GB2310933A

GB2310933A - A method of measuring electromagnetic emissions

Info

Publication number: GB2310933A
Application number: GB9604730A
Authority: GB
Inventors: Peter John Miller
Original assignee: Motorola Ltd
Current assignee: Motorola Solutions UK Ltd
Priority date: 1996-03-06
Filing date: 1996-03-06
Publication date: 1997-09-10
Anticipated expiration: 2016-03-06
Also published as: GB9604730D0; GB2310933B

Description

A METHOD OF MEASURING F"ITIECTROMAGNETIC EMISSIONS Field of the Invention This invention relates to a method for measuring electromagnetic emissions from an electronic device.

Background of the Invention Many international and national standards now require that electronic equipment shall not exceed certain electromagnetic emission maxima.

Therefore electronic equipment is required to be tested for such electromagnetic emissions, to determine compliance with the standards.

Test equipment is used for this purpose. Either the measurements must be conducted in a fully screened environment (such as an anechoic chamber) or they must be conducted "outside" when background electromagnetic radiation will be present. In the first case the anechoic chamber itself will distort the measurements, and any test equipment etc. within the chamber will still emit causing an artificial background radiation. To avoid this either the background radiation effects are ignored (and thus the readings are higher than necessary) or a first reading is taken with the equipment switched off or inactive, and a second reading is taken with the device switched on or active.

For the sake of calculation simplicity, these two readings are typically logarithmic, and are converted to linear values for comparison. The subtraction of the first reading from the second reading removes substantially all background electromagnetic readings. This result is then converted back into a logarithmic scale for interpretation with respect to the standards.

A problem with this arrangement is that if the first and the second signals are almost identical there will be a large negative dB figure, which hinders interpretation of the results. This large negative dB figure is in fact erroneous since in practice any real measurements taken in a place with lower ambient readings will be limited by the equipment's finite noise floor and also the fact that the original readings before they were subtracted have a finite accuracy and resolution(and so two identical readings do not necessarily mean that the signals being measured are exactly identical).

This invention seeks to provide a method for measuring electromagnetic emissions which mitigates the above mentioned disadvantages by providing a very close approximation to the reading that would be obtained using the same measuring equipment if it was used in a place completely free from background electromagnetic radiation.

Summarv of the Invention According to the present invention there is provided a method of measuring electromagnetic emissions from a device to be tested by test equipment, the method comprising the steps of: measuring the electromagnetic spectral noise of the test equipment; measuring a first electromagnetic spectrum with the device in a first state; measuring a second electromagnetic spectrum with the device in a second state; calculating an error spectrum by subtracting the second electromagnetic spectrum from the first electromagnetic spectrum; and, adding the spectral noise to the error spectrum to produce a spectral result, wherein the addition of the spectral noise substantially simulates an electromagnetically free space environment, with the test equipment present.

Preferably the first and second electromagnetic spectra are measured in logarithmic form, and the method further comprises the steps of converting the first and second electromagnetic spectra into linear form before calculating a linear error spectrum and converting the linear error spectrum into logarithmic form.

The first state is preferably an active state, and the second state is an inactive state of the device to be tested. Alternatively, the first state is a first configuration, and the second state is a second configuration of the device to be tested.

In this way a result is achieved which is a very close approximation to the reading that would be obtained using the same measuring equipment if it was used in a place completely free from background electromagnetic radiation.

Brief Description of the Drawing An exemplary embodiment of the invention will now be described with reference to the drawings, in which: FIG. 1 shows a flow chart of a method for measuring electromagnetic emissions in accordance with the invention, and FIG.2 shows a series of graphs of measurements from test equipment, processed in various ways.

Detailed Description of a Preferred Embodiment Referring to FIG. 1, there is shown a flow chart 5 of a test sequence for measuring electromagnetic emissions from an electronic device, such as a piece of computer equipment.

The test sequence commences at block 10. A first measurement A is taken at block 20, of the spectral noise of the test equipment. This is done by removing the antenna of the equipment and replacing it with a resistance equal the antennas impedance (normally 50 Ohms), thereby substantially attenuating noise external to the equipment.

At block 30, a measurement B is taken of an electromagnetic spectrum with the electronic device turned off. This effectively measures the background noise. Any test equipment etc. that should not be measured, but is required for the completion of the measurements in block 40 should be ON at this time.

