IEC 60601-1 contains general safety guidelines for medical equipment. Over the years, these requirements have been refined and made more inclusive. A surge test has been incorporated in the standard for years, designed to ensure that any patient-connected equipment can function after application of a defibrillation pulse.

The 2006 version of the standard introduces the energy measurement test, which is designed to ensure that the patient-connected equipment will not consume more than 10% of the applied defibrillation pulse. As an overview, the energy measurement test has appeared in IEC 60601-2-49, and has been added to IEC 60601-1 in this version. The same 5-kV capacitor is used, but many of the component values are different. Figure 1 shows the circuit used for common and differential mode testing. These are both tests that appeared in IEC 60601-1, edition 2, and are conducted in the same way now. Figure 2 shows the circuit used in the energy measurement test, new for this issue.

Figure 1. Test voltage circuits of individual patient connections for defibrillation-proof applied parts.

The energy measurement test will require different test methods and modifications to the surge generator used to conduct correctly. We have found that older generators used for the common- and differential-mode testing may not be designed adequately for proper evaluation of the energy measurement test. If you are wondering whether your generator can be used, check on the front or rear panel to see if it is equipped with an energy measurement port, located at the node between the 400 Ω and 100 Ω resistors (see Figure 2). If your generator has one, you must also check the resistance value between the port and ground, before and after pulse delivery, to make sure it is within 5% of the nominal value of 100 Ω in both instances. If the surge tester satisfies these two criteria, then it can conduct the energy measurement test.

Surge Generator Design for the Energy Measurement Test

The energy measurement test in IEC 60601-1 Edition 3 requires the operator to conduct two tests: one with the device disconnected to the generator, and another with the device under test (DUT) connected. A passing result is achieved if the two energy measurements are within 10% of each other. Surge tester designers and technicians must use special methods to make sure this test is conducted fairly.
The standard specifies a circuit diagram for the test, but does not specify a resulting waveform. Therefore, the builder of the surge tester must ensure it is built correctly, or the waveform will not be correct. Figure 3 shows a simulation of the output of the energy measurement test without a DUT connected, assuming real-world components.

Figure 2. This voltage model is new to the standard. It tests delivered defibrillation energy.

Because the standard specifies component values and tolerances, an accredited ISO 17025 calibration may not necessarily ensure proper generator output. The inductor or resistor banks must be carefully designed to yield correct results.
A surge tester with a 100-Ω resistor bank that could pass an ISO 17025 calibration may not be in tolerance after a surge is delivered. Because the equation for power incorporates the resistance value, accurate results dictate that this resistor does not change value when it is heated by the delivered surge. Because two tests are required to be performed for the energy measurement test, it is also important that the resistor not change value between tests.

Anytime resistors are used in a circuit, they exhibit heating, which causes its resistance value to decrease. The change in resistance is specified by the resistor manufacturer as the temperature coefficient. This change in resistance, due to the heating effects of previous surges, can alter the power measurement result. If the resistors heat up enough to change value substantially between tests, more of the pulse energy is consumed in the resistor bank in the second test, passing less energy to the DUT. This could allow conforming equipment to show a failing result if the resistor bank is not performing correctly when heated.

If the resistance bank is not correctly sized, heating while the pulse is being delivered could result in incorrect energy measurement. This is because the equation for energy assumes a constant value of resistance for the duration of the test, which is unrealistic. A realistically rated resistor bank will exhibit < 5% change in value during the pulse delivery, and will be rated to stay within 5% pulse to pulse in continuous duty. The 5% pulse-to-pulse accuracy is controlled by the duty cycle assigned to the surge tester. To make energy measurements accurate, the author suggests measuring the resistance of the 100-Ω resistor both before and directly after the test, and using the average value as R in the energy formula.
Inductor design is also important, the standard notes. If the inductor saturates during pulse delivery, the waveform will be affected. The full 360 J of energy is still delivered, but if the inductor saturates, it will be delivered in a different waveform. The waveform resulting from a saturated inductor has a steeper rise and a shorter duration, which results in a more stringent test.

