Laptop Power Supply adaptors - mains on the output!

[akis]

This following investigation and questions started when I realised I was getting mild/medium electric shocks from a laptop power adaptor.

I have a number of laptop power supply adapters, 3 Sony VAIO, and about 10 generic ones, all at 19V-19.5V and 4A-5A. They are all switched mode.

My laptop's own adapter (Sony Vaio) uses a two core mains cable (ie no earth), but the other two Sony adaptors use a 3 core mains cable. All the generic adaptors use 3 core mains cables.

I have plugged the oscilloscope in differential mode on the V+ and V- output of each adaptor. The picture on the scope is terrifying - I am getting mains voltage at the output. The V+ and V- curves are *almost* 180 degrees out of phase, and the amplitudes are not equal so V = V+ + V- gives out tens even hundreds of volts and that is where I am getting the shocks from.

The generic adaptors also produce the same crap at the output.

But there is a difference. All Sony Vaio adaptors, those with mains earth and without, have no connection between earth and output V-. All the generic adaptors have a wire link on the PCB linking mains earth and output V-.

So here I am scratching my head.

1) Why do the replacement adaptors connect mains earth with V- (but the Sony ones do not)? What gives them the right to assume that V- will be connected to earth downstream? What if you have another powered adaptor connected to the laptop somehow, through external devices such as printers, USB, TVs, monitors and so on ? Imagine if you wanted to put two such power supplies in series to double the voltage, as the earth is connected to the V- you would have a short.

2) Why do all the adaptors (almost a dozen or more) give out such terrible voltages at the output ?

[Napalm2002]

Ghost voltage? Bad ripple caps? Did u put a dmm on the output? What ac volts do u have at the Output? What dc volts do u have at the output?

[Zepnat]

I get bit fed up with these 'floating' power supplies these days why cant we have the negative output mains earthed like years ago. I get mild shocks off about half the phone/battery chargers in our house they light up those cheap neon voltage tester no problem, I got a real belt off our Panasonic tv when plugging the aeriel in. Your can see quite a spark when touching the plug the the tv socket in the dark!

[akis]

The SKY satellite box has given me plenty of mild shocks in the past.

I do not think earthing the "negative" is a good idea, because then you have an implied short to the earth - there could be some other equipment in the chain which has the "positive" earthed, or some other, random voltage and then it would blow something up. It is just like the "ground" pin of the oscilloscope probe, which is ... grounded. When you attach the probe somewhere trying to take a measurement you are actually earthing one part of the DUT. And this may well blow something up.

In my case as I said above, I measured the AC voltage at the output using the TWO oscilloscope probes so I did not involve the earth. The AC voltage I saw was something horrendous, 100V or more peaks. I have not tried this under load, but my skin could be the load and it would certainly feel it.

So in summary my questions remain:

1) why do some (laptop) power supply units ground the "negative" and the SONY ones do not?

2) why do all laptop power supply units exhibit very large AC voltages at the output as measured with two scope channels in a differential way? The frequency at the output is 50Hz so I am not talking about HF switching spikes.

[han]

the laptop power chord is typically coaxial cable, using differential probe method is likely have a matching impedance problem

[IanB]

This following investigation and questions started when I realised I was getting mild/medium electric shocks from a laptop power adaptor.

I understand it is rather common to get mild shocks from low voltage power adapters. Apparently the reason is that the output is connected to the mains input via small capacitors (I think for noise suppression). These small capacitors couple the mains voltage through to the output terminals where you can feel it.

[gxti]

Cheap chargers reference the output to earth because it's a cheap way to reduce EMI, not because it's good design. If the first-party chargers have an earth pin but still don't reference the output to it, that's a pretty good sign that they're doing it for a reason.

The voltage you're seeing is coupled through a safety-rated Y1 capacitor which is used to reduce EMI. The current let through by the Y cap is certainly not enough to harm you, unpleasant though it may be. I'm not sure why even the earthed ones are showing a ghost voltage but it will disappear when you add a load to the supply. My intuition is that the Y cap should be connected between output negative and mains earth and thus not produce any ghost voltage, but if they haven't done so then again there's probably a good reason for it.

It's also possible that you're probing it wrong. I would expect that the output might, at worst, float at 1/2 the AC input voltage since there are Y caps to each input leg. Both would be in-phase because if the charger is not broken then the outputs will of course be exactly 12 or 20 or whatever volts DC apart, only the common-mode voltage can float. Subtracting one from the other will yield the desired DC voltage.

[pomonabill221]

Another test....
With a DMM, measure the continuity/resistance between either/both AC pins and either/both connections on the output jack.
You probably will not see anything, or a VERY high resistance on the order of 1meg plus.
Yes, it is a way to reduced EMI and the coupling through the safety cap is enough to give you a tingle and should disappear when a load is connected, but if you are using a laptop that is also floating you could get a little tingle.

