SMPS (Switch Mode Power Supply) AC Leakage Experiment

[For The Popcorn]

Note: This experiment was posted on another forum. The text has been edited to better focus attention to the key points.

From time to time, the question of AC leakage from SMPS (Switch Mode Power Supplies) between the DC output and power-line ground comes up. It may be because someone perceived a shock, or someone measured it with a digital voltmeter. This often results in discussions that the power supply is defective and dangerous and should be thrown away.

This experiment was to investigate the premise that AC leakage of less than one-half line voltage is common, safe, and does not represent a safety issue.

The Experiment
A collection of 15 random USB chargers was gathered for testing. Details of each are contained in the attachments. AC leakage voltage was measured between positive and negative outputs and power-line ground. The measurement was repeated with a 10k load resistor between positive out and ground. AC line voltage was 122VAC at the time of the testing.

The results are shown in the table below. Six of the chargers had leakage voltage between 42 and 57 volts. Three of the chargers had leakage voltage between 12 and 20 volts. The balance of chargers had leakage voltage below 7 volts. None of the chargers had leakage voltage over 47% of line voltage.

Leakage current was calculated across the 10k resistor using Ohm's Law. The worst-case leakage current observed was 0.07mA. According to IEC Standard EN 60950-1:2006, the most stringent limitation for "Touch Current" is 0.25mA. The worst-case leakage observed during this experiment was 28% of the most stringent standard.

Conclusions

1. Leakage voltage below 50% of AC line voltage is common, safe and does not indicate a fault in a switch mode power supply.

2. Leakage current observed with leakage voltage below 50% of line voltage is safe.

3. Leakage voltage of 50% or more of AC line voltage is not typical and should be considered a hazard.

 

[Diver300]

I would agree that none of those chargers is remotely unsafe, even though you could probably feel the leakage in certain circumstances.

However, I don't think that voltage is a good measure at all, and whether the voltage is above or below half line voltage is more down to good luck than whether the appliance is safe or not.

A lot of appliances have a Y capacitor between the input and output. Some have two, one to line, one to neutral. The current to ground depends on the value of the Y capacitor that is connected to live, and that is what makes it safe or not.

The voltage, when measured with a high-impedance voltmeter, depends on the ratio between the live side Y capacitor and the neutral side Y capacitor. If the neutral side Y capacitor is small or missing, that will cause the voltage to be above half the supply, but that alone tells you next to nothing about the leakage current if there is a fault. Conversely, there could be very large values of Y capacitor to both live and neutral, so the secondary voltage would be near to half the supply voltage, but the current to ground could be dangerous if the capacitors are too large.

 

[For The Popcorn]

My comments specifically refer to switch mode power supplies, not appliances in general.

My point is, that in general, measuring up to nearly half of line voltage leakage voltage does not indicate a fault, and if there is doubt, it's easy to confirm by loading the circuit down with a 10k resistor. If the voltage across the resistor is 2.5 volts or less, the leakage would be considered acceptable in accordance with appropriate standards.

It's possible that any switch mode power supply may be defective and hazardous either through defect or abuse. Within the parameters I have stated here, this is unlikely.

My goal is to prevent uninformed individuals from declaring a power supply or charger exhibiting typical characteristics as defective and hazardous.

 

[Diver300]

My point is, that in general, measuring up to nearly half of line voltage leakage voltage does not indicate a fault, and if there is doubt, it's easy to confirm by loading the circuit down with a 10k resistor. If the voltage across the resistor is 2.5 volts or less, the leakage would be considered acceptable in accordance with appropriate standards.

My point is that the 10 k resistor measurement is the important one, far more so than the voltage. I think that some power supplies have just one Y capacitor, and if it can be plugged in either way round, the voltage to ground could be quite large.

Switch mode power supplies tend to be the ones that have the Y capacitors to reduce noise, but there can be Y capacitors in anything.

 

[Nigel Goodwin]

I don't really see the point? - they are class II devices, working exactly as intended - I can't see why anyone would waste their time measuring the designed 'leakage' of them. The reasons for it are perfectly obvious, and an essential part of the design.

 

[For The Popcorn]

I guess you don't read all the posts here Nigel Goodwin – there have been a number of posts along the lines of "I got a shock from my (charger, phone, laptop). When I measured the voltage between the (power supply, point I touched, etc.), it was almost half the powerline voltage!"

This inevitably generates a bunch of replies that the device is hazardous and should not be used. That plugging it in will certainly result in injury and death.

I did actually explain this in my first paragraph.

From time to time, the question of AC leakage from SMPS (Switch Mode Power Supplies) between the DC output and power-line ground comes up. It may be because someone perceived a shock, or someone measured it with a digital voltmeter. This often results in discussions that the power supply is defective and dangerous and should be thrown away.

Just trying to demonstrate that this is a common safe condition to save people worry or replacing perfectly good devices.

 

[ronsimpson]

There are different ways to measure line leakage. Amy one with a meter can measure 50/60hz voltage. Power supplies often have two Y caps that sit from line to output and back to the other line. This forms a 2:1 ac voltage divider. It is somewhat common to get a 1/2 line voltage reading using a high impedance voltage meter. I worry about the current not the voltage.

There are two different frequencies to measure in. The 50-60hz and the first harmonics. I spend most of my time chasing current at the switching frequency and high frequency ringing. (100khz-500mhz switching) and (1 to 30mhz ringing) Most meters will not measure 500khz ac current.

Medical power supplies have a very low leakage rating. The Y caps are small to keep ground current down.

I think much of the "got shocked from my phone charger" comes from static electricity and would have happened even if the main power was off.