USB Charger (and SMPS) AC Leakage Experiment

To confirm my beliefs and to correct some erroneous information presented in this post My Mobile Charger Shocked Me While Charging, and other posts in this forum, I conducted an experiment. Call it put up or shut up.

Belief #1
USB chargers and by extension SMPSs often have AC leakage between the DC output and ground, being approximately one-half the AC input voltage. This leakage is at an extremely low current, and while it rarely may cause a perceptable shock, it is not hazardous and does NOT indicate a fault in the power supply. It's caused by filter capacitors in the power supply.

Belief #2
@MisterBill2 made the following statement which I believe to be entirely erroneous.

"A charger for a cell phone is intended to only be used for that purpose and so there is no reason at all that the circuit needs to be isolated from the mains. So from some chargers, depending on which way they are plugged in, connecting with the 5 volt DC output could connect you with the hot side of the 220 volt mains. THAT is the reason that I never suggest using them for other than charging a phone.
But folks are welcome to ignore my advice , and let natural selection 'do it's thing.'"

While this may have been a danger back in the days of 5 tube AC/DC radios, I don't believe it's true of any modern-day USB or "wall wart" SMPSs. Capacitive droppers and other non-isolated power supply designs are permitted only where human contact is impossible under normal circumstances. An Apple Lightning iPhone connector certainly exposes live contacts, as does a DC barrel jack. Even the contacts of a USB A connector could be touched with a slight bit of effort.


The Experiment

A collection of 15 random USB power supplies in my possession (the easy to find ones) was selected for this experiment. AC voltage between the positive output and AC power line ground and the negative output and AC power line ground were measured using a BK Precision Test Bench 391A high input impedance digital VOM. Next, a 10k Ohm resistor was shorted between the positive output and AC ground and the resulting voltage measured.

Information on the chargers is shown in the attachments.



I must admit that I was surprised by the results. Six of the chargers have AC leakage of over 40 volts - I had expected this would be almost universally true. Three of the chargers have leakage voltage between 12 and 20 volts, with the balance having leakage voltage of less than 10 volts. The results are shown in the table below.



When loaded down by a 10k Ohm resistor, the leakage voltage dropped to less than 1 volt in every case.


Conclusions

1. AC leakage of approximately one-half power line voltage is common, and does not reflect defects in a SMPS or a safety

2. The leakage current is low as evidenced by the effect of loading the output to ground with a 10k resistor. If there was any current behind it, the voltage would not have dropped below 1 volt. While a tingle may be occasionally felt, this does not represent a hazardous condition.

3. No evidence was seen of a direct connection between DC output and AC power input pins.

4. The contention that "chargers are only for phones...." is against the current philosophy of using USB and USB C PD (Power Delivery) chargers for all manner of DC-operated devices. In fact, many devices do not even include a charger due to prevalence of USB power supplies.

 

I applaud your scientific approach, and publishing real results.

However there are occasional reports of things like people taking their phones into a bathtub while the phone is still charging. Good way to get dead quickly.

As is implicit in @MisterBill2 ’s comment these chargers were not designed to prevent electrocution when inappropriately used.

 

Ditto to Dick’s comments. I’ll add that your sampling of chargers doesn’t necessarily represent the entire population of chargers out there. The fact that you already owned them means your sample is biased towards ones that are acceptable to you, and presumably you applied some quality standards when buying them. Your data implies your standards “worked” at least in regards to this topic. Your data sheds no light on the chargers you chose not to buy.

 

Obviously, I haven't tested every type of charger in existence, and if I had your next comment would be just because the one you tested is safe doesn't prove that EVERYONE is safe.

The chargers I tested are from a variety of sources and a variety of manufacturers. They came with products I've purchase new, products I've purchased used and even boxes of "cables and junk" purchased at garage sales. These were the ones easily gathered when I decided to do this experiment for the benefit of others. If you look at the attached pictured, you'll see these weren't "high-graded".

Please feel free to take what knowledge you like from this effort or none at all. If nothing else, perhaps you'll at least hesitate a moment before telling someone they are in mortal danger because they measure a common characteristic on a perfectly good power supply.

 

However there are occasional reports of things like people taking their phones into a bathtub while the phone is still charging. Good way to get dead quickly.

Sadly, you can't outlaw stupid. I don't think I want to live in a world designed to protect the stupidest of the stupid. It would be a pretty boring place.

Besides, stupid is supposed to hurt – it's a good learning tool.

 

The European standard (EN60950) requires that all accessible parts pass a 3kV flash test, and that in normal operation the leakage current is <250uA (It is 750uA in some circumstances).
The official way of measuring the leakage is by measuring the voltage across this network.
The concept of an electric shock being "acceptable" when plugging in phone chargers is propostorous.

