Controlling Cardiac Probe Leakage Current

Stephen D. Anderson, Consultant

Medical procedures sometimes require direct connection of electronic instrument probes to the heart. Examples are direct EKG electrodes and thermocouples. The probe wires usually pass through a catheter to some external measurement equipment that sits beside the patient. When this equipment must be operated from the AC mains, a major design problem is leakage current at the mains frequencies of 50 Hz and 60 Hz. Currents of only a few tens of microamps through the heart can be fatal. Since leakage current is normally limited by small capacitances in isolating components; the challenge is to choose the isolating components, arrange the board layout, and separate components for minimum capacitance. This article shows you how to design for minimum leakage and pass IEC-601 leakage tests.

IEC-601 (Reference) completely specifies the requirements and tests for medical equipment in many countries. For electrical equipment operated from the AC Mains, the standard focuses on two issues: high voltage breakdown and current leakage. Current leakage directly affects the patient, while breakdown can indirectly affect the patient by increasing leakage.

The standard defines the probe and its electrical connections as an "Applied Part." A probe connected to the heart is a "type CF Applied Part." The probe and associated equipment are referred to as "Class 1 Equipment" which simply means it's operated from the AC Mains and normally has a safety ground wire (3rd wire).

There are four different leakage currents defined. The worst of these in terms of difficulty in passing the tests, is the current through the patient. This is defined, appropriately, as "patient leakage current." Different types of Applied Parts have different limits on patient leakage current. A type CF Applied Part has two leakage limits: 10 uA and 50 uA RMS. The 10 uA applies under normal conditions, while 50 uA applies in the event that the patient becomes connected to the Mains voltage through failure in some other equipment.

The leakage should be verified through measurement. But at these small current levels, stray coupling can render the measurements useless. Special shielding and isolation techniques are needed. This article shows how to minimize leakage and how to predict and measure it under the prescribed IEC-601 tests.

Two IEC-601 Leakage Tests

The leakage paths for the two tests are illustrated in figures 1A and 1B. In figure 1A current leaks into the grounded patient from the probe. The leakage source is the AC Mains source that powers the device. Here, the leakage passes through the capacitance between power transformer primary and secondary windings.

In figure 1B the patient is connected to the AC Mains. The leakage occurs from the patient into the probe. From there it takes various paths to ground -- some through direct capacitances to ground, and others through the power transformer and then through the AC Mains source that powers the equipment.

Of the two types of patient leakage current, the figure 1B leakage is invariably larger, since AC Mains voltage is effectively connected directly to the probe. Moreover, the leakage present in figure 1A is still present in figure 1B. There is the possibility in figure 1B that the leakages due to the combined voltage sources will be greater than leakage due to either source acting alone. At worst, the available voltage driving the leakage in figure 1B is twice the Mains voltage.

The Mains voltage sources are shown with one side grounded. IEC-601 tests require this, regardless of actual practice in Mains wiring.

Figure 1A -- Leakage Current Path Through Patient due to AC Mains Voltage that Operates the Equipment

Figure 1B -- Leakage Current Path Through Patient due to Both AC Mains Voltage that Operates Equipment and
AC Mains Voltage Applied to Patient

Minimizing Patient Leakage Current

The job would be easier if you could use a battery operated circuit encased in a good (low capacitance to ground) insulator. Since this eliminates the figure 1A leakage and reduces the figure 1B leakage, you should at least explore the possibility. If you decide it just isn't an option, other ways to "break" the leakage path must be used. In figure 2, a combination of an isolation amplifier and high-isolation DC-DC converter are used. Some initial conditioning of the signal occurs at the isolated side, while most of it occurs downstream at the non-isolated side. The non-isolated side would include any display, computing, recording, or data communication functions.

Figure 2 -- Equipment Circuit Divided into Isolated and Non-Isolated Sides and Showing Isolating Components

Figure 3 -- Physical Circuit Layout

Physically the circuit is as shown in figure 3. The circuit board is mounted on standoffs inside the cabinet. The isolated and non-isolated sides each have their own ground planes. A feedthrough connector brings the probe wires through the device cabinet to the isolated side of the circuit board. Three grounds are identified. G1 is the isolated ground plane on the board. G2 is the non-isolated ground plane on the board. G3 is enclosure ground + safety ground. (The resistances of the isolating components are usually so high that they can be ignored. If you don't have this situation then you may have to include the resistances as leakage paths.) The resulting capacitances are shown in figure 4. They are:

C1 = Combined capacitance to ground of the probe feedthrough connector, connector-to-board cable, G1 to G3 capacitance,
and capacitance from isolated components to G3.
C2 = Combined capacitances through isolation amplifier and isolating DC-DC converter (capacitance from G1 to G2).
C3 = Combined capacitance from non-isolated ground plane to enclosure (capacitance from G2 to G3).
C4 = Capacitance from non-isolated ground plane to power supply AC Mains.