I don't like this because this leakage current is passed to anything you connect to the laptop like a usb device that may be earthed!  Really piss poor way to reduce emi.

If you get a tingle, try reversing the AC plug if it isn't polarized, and see if the tingle goes away.
If it does, then either the supply's EMI cap is wired to the wrong side of the line, you are on a 220v circuit where both lines are hot to earth, or your AC plug in the wall is wired wrong.

Going back years with old tv sets, sometimes one side of the line was connected DIRECTLY to the chassis and there wasn't a power supply transformer.
The line was rectified, filtered and divided for the different voltages, and when you took the back off the service the set, you got a BIG REAL shock if the plug was in wrong... not to mention any test equipment that you connected the ground lead to the chassis was nailed!  I know as I found out on a RCA XL100 solid state CRT tv!  YEEEOU!

[amyk]

This behaviour is well-known: http://www.aplomb.nl/SMPS_leakage/Doc_ie.html

You should not be able to draw much current out of it, maybe a hundred uA at most.

[akis]

Just to summarise then, the reason why I am seeing AC voltages on ALL power adaptors' outputs I have tested is because of a capacitor placed between the output and the mains which is to try and reduce EMI. I will repeat some more tests tomorrow.

One more thing about mains: I was under the impression that the "neutral" lead was connected to the earth at the consumer unit (fuse box) at every house (can't go check now as everyone is sleeping in the house).

[gxti]

Yes, neutral is tied to earth. However many plugs will not be polarized so if it only has a single Y cap between the primary and secondary there's no guarantee whether it will be hot or neutral. So maybe try flipping the plug around. Of course the earthed ones ought to be connecting to earth not the lines, and even if not then US and UK earthed plugs are always polarized so it ought to be connected to neutral.

I still suspect that if you're seeing big voltages on the earthed chargers then there is probably something wrong with your test method and/or house wiring. The entire point of having that earth wire there is to have something to use as a shield and EMI sink, if they're not gonna use it then they would just use a non-earthed plug.

[IanB]

Yes, neutral is tied to earth. However many plugs will not be polarized...

In the USA this is true, but in the UK all plugs and sockets are three pin polarized. (Even if the earth pin is not used there is a plastic dummy pin present so the plug or wall wart can only be inserted one way round.) For any device that complies with appropriate regulations the live and neutral connections to the mains plug are reliably identifiable.

(Lest I get ambushed by pedants, two pin shaver plugs are non-polarized, but virtually nothing but electric shavers uses that two pin plug.)

[rolycat]

One more thing about mains: I was under the impression that the "neutral" lead was connected to the earth at the consumer unit (fuse box) at every house (can't go check now as everyone is sleeping in the house).

Not necessarily - until the 1970s houses in the UK were wired with separate earth and neutral connections back to the sub-station, and many still use this system.

[akis]

We are discussing two slightly different issues then.

1) Earthed negative in all cheap, replacement adaptors.

2) Blue "Y1" capacitor linking the primary and secondary sides of the transformer. Otherwise there is a complete separation of tracks and circuit under the transformer, with the secondary being on the output side of the PCB and the primary being on the mains side.

I have removed the earth in a few of these adaptors and thus made them floating, like the SONY ones. I have even connected some in series and it's fine.

I have no solution about the Y1 capacitor which seems to be responsible for the mild electric shocks.

[akis]

Here are two snapshots of the effect of that Y1 capacitor - the cheaper adaptor produces a much larger wave. I wonder what will happen if I simply remove that capacitor.

[Zbig]

Genuine Dell PSUs seem to have approximately 1k resistance between output's GND and mains plug's Earth. What could be the reason and consequences of that?

[akis]

Yes that is an excellent idea. Some PSUs I have seen have a wire link on the PCB between negative output and mains earth. One could replace the wire link with a resistor.

Except that depending on how the PSU is used, it could be floating at an unknown voltage level with respect to earth. In that case the 100R resistor would be seeing an unknown voltage across it, and there will be a lot of power spent on it, maybe burning the resistor or causing problems to the equipment that is providing this power. As an example say you take 2 laptop adaptors at 19.5V each and put them in series, you then have 19.5V over the 100R resistor, which is 195mA and almost 4 Watts. A 1K resistor would be better in this case. But the PSU will never be truly floating.

A question I asked above was "what happens if I remove the blue Y1 rated capacitor that links the two sides of the transformer" ? Well I did it and blew the PSU, it's not working anymore even with the capacitor back in. What is the exact purpose of this capacitor?

[akis]

I was asking in such a way because from some of the posts above it seems the purpose of this capacitor is well documented/understood.

[Neilm]

Caps across a SMPS transformer will provide a low impedance path for any high frequency noise. This will reduce the EMI it produces. Depending on the system, I have seen them also be used to snub voltages hence why removing them ended with a bang.