 

But folks are welcome to ignore my advice , and let natural selection 'do it's thing.'"

In that case, natural selection would rather favour countries with 120V mains!

 

Your drawing is very nice, but taken out of context, I don't understand what it's trying to show.

From the same standard, I found this drawing under section 5.1, "Touch Current". This drawing I can understand. My "measuring network" between points A and B consisted of a 10k ohm resistor, with leakage voltage measured across this resistor. Leakage current may be calculated using Ohm's Law (which I won't reiterate here, as I'm somewhat sure some of you were his laboratory aids or drinking buddies). The network isn't per the standard but will serve to get a reasonable result.


The standard for "Touch Current" is shown in the table below. The maximum for "All Equipment" is 0.25 mA rms, with higher exceptions for certain applications.



The greatest leakage voltage I measured as shown in my table above is 0.7 VAC rms. Ohm's Law, in case you've forgotten, is:

V = i × r, which may be rearranged to:

i = V / r

i = 0.7 / 10000 = 0.07mA.

My worst-case measurement was 3.5 times less than the most stringent standard.

The point I have attempted to make (apparently without much success) is that it is common to measure as much as one-half AC line voltage between a SMPS output and power-line ground, and that this voltage does not indicate a defect or safety hazard in the power supply.

 

The point I have attempted to make (apparently without much success) is that it is common to measure as much as one-half AC line voltage between a SMPS output and power-line ground, and that this voltage does not indicate a defect or safety hazard in the power supply.

First, thanks for the excellent post. Very nicely done and some of the best content I've seen on AAC.

Yes, the voltage isn't the issue, it's how much current it can provide that's the issue. Even the best chargers can a leakage current but that current is tiny, and can produce no more than a tingle.

When a design is such that either the "working" configuration or a foreseeable fault can cause it to produce mains or ½ mains at substantial current, it is dangerous and for that reason the standards provide a mechanism to limit such cases as much as practical.

 

Your drawing is very nice, but taken out of context, I don't understand what it's trying to show.

From the same standard, I found this drawing under section 5.1, "Touch Current". This drawing I can understand. My "measuring network" between points A and B consisted of a 10k ohm resistor, with leakage voltage measured across this resistor. Leakage current may be calculated using Ohm's Law (which I won't reiterate here, as I'm somewhat sure some of you were his laboratory aids or drinking buddies). The network isn't per the standard but will serve to get a reasonable result.
View attachment 299965

The standard for "Touch Current" is shown in the table below. The maximum for "All Equipment" is 0.25 mA rms, with higher exceptions for certain applications.

View attachment 299966

The greatest leakage voltage I measured as shown in my table above is 0.7 VAC rms. Ohm's Law, in case you've forgotten, is:

V = i × r, which may be rearranged to:

i = V / r

i = 0.7 / 10000 = 0.07mA.

My worst-case measurement was 3.5 times less than the most stringent standard.

The point I have attempted to make (apparently without much success) is that it is common to measure as much as one-half AC line voltage between a SMPS output and power-line ground, and that this voltage does not indicate a defect or safety hazard in the power supply.

It's the "measuring network" from your drawing.

 

According to Google, anything above about 1mA through the human body is usually detectable.
So if you can sense a leakage current from a device, then it would appear to be exceeding the 250µA maximum spec requirements.

 

Excellent experiment, thanks for sharing.
However as evidenced by teardowns from Big Clive’s and Diodes Gone Wild’s YouTube channels, there are plenty of “suicide chargers” out there.
With minuscule creepage clearances.

 

Thanks for the educational research, and for posting the results. While the results point towards the shock hazard for those devices being minimal, I am guessing that all of those supplies were in good working condition, and had never been underwater for any length of time. But every one of them would have been able to put an undesirable hum voltage into an audio system. That is not a safety issue, but often quite a bother.

And I would also point out that intentionally contacting the connection portions of an electrical device connector while it is powered is not the good choice, no matter what.

 

You are absolutely correct. I didn't soak any of the supplies in water before testing them. I also didn't pound on the case with a hammer repeatedly before testing, nor did I barbecue them with a tangy honey glaze until the case melted. Didn't drag any of them behind the car for miles (1.6 kilometers) before testing or connect the output to line voltage. Didn't pound a nail through the middle of the case. Didn't borrow the neighbor's rabid Rottweiler to chew on the charger first.

Stupid people do stupid things and it's impossible to stop them. Sorry, stupid should hurt.

Again, I wanted to show that minute leakage current is a normal, safe thing with SMPSs and is not a cause to panic or to replace a perfectly functioning charger. The response on this forum is all too often shrieking overreaction when somebody, for whatever reason, measures leakage voltage on a SMPS.

