# Galvanic isolation - threshold for considering a voltage to be dangerous

I have been wrestling with the philosophy behind galvanic isolation, and whether or not it makes sense to employ it in an application I am working on.

I understand galvanic isolation is required for wired communications between two circuits operating at different ground references (floating w.r.t. one another). The need for galvanic isolation in this case is self-explanatory (comms won't work, stuff gets fried, etc).

I also understand that the concept behind galvanic isolation for safety is to eliminate the conductive path between a power source and a point in the circuit that the user might touch (e.g. a jack or other wire-to-board connector). In this case, if there is a problem with a circuit that the user touches, no current will flow through the user to earth ground.

My question is this:

At what DC and AC voltages is galvanic isolation required from a safety perspective?

For example, if I want to measure AC and DC signals, at what point do I need to isolate the ADC section of my board from the section that supports user interface peripherals (see below diagram)?

In other words, how big does V1 or V2 have to be before I need galvanic isolation between the ADC section and the rest of the board?

simulate this circuit – Schematic created using CircuitLab

• Where is your system installed? 30VDC across your chest can kill you if your skin is wet and conductive. If you're dry, 30VDC probably won't do anything to you.
– vofa
Commented Jan 4, 2017 at 21:02
• System will be inside an enclosure, in a building, in rural Africa (so, probably pretty dry). Commented Jan 5, 2017 at 13:44

In Europe, the Low Voltage directive points to documents (UL 60950-1) that define the Safety Extra low voltage (SELV) as being: -

The voltage between any two accessible parts/conductors or between a single accessible part/conductor and earth must not exceed a safe value, which is defined as 42.4 VAC peak or 60VDC for no longer than 200 ms during normal operation. Under a single fault condition, these limits are allowed to go higher to 71VAC peak or 120VDC for no longer than 20 ms

Definition not taken directly from standard but from here. But there are other definitions that are slightly different. "ELV" covered by Wiki is found here.

So, do some research on it because there isn't common agreement in the world.

Maximum allowed voltages are defined by law, IIRC it is about 50V AC and 120V DC for dry areas.

Anything that connects to the mains voltage is considered not safe unless it uses an isolated power supply (most electronic devices) or is made in such a way that the user cannot touch any metal part that may have voltage on it (simple devices like a fan or a heater).

A lot of countries use non-polarized mains plugs, so the "neutral" only has 50% chance of being neutral. In addition, if there is a fault in the wiring of the house (say, the neutral wire fails), then the "neutral" may become hot.

• Assuming your hands are wet and one of them is solidly connected to system ground, a conservative analysis is: 120VDC/R_human = 120VDC/500 Ohms = 240mA. You will die.
– vofa
Commented Jan 4, 2017 at 21:08
• @vofa some guy managed to kill himself with a 9V battery. Anyway, Wikipedia also specifies this en.wikipedia.org/wiki/Extra-low_voltage Commented Jan 4, 2017 at 21:14
• Yes, 9V can kill you if you connect the electrodes to your bloodstream or have gaping open wounds on your hands. All the juices inside the human body are very conductive. The difference is that 120V can kill you even if you aren't being an idiot. Note that those ELV levels are for dry areas only, and wet areas have stricter rules. The technical capability of the user/service personnel is important too. In my opinion, even clueless end-users should not be exposed to any hazard.
– vofa
Commented Jan 4, 2017 at 21:20
• @vola, I wrote my answer according to the standard definition of ELV. Anyway, I edited my answer to mention that this is for dry areas. Commented Jan 4, 2017 at 21:22
• Is the theory that a failure of one of the electrical components could tie the high voltage to a power supply rail? For example, if a biasing resistor in the offset stage failed as a short and conducted high AC voltage onto a 5V power rail while someone was using the keypad. Commented Jan 4, 2017 at 22:05