for each configuration. In other words, correction is -8 dB to added algebraically to if you want to use the method for maximum the meter reading to obtain the correct reading sensitivity, the calibration is performed with just in dB above one microampere. a 50 ohm load on the primary winding simulating If the 10 dB pad has been used, this loss must be the EMI meter. If the flattening networks will be accounted for in deriving the correction. If the used, then they must be connected to the primary pad will be used in the actual test setups, its loss winding during the calibration and must be becomes part of the correction factor. In this case, further loaded with 50 ohms to simulate the EMI the meter reading obtained in the foregoing FIGURE 4 – TEST SETUP meter input. example would be 78 dB above one microvolt FOR DETERMINING CORRECTION FACTOR At each test frequency, the output of the audio and the correction factor would be +2 dB for signal generator is adjusted for a level which narrowband measurements. delivers the same current to the secondary (S) of Repeating this procedure at a number of test MIL-STD-461A, the range covered will depend the transformer. This is accomplished by setting a frequencies will produce enough data to plot upon the cutoff frequency of the filter. For constant voltage across the 10 ohm resistor. A a smooth curve for use when actual tests are example, on 60 Hz power lines and using Solar convenient level is 0.1 volt across 10 ohms which being conducted. Type 7205-0.35 High Pass Filter between is 10,000 microamperes (80 dB/uA). the 6220-1A Transformer and the EMI meter, Adjust the gain of the EMI meter to assure a one DERIVING THE BROADBAND obtain the average narrowband correction microvolt meter reading for a one microvolt R.F. CORRECTION FACTOR between 350 Hz and 14 KHz and add the band- input from a standard signal generator. Then When making broadband measurements as width correction factor of 37 dB. On 400 Hz lines connect the 50 ohm input circuit of the EMI meter required by MIL-STD-461A in terms of “dB above when using the Solar Type 7205-2.4 High Pass to the primary of the 6220-1A. If the EMC-10E is one microampere per megahertz,” use the Filter between the transformer and the EMI used, insert a 10 dB pad in series with the input. If average of the narrowband factors over the range meter, determine the average of the narrowband the calibration is for maximum sensitivity, no 30 Hz to 14 KHz and add a bandwidth correction factors in the range of 2.4 KHz and 14 KHz and additional loading is necessary. If the calibration factor of 37 dB. add the bandwidth correction factor of 38.5 dB. is for the ţattened versions discussed above, the In the case of Method CE01 of MIL-STD-461A, use appropriate resistance must be connected across the 20 KHz wideband mode of the EMI meter, SUMMARY the primary of the transformer. determine the average of the narrowband factors Some of the material given in this Application At the frequency of the test, set the output of the over the range 30 Hz to 20 KHz and use this Ţgure Note is terse and given without much explana- signal source to obtain 1.0 volt across the 10 ohm as the bandwidth correction factor. tion. If your are confused by this simplification, resistor. Carefully tune the EMI meter to the test just call us. Incidentally, the Signal Corps liked this When using high pass filters at the input to the frequency and note the meter reading on the dB method so well that they included it in Notice #3 EMI meter to eliminate the first few harmonics scale. The difference between the meter reading to MIL-STD-462 date 9 Feb 71. of the power line frequency as allowed by in dB and 80 dB represents the correction neces- 62