Figure 4 -- Circuit For Leakage Calculation

All of the circuitry on the isolated side has been lumped into isolated ground G1. Similarly all of the circuitry on the non-isolated side has been lumped into non-isolated ground G2. The probe wires have been modeled as just a single wire. For figure 1A leakage the probe goes to ground. For figure 1B leakage it goes to the 2nd Mains source.

The circuit is deliberately simple and may omit elements that help reduce leakage. For example, if your circuit has a large resistance or very small capacitance between the probe and G1 (The high impedance must be present in all of the probe wires.), then you may want to do a more complete analysis and include this impedance.

The equipment power transformer is generally a complex structure, with distributed capacitance from primary to ground (G3), secondary to ground, and from primary to secondary. This makes it difficult to assign values to C3 and C4. It can also create a voltage divider that may apply only a portion of the Mains voltage to C4, rather than the full Mains voltage as shown in the figure. If an off-line switching power supply or a shielded transformer is used, this further complicates the leakage model at C3 and C4. A conservative way of dealing with this situation is to assume that C4 is short-circuited and that the full Mains voltage is applied to G2. This rather sweeping assumption is supported as follows. First, assume that most of the isolation is provided by the isolating components. That is, C2 is much smaller than C4, making C4 a short-circuit compared to C2. Second, the figure 1B test generally produces a higher leakage current because of the direct connection of the second AC Mains source to the probe. If, when both sources are present, most of the current is due to this second source, then the exact value of the first Mains source is less important. To be conservative, set it equal to the Mains voltage.

Calculating the Current

In figure 1A there is only one voltage source. To calculate the leakage, ground the probe (to G3), short-circuit C4, and assume that the whole Mains voltage is applied from G2 to G3. Then the current in the probe is simply

Figure 1B leakage is the leakage found for figure 1A combined with the leakage due to the 2nd Mains source. To calculate the leakage due to only this source, short-circuit C4 and replace the first Mains source with a short-circuit to ground (G3). The leakage flowing through the probe is then

You should assume that the combined leakages due to the two sources could add together in the worst case. Then the total leakage for figure 1B is

The leakage is minimized by minimizing C1 and C2. First of all, select isolating components for lowest possible capacitance. (Voltage withstand capability of these parts is also important and may narrow the choices.) Next, make sure that the feedthrough and cable are well separated from ground G3. There should be a wide separation between components and ground planes (G1 and G2) on the board. Finally, keep the circuit board and its components away from the enclosure to minimize G1 to G3 capacitance. Try to avoid putting physically large components such as a heat sink on the isolated side of the circuit. If the probe cable must run along and be clamped to a surface, avoid using the enclosure walls. If an enclosure wall is the only reasonable choice, then separate the wall and cable with a thick piece of plastic.

Using this approach the calculated figure 1B leakage will always be larger than the calculated figure 1A leakage. The goal is to have less than 10 uA for figure 1A and less than 50 uA for figure 1B.

An Example

As an example of the leakage current calculations, assume that there are 8 isolating components: 4 isolation amplifiers, and 4 DC-DC converters. These are rated in the following table.

COMPONENT	CAPACITANCE	RESISTANCE
Amplifier	2 pf	10¹⁴ ohm
Converter	6 pf	10¹⁰ ohm

The combination is then 32 pf in parallel with 2.5 Gohm. The circuitry at the isolated side is estimated to have a board area of 9 square inches and the board is to be mounted about 1/4 inch away from an enclosure wall. This creates an air-dielectric capacitor of about 8 pf. Suppose that the feedthrough was measured and found to have a capacitance to enclosure of 2 pf. The cable-to-enclosure capacitance is assumed to be 5 pf. The capacitance between ground planes G1 and G2 due to their proximity is estimated to be 1 pf. Then the capacitances are C1 = 15 pf, C2 = 33 pf. Use V_mains = 266 volt and assume the equipment must operate at either 50 Hz or 60 Hz. Use f = 60 Hz to produce the greater leakage. Then the calculated figure 1A and 1B leakages are 3.3 uA and 8.1 uA.

