English is currently available.
Biomedical Engineering DISCUSSION
Our project group is building a wearable that has electrodes in contact with skin, and the plan was to power and read it through a USB cable from a desktop PC. Our supervisor says this is not acceptable for anything attached to a person and mentioned leakage current limits and isolation.
What are the actual limits, why are they so low compared with ordinary appliance safety, and what design approach satisfies them?
The concern is current through the body if something fails, and the limits come from physiology. At mains frequency, a current through the body of around 1 mA can be felt, around 10 mA can stop a person from letting go, and currents in the region of 100 mA across the chest can cause ventricular fibrillation. Those figures are for current entering through intact skin.
If a conductive path leads directly to the heart, for example through a catheter, currents of the order of tens of microamps are considered dangerous. Medical standards are written around that case and around patients who cannot react or pull away.
IEC 60601-1, the base safety standard for medical electrical equipment, classifies the parts that touch the patient as type B, BF or CF applied parts. The patient leakage current limits usually quoted are 100 µA in normal condition and 500 µA with a single fault for types B and BF, and 10 µA and 50 µA for type CF, which is the class intended for direct cardiac connection. BF and CF parts are floating, meaning isolated from earth and from the rest of the equipment.
Read the current edition for the exact test conditions; the point is the scale: microamps, not the milliamps tolerated elsewhere.
A USB connection to a desktop PC fails on isolation. The USB ground is tied to the PC chassis and therefore to protective earth, and the PC power supply is built to information-technology safety rules, not medical ones. Your electrodes would give the person a low-impedance connection to earth, so any other faulty equipment they touch has a return path through them, and a fault in the PC supply could put voltage on the cable.
A laptop on battery is better than a desktop, but the moment its charger is plugged in the same objection applies.
The standard approach has two parts. Power the patient-connected side from a battery, or from a medical-grade isolated DC-DC converter, and pass data across an isolation barrier (digital isolators or optical) or over a radio link. The barrier is rated in means of patient protection: for mains working voltage, two means of patient protection are commonly specified as a 4 kV AC test voltage with 8 mm creepage and 5 mm clearance. Check the figures for your working voltage in the standard.
For a student prototype, battery plus Bluetooth is the simplest arrangement that avoids the question entirely. Make charging impossible while the device is worn, for instance by using the same connector for the electrodes and the charger.
Add current limiting at the electrodes as a second layer: series resistors sized so that even with the full battery voltage across a lead, the current stays in the microamp range, plus input protection that cannot itself become a source. With a 3.3 V supply and 330 kΩ in series, the worst-case current is 10 µA.
Finally, separate the engineering from the permissions. Meeting the numbers on paper does not make a device approved for use on people. Trials on volunteers normally need review by your institution's ethics or safety committee, so involve your supervisor before anyone wears it.
Back to Biomedical Engineering community