Although IEC 60601-1 makes special mention of the importance of correct inductor design, the author has found that the design of the resistor bank has a much greater impact on the pulse delivered to the device under test. If an incorrectly designed resistor bank is used, power delivery can be much less than intended. Unlike the case of the saturated inductor, which makes the test more stringent, unintended heating of an incorrectly sized resistor bank can invalidate the energy measurement test. As can be seen from the following formula for energy, a heating resistance bank will have a direct effect on the energy calculation between tests:

Because operational problems might not be caught in the calibration process, the generator’s waveform should be checked. The output of the surge generator should be compared with the generator’s calibration waveforms, and any anomalies should be investigated before the tester is put back into service.
The generator’s switch technology is also of some importance. When switching the energy from the bulk cap to the device under test, the switch used must pass the energy and provide the correct waveform. Solid-state switches are the best way to handle this task, but those rated for 5 kV are prohibitively expensive. Technology is moving fast on this front and within a few years perhaps an economically feasible solid-state switch will be available for 5000-V surge testers. In the meantime, we are left with two other switch type: the mechanical relay and the vacuum relay. The vaccum relay is a specially constructed mechanical relay with the contacts in a vacuum, so arcing is minimized.

The author has evaluated performance of the mechanical and vacuum relays used in this generator and has found the results at 5 kV are similar between the two methods. However, experience has shown considerable improvement using the vacuum switch at very low voltages around 400 V. If testing at this voltage is contemplated, then the vacuum switch can show improved waveforms. Some generators on the marketplace offer the vacuum switch either as standard equipment or as an option, if a cleaner waveform rise at lower voltages is desired. 

Duty Cycle Considerations

The manufacturer (or builder) has presumably designed and tested the resistor bank within the defibrillator-proof surge tester, and has determined the duty cycle that can preserve the resistance within the tolerance stipulated by IEC 60601-1 during continuous testing. The user should ensure that all test procedures take this duty cycle into account to confirm that the resistor is within tolerances for all testing. 

However, the energy measurement test is an exception to this duty cycle, and it is suggested that the manufacturer not conduct the test with the DUT connected until enough time has elapsed to allow the resistor bank in the tester to recover completely. Some examples are featured here to illustrate the point.

Example 1

Because the energy measurement test method in IEC 60601-1 specifies that the referee test (with the DUT disconnected) is to be conducted first, any resistor bank heating means that less energy is available for consumption by the device.

Therefore a possibly passing device could erroneously show as failing. For example, consider an in-tolerance resistance bank shift of 0% between the two tests, which would occur if using perfect resistors, or if the resistors were allowed to cool completely between tests (and as anticipated by IEC 60601-1):

Test 1 (referee test to see how much energy is available):

•    Energy consumed by the resistor bank: 360 J.
•    Energy consumed by the DUT: not applicable (N/A).
•    Theoretical energy allowance for DUT: 36 J.

Test 2 (test with DUT connected):

•    From Test 1 above, we know that 360 J is developed by the tester.
•    Change in resistance between tests: 0 Ω.
•    Change in power delivered by the tester: 0 J.
•    Theoretical energy allowance for DUT (assumed from Test 1): 36 J.
•    Theoretical max energy allowed in resistor bank: 360 J – 36 J = 324 J.
•    Actual power delivered by the tester: 360 J.
•    Actual energy allowance for DUT (from actual): 36 J.
•    Actual energy allowance for DUT: 360 J – 36 J = 324 J.

Based on these numbers, the 10% energy consumption allowance for the DUT works out fine. If the resistor bank consumes less than 324 J, then the DUT fails, because it is consuming more than 10% of the power delivered and is therefore not in compliance with IEC 60601-1.

Example 2

In this scenario, the resistor bank heats between tests, so the resistance is 5% less in the second test. The first test shows the same results as before.

Test 1 (referee test to see how much energy is available):

•    Energy consumed by the resistor bank: 360 J.
•    Energy consumed by the device under test: N/A.
•    Theoretical energy allowance for DUT: 36 J.

But the second test is taking place with a hot resistance bank for which the value has decreased by 5% (still within tolerance limits shown in IEC 60601-1). The problem is that the result is calculated using the results of Test 1, and the technician assumes that the resistor bank is still cranking out 360 J. Let’s see if that is true.