Performing the Tests

The tests for the two leakages are parts 19.4.h.1 and 19.4.h.2 of IEC-601. The test setups are figures 20 and 21 of IEC-601. These are reproduced here in simplified form as figures 5A and 5B. The actual IEC test setups have several switches that must be thrown to reverse the polarity of applied voltages and to simulate fault conditions. "MD" is the current measuring device prescribed by IEC-601. It consists of an AC voltmeter and a small RC network that are shown in figure 6. At 50 Hz or 60 Hz the MD looks like a 1000 ohm resistor. Divide the voltage reading by 1 kohm to get leakage current Comparing figures 1 and 5 you can see that the MD replaces the patient. The impedance of the MD is so small as to be almost a short-circuit compared to the isolating impedances.

For both tests the Mains sources must be adjusted to 110% of the highest rated Mains voltage. For Europe this is probably 242 volt, which results in a test voltage of 266 volt RMS. IEC-601 shows the voltage sources implemented as adjustable isolation transformers. If these aren’t readily available, a practical alternative is a fixed isolation transformer followed by an autotransformer. If you're building a permanent test setup you'll want to include the autotransformers to adjust the voltage and panel meters to show you what the voltage is. However, if a 3rd party will eventually do the testing and you'd like just an estimate of the leakage current, you can simplify things by using only fixed isolation transformers.

Using fixed transformers and the available Mains voltage gives you a leakage measurement that is low. You can use the measured nominal Mains voltage to calculate the current that would exist at the required test voltage. If you have only 120 VAC available and you need 240 VAC, use fixed transformers that have a 1-to-2 step-up ratio.

In figure 5B a 100 kohm current limiting resistor is used to prevent the flow of a large current in the event of a failure in the equipment under test. This is small enough that it doesn’t affect the test.

You can include the prescribed IEC-601 switches in your test setup. Or you can perform the switching manually by re-arranging the wiring. Using switches is probably safer. The switches are used to change polarity of the Mains sources, remove the safety ground wire, introduce fault conditions, etc. Some switch positions cause the equipment to be powered. Others turn it off. The leakage must be measured for every combination of switch positions. The highest of these measurements is your leakage value. Switch positions that simulate fault conditions allow a higher leakage value. Normal and fault leakage values are given in table IV of IEC-601. IEC-601 also specifies various environmental conditions for the tests.

When selecting the transformers, remember that one of them has to be sized large enough to power the equipment under test.

Figure 5A -- Test For Leakage of Figure 1A

Figure 5B -- Test For Leakage of Figure 1B

Figure 6 -- Measuring Device

The MD deserves special attention, since stray coupling can easily render measurements useless. This occurs because the DVM has a large capacitance between its circuitry and any nearby conductor, such as ground. This is illustrated in figure 7 for figure 1B leakage. Part of the current goes into the equipment probe, while the rest goes into the DVM and back to ground. To avoid this you can use a battery-operated DVM and put the whole MD inside a metal box with a window. The box, itself, becomes the connection to the applied voltage source for figure 1B measurements or to ground for figure 1A measurements. A small probe through a hole in the box serves as the connection to the equipment under test. This is illustrated in figure 8.

Figure 7 -- Current Path Through MD Leading to Incorrect Measurement

Figure 8 -- Removal of DVM Current Path Through MD

Final Remarks

If one or more of your measured currents is much larger (or smaller) than the calculated value, then either your test setup is wrong or the circuit isn’t what you think. Are there current paths that aren’t included in the circuit of figure 4? Is the printed circuit board free of all metal in the area between grounds G1 and G2? Could one or more of the isolation components have failed? Normally, leakage tests are done after HIPOT testing. But if you’re experiencing problems, you should do leakage tests on a few units that haven’t been HIPOT tested, just to make sure that they start out OK. At worst, you may have to dismantle the equipment piece-by-piece to locate the offending leakage path.

We all tend to focus on the functions of a given device and not on safety features. However, devices of the type described here leave very little maneuvering room if you don’t consider both leakages and breakdown voltages up front.

The focus of this article has been on a specific type of equipment and on the IEC-601 leakage current tests that are the most difficult for this equipment to meet. The two tests described here are only a few of many that IEC-601 requires. You should consult the actual document before proceeding with any design.

Reference

"Medical Electrical Equipment, Part 1: General Requirements For Safety," IEC-601-1, 2nd Ed., 1988, International Electrotechnical Commission, Geneva, Switzerland.

For more information on how Analog Services, Inc. can help solve your circuit/system problems, call or e-mail us today.

Contacting Analog Services, Inc.:

VOICE 612-927-7112 FAX 612-929-7503 E-Mail: stevea@analogservices.com

home