At block 40, a measurement C is taken of an electromagnetic spectrum with the electronic device turned on. This effectively measures the background noise combined with the electromagnetic emissions of the device.

At block 50, the measurements B and C and the spectral noise measurement A are converted from logarithmic scales to linear scales, to facilitate easy manipulation thereof.

At block 60, the linear equivalents of the measurements A, B and C are processed as follows. Measurement B is subtracted from measurement C.

This removes the background noise content from the electromagnetic emissions of the device. Then measurement A is added. This injects the low level noise associated with the test equipment back into the overall result (D).

At block 70, the result D is converted back into the logarithmic scale, for ease of interpretation. The test ends at block 80.

The invention is further illustrated with reference to the following example which gives actual measurements taken within a semi-anechoic chamber using an artificial background electromagnetic source which can thus be switched off to enable the "real answer" to be obtained. The figures below show the calculations for one frequency rather than the many frequencies normally present in a complete spectra, to save space and improve clarity.

Also, again for clarity, all results are directly shown in linear units rather than the more normal dB.

Results 1). Signal+interference+noise 102.09uV A Interference+noise 52.12uV B Just equipment noise 27.38uV C Real measures reading (signal+noise) 76.82uV D Calculations - subtraction (A-B) 45.72uVE with added noise (A-B+C) 73.lug F 2). Signal+interference+noise 127.06uV A Interference+noise 52.12uV B Just equipment noise 27.38uV C Real measures reading (signal+noise) 105.93uV D Calculations - subtraction (A-B) 74.94uVE with added noise (A-B+C) 102.32uV F In case 1), ignoring the background interference gave a reading 33% high, just subtracting gave a reading 40% low, while the new approach gave a 5% error. In case 2), ignoring the background interference gave a reading 20% high, just subtracting gave a reading 29% low, while the new approach gave a 3% error.

Note that in a "normal" case, measurement D would not be possible (this is the reading obtained if the interface were turned off).

It can be clearly seen that the new approach gives results very close to those that would be obtained if the measurements could be made directly. It can also be seen that this advantage improves the closer the measurements are to the test equipment's noise floor or resolution limits.

Referring now also to FIG.2, there are shown three graphs, 100, 200 and 300, of results from an example piece of test equipment. Graph 100 shows uncorrected raw data. Graph 200 shows a basic subtraction result. In graph 200 no values are shown below about 70Mhz (the over curves go to 30MHz) because large negative values were produced that were well off the curve.

Graph 300 shows the result with the added test equipment noise. It is to be noted that this graph substantially simplifies the result, without departing from the "real" values.

It will be appreciated by a person skilled in the art that alternate embodiments to the one described above are possible. For example, instead of the "linear" addition/subtraction described above which would normally be used for peak and quasi-peak measurements, an "rms." +/- could be used (this would normally be correct for average measurements). In this case the sum is changed from D=(B-C)+A to D=nisquare root(B2-C2+A2).

Other statistical means to add/subtract could also be used.

Furthermore, the measurements taken need not merely reflect an 'on' and 'off state of the electronic device. For example, a device having two or more power settings or configurations could be measured at each setting or configuration, thereby determining the change in electromagnetic noise output between the two settings (configurations).

Claims

1. A method of measuring electromagnetic emissions from a device to be tested by test equipment, the method comprising the steps of: measuring the electromagnetic spectral noise of the test equipment; measuring a first electromagnetic spectrum with the device in a first state; measuring a second electromagnetic spectrum with the device in a second state; calculating an error spectrum by subtracting the second electromagnetic spectrum from the first electromagnetic spectrum; and, adding the spectral noise to the error spectrum to produce a spectral result, wherein the addition of the spectral noise substantially simulates a electromagnetically free space environment, with the test equipment present.

2. The method of claim 1 wherein the first and second electromagnetic spectra are measured in logarithmic form, and the method further comprises the steps of converting the first and second electromagnetic spectra into linear form before calculating a linear error spectrum and converting the linear error spectrum into logarithmic form.

3. The method of claim 1 or claim 2 wherein the first state is an active state, and the second state is an inactive state of the device to be tested.

4. The method of claim 1 or claim 2 wherein the first state is a first configuration, and the second state is a second configuration of the device to be tested.

5. A method of measuring electromagnetic emissions substantially as hereinbefore described and with reference to the drawings.