EMI Testing in Medical Electronics

Medical Device & Diagnostic Industry Magazine
MDDI Article Index

An MD&DI August 1997 Column

Although electromagnetic compatibility compliance testing is a necessary time and cost expenditure, manufacturers can save themselves much of the hassle of this final stage by doing in-house precompliance testing and troubleshooting.

EMC compliance testing requires a major investment, generally more than most companies can justify for in-house testing facilities. But less rigorous precompliance testing and troubleshooting can be done with a much smaller investment. These tests, performed in the manufacturer's facilities, allow for much more effective use of engineering time, speeding up the product development cycle. Meticulous work performed in these manufacturer-controlled sessions can enable the company to work out kinks in the device processing or function. Finding flaws during these early stages can save the manufacturer substantial amounts of time and money during the compliance stage, when errors are more costly to fix and testing-facility costs are at a premium.

Radiated emissions are usually tested in an open area test site (OATS), such as this one owned by TÜV Product Service in Pinewood Springs, CO. A log periodic antenna inside is at approximately 3.7 m.

However, without an understanding of compliance testing and how it operates, it may be difficult for a manufacturer to take advantage of the benefits of the precompliance and troubleshooting stages. A brief discussion of compliance testing may help put the other two steps into perspective.

COMPLIANCE TESTING

Compliance testing is what most engineers equate with electromagnetic interference (EMI) testing. These tests gather the necessary data to verify compliance with the appropriate EMI requirements, usually the International Electrotechnical Commission's IEC 601-1-2 for medical devices. The testing emphasis is on accuracy and repeatability, both of which require a carefully controlled test setup. Comprehensive tests are performed for each device.

Establishing a complete compliance test facility could easily run up several million dollars in facility, capital equipment, staff, and overhead costs. Few companies can justify the cost of running these tests themselves. Most go to an independent test facility, generally one that has connections to a competent body for the European Union (EU) requirements. A complete facility includes a testing range for radiated emissions, a shielded facility for radiated immunity, plus a plethora of transient test instruments. These houses are well versed in the various testing requirements, but they are expensive, typically renting for $1400 per day. Add in the cost of preparing for and supporting the test, and costs may run up to $5000 per day.

Radiated emissions are generally tested in an open area test site (OATS), and the emissions must be separated from the ambient provided by nearby radio transmissions. Anechoic shield rooms large enough for 10-m emission tests are also expensive but increasingly available. One or more receiving antennae are mounted on a mast at various positions, and measurements are taken using a wideband receiver or a spectrum analyzer. The equipment being tested is mounted on a turntable so that all azimuths can be measured. Conducted emissions are easier to handle using a receiving device and an LISN (line impedance stabilization network) to isolate the equipment from the power source.

Radiated immunity testing subjects the equipment to a radiated field and must be done in a shielded enclosure, usually a shield room. The required field uniformity is not trivial, generally requiring anechoic tiles to minimize standing wave patterns. A signal generator creates modulated waves in the appropriate frequency range to feed an amplifier, and antennae provide the necessary field. A means of measuring the field strength is needed as well.

The electrostatic discharge (ESD) test employs an ESD gun capable of generating a calibrated transient pulse. The electrical fast transient (EFT) test injects a calibrated high-frequency pulse train into the power lines and external data cables, simulating the showering arc. The surge test injects a calibrated low-frequency pulse, simulating a lightning-induced surge.

Other tests are under consideration, including the low-frequency magnetic field tests, a power dropout test, and a conducted radio frequency (RF) test. There is no apparent end to the number of new tests, and these will place a cost burden on the manufacturer, whether the testing is performed in-house or at a test facility.

PRECOMPLIANCE TESTING

Given the high cost of compliance testing, consider running in-house precompliance testing and troubleshooting. Precompliance testing identifies and fixes trouble spots in the design phase, when procedures are more flexible and changes are easier and cheaper to make. The emphasis is on quick results, sacrificing accuracy for expediency. Precompliance testing is generally done in two phases. The first is done early in the design stage, when there are a number of unanswered questions. What are the vulnerable design areas, and how can they be protected? It is not too hard to identify the risk areas, but it is impossible to accurately quantify them. Representative samples of the vulnerable elements should be subjected to a test early in the design stage. For example, the input amplifier of a patient-connected device is notoriously vulnerable to radio-frequency interference (RFI). By building and testing this portion of the design before the rest of the system is working, problems can be identified and fixed in a timely fashion.

Generally, accuracy is sacrificed for expediency during prequalification. Concentrate on the identifiable trouble spots rather than doing a full scan. Knowing that the test is sacrificing accuracy, factor in a safety margin of approximately 6 dB, depending on the test. This generally works but risks overdesign: a lot of hard work may go into getting a little extra margin that really isn't necessary.

To minimize unnecessary work, examine the test setup. Moving the equipment cables even a little can cause significant changes in the results. Try several setups (cable, equipment, or antenna placements, etc.) to get a feel for the test's validity. If there are concerns about test accuracy, run the procedure at the test house for a more accurate reading. In such a case, bring along a fix, unpalatable as it might be: test houses seem to find the weak spots in equipment and often force manufacturers into meeting more stringent values.

