DC adaptors - AC voltage between positive/negative and earth

[Sawkins]

I was testing a newly bought 12V DC adaptor yesterday and found a nice steady 12.1V between its positive and negative terminals.

For no particular reason I then tested the voltage between one terminal and earth and got an odd fluctuating reading. I switched the meter to measure AC and found 115V AC between the two. It was the same with the other terminal.

That can't be right surely? I don't believe this to be safe so I'm not going to use the adaptor.

I've done the same test since with a couple of other adaptors I've got round the house and I found 5V and 20V AC respectively between their terminals and earth.

 

[RF Lighting]

Are you testing for voltage with a high impedence digital meter??

 

[ricicle]

It is possible to see some AC component on the DC side as the DC output is derived from a switched or pulsed AC.

 

[Sawkins]

RF, yes I am.

ricicle, that's interesting, as the supply that shows 113V is a switched mode one.

 

[RF Lighting]

Try testing with a low impedence analogue meter or a test lamp.

The voltage will most likely be some sort of capacitive voltage that isn't really there.

 

[ericmark]

When working in Hong Kong I worked on an IT earth system. This is where earth is isolated from supply and we had some really odd readings. It was a case of forgetting about any measurement to earth.

With the small scale you are using although there is no direct connection there will be capacitive and inductive links. Since it will have diodes the test voltage would need to be low and to ensure no capacitive and inductive link measured with low voltage DC.

Simple matter of measuring between the input and output and there should be no DC link.

However switch mode power supplies do tend to have a small leakage between the input and output but it is well within the industrial limits. Measured with a AVO Mk8 you should see no more than 25 volt measured on the 25 volt scale at 20,000 ohms per volt.

However my digital meter is basically a valve volt meter and because it draw little or no current the voltage measured can be a lot higher. To measure with a valve volt meter or digital meter one needs to use a 1 Meg Ohm resistor across the leads to remove any capacitive and inductive linked voltages.

One would expect any voltage across a 1 meg ohm resistor to be less than 12 volt. Although in theory it can be higher. The transformer should have a shield with two linked circles which show it is an isolation transformer.

 

[Sawkins]

RF, have done with a 2K ohm/volt analogue meter. I didn't get consistent results - 50V in the 500 and 250V ranges, 10V in the 50V range. That means your suggestion is right doesn't it?

Having said that, it feels like it's there - if I hold the meter lead instead of earthing it I get a tingle off the point, like with one of the old neon mains testers.

 

[TicklyT]

What happens when the output is loaded with whatever you intend the PSU to drive? Testing with the output unloaded will almost always give spurious results, even with a relatively low impedance meter.
If the PSU is to supply, say, 12V DC, 0.1A, the load across the output would be about 120 Ω, several orders of magnitude lower than suggested so far.

 

[bernardgreen]

When there is no earth connection to the PSU (ie it has a two core lead, a plastic earth pin or metal pin with no connection then the "ambient" voltage inside the unit is the mid-point of the mains supply, namely 115 volts above ground.

Any leads coming out of the unit will be capacitively coupled to this "ambient" voltage as will any equipment supplied by the unit where there is no earthing to the equipment.

Most PSUs are designed to have a very low capacitive coupling so the energy at 115 volts that can be taken out of the unit is too small to be harmful to anyone or any thing. But it can still be felt as a tingle.

 

[Sawkins]

That's it exactly. Today I connected the PSU to the device I bought it for, and it worked fine, but the device did indeed have this same apparent 115V between its chassis and earth, and between its leads and earth.

Can I do anything about this? I don't like getting a nip off the device phono sockets when I happen to touch them, even if it isn't dangerous as such.

 

[Monkeh]

Earth the chassis?

 

[Gees]

Earth the chassis?

I Agree.

 

[Monkeh]

There's a reason that tag is there.