Test 2 (test with DUT connected):

•    From Test 1 above, the operator assumes that the tester develops 360 J.
•    Change in resistance between tests: –5% Ω.
•    Change in power delivered by the tester: –18 J.
•    Theoretical energy allowance for DUT (assumed from Test 1): 36 J.
•    Theoretical max energy allowed in resistor bank: 360 J – 36 J = 324 J.
•    Actual power delivered by the tester: 342 J.
•    Actual energy allowance for DUT (from actual): 34.2 J.
•    Actual energy allowance for DUT: 342 J – 34.2 J = 307.8 J.

This result shows that through resistor heating with an intolerance resistor bank (5% tolerance is within the tolerance allowed by the Standard), half the energy allowance for energy consumption by the DUT in the energy measurement test is consumed by the tester itself. In Test 2, the total energy delivered was not 360 J, but 342 J, the output difference solely due to the hotter resistors. The operator has no way of knowing this, so he is assuming the pass point is (360 × 0.9) = 324 J. 

However, the real pass point is (342 × 0.9) = 308 J. If the test result is between 308 and 324 J, the tech may erroneously conclude that the DUT failed the energy measurement test. The conclusion here is if the resistor bank is not allowed to cool properly between the two tests, complying DUTs may show failing results.

Based on these conclusions, the technician should measure the resistance bank using the energy measurement port, and wait to complete the second test until the resistor bank has recovered from the first test. Remember that the result in Test 2 is with a complying resistor bank, so the goal is not to worry about compliance of the tester, but to make sure that the resistor bank is completely recovered to its original resistance before the second test is conducted. Doing so ensures that the full 10% energy allowance is available to the DUT.

Another method to maintain accuracy is to measure the resistor bank before and after each of the two tests, and to use the two resulting average resistances to calculate power for each instance. The operator should then normalize to simulate equal resistance values for each test. Although this method corrects for inevitable treating effects of real-world resistors, it is not supported in the standard.

Therefore the safest course to follow is to wait for resistor bank recovery between tests.

Finally, the sequence of tests stipulated in IEC 60601-1 is important. The referee test with the DUT disconnected must be conducted first, followed by the test with the DUT connected. If these tests are run out of order, or if the resistor bank is not allowed to recover completely, failing DUTs may exhibit passing results due to heating of the resistor bank within the surge tester.

Computer Interface Considerations

Surge testing puts a large amount of voltage and current on a DUT for a very short period of time. This power is difficult to contain entirely within the tester and the DUT. A portion of the pulse energy will flow back through the power line, and pmight affect other equipment in the laboratory.

This problem is exacerbated with any equipment connected to the surge tester. In many environments, computers are used for surge tester control and data acquisition. The author’s particular environment includes many surge testers being developed, tested, and calibrated. This experience has resulted in the following observations:

•    Personal computers in proximity to ongoing surge tests tend to have problems. The computers don’t have to be directly connected to the surge tester to experience these problems, and surge suppressors are not that helpful. If a lab has a combination of surge testers and computers, it is recommended that a separate power line be run for the surge testers. Doing this has helped reset problems, but any long-term improvement of the computers when this modification is implemented is unknown. It is suggested that users treat the computer workstations in the surge lab as expendable. Back up often to the server and don’t keep any data on the workstations. 
•    USB connections between surge test equipment work best when the surge test equipment is connected to a dedicated power line. Another option is a hybrid connection scheme that presents the USB connection to the computer and a robust RS-232 connection to the test equipment.

Conclusion

IEC 60601-1 stipulates surge testing to ensure defibrillator-proof interfaces on patient-connected equipment. Surge test design appears straightforward, but in the absence of waveform guidance from IEC 60601-1, caution must be exercised to ensure correct output. The operator must be careful to follow duty cycle times in accordance with the manufacturer’s specifications, but the energy measurement test requires stricter guidelines to ensure fair testing. 

The energy measurement test, new to IEC 60601-1:2006, presents significant challenges for both the surge tester designer. However, by following the guidelines of this article, equipment manufacturers can be judged fairly under the new requirements.

Jeffrey D. Lind is president of Compliance West USA (Del Mar, CA). He can be reached at jlind@compwest.com.