I am making no contention that there are not hazardous chargers around. Either because of shoddy manufacture or something stupid like I listed above. But the indications of a hazardous charger are usually not half line voltage to ground.

 

Maybe this forum is better suited to endless posts like:

"Can anybody tell me what part goes here?

 

You are absolutely correct. I didn't soak any of the supplies in water before testing them. I also didn't pound on the case with a hammer repeatedly before testing, nor did I barbecue them with a tangy honey glaze until the case melted. Didn't drag any of them behind the car for miles (1.6 kilometers) before testing or connect the output to line voltage. Didn't pound a nail through the middle of the case. Didn't borrow the neighbor's rabid Rottweiler to chew on the charger first.

Stupid people do stupid things and it's impossible to stop them. Sorry, stupid should hurt.

Again, I wanted to show that minute leakage current is a normal, safe thing with SMPSs and is not a cause to panic or to replace a perfectly functioning charger. The response on this forum is all too often shrieking overreaction when somebody, for whatever reason, measures leakage voltage on a SMPS.

I am making no contention that there are not hazardous chargers around. Either because of shoddy manufacture or something stupid like I listed above. But the indications of a hazardous charger are usually not half line voltage to ground.

My point is that with just a bit of abuse and misuse, many devices can become hazardous but still function and not appear damaged. It is very easy to drop one of those chargers quite by accident. And getting them wet accidentally is also possible. So abuse does happen even among those folks not excessively stupid.

 

My point is that with just a bit of abuse and misuse, many devices can become hazardous but still function and not appear damaged.

Could not the same be said for ANY electrical device or extension cord? The grinding disk on an angle grinder bumped and fractured so it explodes when you pull the trigger? A piece of paper jams in your laser printer and bursts into flame?

Seems like you're working awfully hard to rationalize SMPSs into dangerous objects. Accidents happen, and hopefully people have enough common sense to recognize when something should not be used (Ok. This is too often not my experience. I have been known to say in the past "You can't legislate common sense.")

 

Could not the same be said for ANY electrical device or extension cord? The grinding disk on an angle grinder bumped and fractured so it explodes when you pull the trigger? A piece of paper jams in your laser printer and bursts into flame?

Seems like you're working awfully hard to rationalize SMPSs into dangerous objects. Accidents happen, and hopefully people have enough common sense to recognize when something should not be used (Ok. This is too often not my experience. I have been known to say in the past "You can't legislate common sense.")

Safety, like all engineering challenges, is an optimization problem. Perfection is not possible and practical solutions will always require balancing conflicting aspects.

The biggest problem is choosing how to weight the various things in conflict. The errors occur when the choices are overly influenced by either unquestioned assumptions or ethical failings. The manufacturer who ignores safety in favor of profit will cause harm if successful at selling but doesn't care because of the profit.

The manufacturer that has a culture of excess caution with increase the cost of their products so they either fail in the market of restrict availability to the more fortunate. If the ideal in this case is defined as the solution that helps the most people, or maximally reduces harm, it will have to include safety features that might reduce the profits of the manufacturer while maximizing the value to the customer.

In the current ethical climate, this is hard to justify. The idea that a manufacturer has some ethical responsibility to make no more from a product than allows the product to not kill anyone is seen as "idealistic". All industrial scale manufacturers use actuarial tables to decide how many deaths are acceptable as a trade-off in profits. If you actually look at it hard, it's quite ugly.

But, it's how we (corporately) see "reality". I am not convinced in the inevitability of the state of affairs, but this is not intended to be a political discussion at all, only a practical one, and it comes down to this: how safe things are made is a matter of the prevailing Weltanschauung, about things that we as a society can't even see to question.

Do things have to be perfectly safe? No, they can't be. But, I believe manufacturers are morally obligated to design and build things in such a way that when they are used as expected and intended they have as little potential for harm as is practical. And that if there is danger inherent in a product, its cost-benefit ratio needs to be made very clear to the people buying and using it.

 

"Y" has stated close to what I think, which is that the products are aimed at being safe when used as intended, or fairly close to as intended. But none of them are perfectly safe "at all times and under all conditions". Besides that, it is certainly true that You Can't Fix Stupid, but the laws prohibit natural selection.
All designs are a compromise of some sort, and in some imports, it is quality that loses.

 

At Apple’s Singapore plant every computer monitor made there went through a ground fault current test. If a monitor did not pass a buzzer sounded, lights flashed, and the production line stopped until an engineer reset the alarm and restarted the line.

I was very impressed with those guys in manufacturing.