Generally, the radiated emissions and immunity tests are the most difficult to do accurately and, thus, need the largest test margin. Conducted tests are more predictable, so margin needs are lower. ESD testing has a reputation for inconsistency, but the contact test is much less of a problem. Hold ESD testing on successive days to establish a confidence level. Transient testing, especially ESD testing, tends to be destructive, so keep a liberal supply of spare parts available. Since transient tests tend to create latent defects (the walking wounded), never test a unit intended for sale.

What are the minimum requirements for in-house testing capability? For emissions, start with a spectrum analyzer. Although test houses are turning to receivers as the instrument of choice, spectrum analyzers are more flexible diagnostic tools for engineering tests. Buy one or two antennae of the appropriate frequency range. Consider a biconical antenna for the lower frequencies (30 to 200 or 300 MHz) and a log periodic for the higher frequencies (up to 1000 MHz); other choices are available. An antenna mast is optional for raising the antenna to higher elevations. A turntable is necessary to rotate the equipment being tested. Inexpensive turntables can be made from large cable reels, like a lazy Susan. The equipment can be set up in a parking lot, weather permitting, or even in a large conference room away from other emitting equipment. If the ambient is too high, move the antenna closer and extrapolate, using a reciprocal falloff. If the antenna is moved to a 1-m distance, subtract 20 dB to arrive at a 10-m reading. This extrapolation will increase the error, so allow a 6-dB margin.

For radiated immunity, consider a TEM cell or G-TEM! cell (EMCO, Austin, TX), which will accommodate small boxes. As mentioned, a shield room is expensive, especially one with anechoic tiles. A calibrated ESD gun is essential for testing ESD. Homemade ESD generators are not suitable, nor are the inexpensive hand-held Tesla coils. Similarly, the various power-quality tests can only be adequately tested by a commercial tester. Fortunately, ESD and power-transient testers are becoming relatively affordable. The cost of an integrated transient test system is approximately $20,000.

TROUBLESHOOTING

The second phase of precompliance testing is usually associated with troubleshooting. Potential problems are identified prior to packing up the equipment and heading to the test house. At this stage, use diagnostic techniques to create conditions that force failures. Accuracy is generally not too important; after a problem has been identified and fixed, a calibrated test will be conducted. Without this preliminary troubleshooting, there may be a nasty surprise at the test house: the equipment may fail the test, and the problem will need to be identified and fixed. Depending on the nature of the equipment and time availability at the test house, the problem-solving phase could take place either at the test house or back at the corporate facility. Having in-house precompliance testing capability can reduce the number of test failures, providing manufacturers with a high confidence level prior to compliance testing.

For troubleshooting, add some sniffer probes, which are small loop and dipole antennae used for detecting emissions. Emission leaks can be found by moving the sniffers around the equipment, especially near seams, openings, and cable entry points. Sniffers are commercially available, but crude ones can easily be made. Make a loop antenna by forming a loop at the end of a length of coax. The loop may be a single turn of a 1/2-in. diameter or several turns of a 3-in. diameter. The smaller loops are less sensitive but more selective. The best bet is to make several loops in various sizes.

These loops can also be used as a field source with the addition of a terminating resistor. Generally, use a small loop to inject RF onto an individual wire or trace or a larger loop to inject RF onto a cable. A baseline needs to be established because these are not calibrated tests (not even if calibrated sniffers are used), but they do force a failure on sensitive analog devices. Use an amplifier, because signal generators are unlikely to have enough power to cause an upset.

A crude RFI test can be done with handheld radio transmitters, usually transmitting at approximately 150 or 460 MHz. You can estimate the field strength based on the equation for power and distance:

in which electric field strength is in volts per meter, P is transmitter power in watts, and R is distance from the radio in meters. This test is definitely not calibrated and tests only two frequencies, but it is better than nothing. The test can generally identify weak spots in analog devices.

A crude power quality test uses a chattering relay, in which the normally closed contact is wired in series with the coil, then connected across line voltage and neutral. It will generate transients, but they are completely uncalibrated. This test can force failures. If a device can pass this test, it's fairly certain to pass the surge and EFT tests.

CONCLUSION

Full EMC compliance testing requires a major investment, generally more than medical manufacturing companies can justify. But compliance testing is only needed for the final product--less rigorous precompliance testing and troubleshooting can be done with a much smaller investment, allowing problems to be identified and remedied early in the design cycle. These tests, performed in the manufacturer's facilities, are a more effective use of engineering time, speeding up the development cycle. The result is reduced engineering development costs, reduced test costs, and, perhaps most importantly, shortened time to market.

William D. Kimmel and Daryl D. Gerke are principals in the EMI consulting firm Kimmel Gerke Associates, Ltd., based in St. Paul, MN.

Copyright ©1997 Medical Device & Diagnostic Industry