There have been several discussions on the forum about the problems with noise from earth loops—particularly with USB connections.
This is an opposite problem!.... 

The MOX uses the PA 150A AC power adaptor as its external DC power supply. This ‘wall wart’ type power unit has no ground connection, and relies on the the internal circuitry to isolate it from the household AC supply. This configuration allows for a small amount of ‘leakage’ current to flow to earth.  Not enough to hurt you of course—although you will get a ‘tingle’ if you happen to be holding an output cable and an earthed chassis at the same time!

Measuring between any of the MOX ‘ground’ connections (output connector screens etc) and household supply earth with a Digital Multi-meter gives reading of about 100VAC on the meter (here in the UK .. probably less in the US).  Fitting a 100K resistor in series with the earth and a MOX output ground reduces that to about 30V, suggesting that a leakage current of some 3mA is flowing.(Edit: should be 0.3mA—correction in post #5 below)

I checked with a second PA 150A, to confirm it was not a random reading from one unit..... It also seems to be typical of other ground free ‘wall warts’ I have tried.

Where a problem can arise is when connecting an output from - say - a grounded mixer to the A/D input, with the MOX already switched on.  I recently connected a signal from the headphone output of my Behringer mixer to the MOX A/D input, and the result was a blown headphone amplifier in the mixer. (The MOX was fine!)

So if, like me, you use your MOX normally with just headphones connected, remember that the ‘ground’ side of the external connections will ‘sit’ at about 100VAC. If that touches something like a mixer headphone output, it can destroy it.

I would suggest that its quite important to make sure that all connections are made before you turn everything on.... That’s best practice anyway of course.... silly of me not to follow it!

And if you only have headphones connected, and you can touch any metalwork on the phones which is connected to the ‘common’ connection, remember not to touch any ‘real’ nearby earth point (grounded chassis, etc). You will get a ‘tingle’!....

For my own peace of mind, I have connected a permanent mains earth connection to my MOX now. It introduces a (very) small amount of hum, but no more unexpected ‘tingles’....and hopefully no more blown headphone amps!

rogs - 14 August 2014 11:58 AM

[...] Measuring between any of the MOX ‘ground’ connections (output connector screens etc) and household supply earth with a Digital Multi-meter gives reading of about 100VAC on the meter (here in the UK .. probably less in the US).  Fitting s 10K resistor in series with the earth and a MOX output ground reduces that to about 30V, suggesting that a leakage current of some 3mA is flowing.[...]

If your meter can measure AC current, instead of a calculation based on voltage drop across a resistance, what do you measure directly (meter connected from MOX “ground” to earth ground)?

5pinDIN - 14 August 2014 09:42 PM

rogs - 14 August 2014 11:58 AM

[...] Measuring between any of the MOX ‘ground’ connections (output connector screens etc) and household supply earth with a Digital Multi-meter gives reading of about 100VAC on the meter (here in the UK .. probably less in the US).  Fitting s 10K resistor in series with the earth and a MOX output ground reduces that to about 30V, suggesting that a leakage current of some 3mA is flowing.[...]

If your meter can measure AC current, instead of a calculation based on voltage drop across a resistance, what do you measure directly (meter connected from MOX “ground” to earth ground)?

Using the AC current range in the configuration you describe gives a reading which fluctuates between 2.8mA and 2.9mA, which is what I would expect, based on my observations above.
Were you expecting that measurement to give a different result?…

rogs - 15 August 2014 04:23 AM

5pinDIN - 14 August 2014 09:42 PM

rogs - 14 August 2014 11:58 AM

[...] Measuring between any of the MOX ‘ground’ connections (output connector screens etc) and household supply earth with a Digital Multi-meter gives reading of about 100VAC on the meter (here in the UK .. probably less in the US).  Fitting s 10K resistor in series with the earth and a MOX output ground reduces that to about 30V, suggesting that a leakage current of some 3mA is flowing.[...]

If your meter can measure AC current, instead of a calculation based on voltage drop across a resistance, what do you measure directly (meter connected from MOX “ground” to earth ground)?

Using the AC current range in the configuration you describe gives a reading which fluctuates between 2.8mA and 2.9mA, which is what I would expect, based on my observations above.
Were you expecting that measurement to give a different result?…

Among other factors, a resistor limits current. Yes, I was expecting a somewhat higher reading. Here’s why…

You reported that the leakage measured about 100VAC using a digital multimeter. Most DVMs have an input impedance of 1 or 10 Megohms, and therefore shouldn’t materially influence that measurement. When you measured the voltage across a 10kohm resistor, you reported 30VAC, and that it indicated a 3mA current. I agree that with the resistor in place (assuming reasonably close tolerance of the resistor’s value), the calculation is correct. However, if the “open circuit” voltage is 100V, then 70V was being dropped across the leakage impedance. With 3mA causing a 70V drop, the leakage impedance would appear to be about 23kohms. Since a 10k resistor is almost 50% of the apparent leakage impedance, it should substantially affect the measurement.

Based on what you previously reported, I would have expected a direct measurement of the leakage current (without the 10k resistance) to have been closer to 4.3mA. Perhaps the meter’s AC current range is inaccurate, or the 10k resistor is off value, etc.

Under any circumstance, even 3mA is a surprisingly high leakage current measurement. I don’t own a MOX or MOXF, and don’t have a PA150(A) supply, but leakage currents I’ve measured on most ungrounded “wall wart” supplies typically are not more than 1/10 of what you are measuring, and many are closer to 1/100 of it. You might want to address your findings with Yamaha Support. In the meantime, not connecting to or disconnecting from the MOX while it’s powered does seem prudent.

By the way, as opposed to switch-mode power supplies, linear ones I’ve measured only have leakage of around 1-2VAC, and correspondingly low leakage current.

5pinDIN - 15 August 2014 09:37 AM

Under any circumstance, even 3mA is a surprisingly high leakage current measurement.

As you say, very high..... which is why I felt it best to check my calculations.....

And, embarrassingly, I need to advise some corrections.
Most importantly, my original post suggested a 10K load resistor.  That should have read 100K.  Which of course makes the current calculations derived from the voltage across the resistor 10 times too high, and explains why the current and voltage readings were not very different, as you advised they should be. 
My sincere apologies for that error. 

Having been a bit annoyed at the damage to my mixer, I made the measurements immediately, to the rig as connected.  That included 2 x PA 150As, plus two extra non Yamaha ‘wall warts’ used for other kit. 
So although my voltage readings were correct, my current ones for each PA150A were clearly not. I was actually measuring all the leakage currents in parallel!

In addition, I originally used a meter which I now know is less than accurate!

So....Re-doing my measurements with a calibrated meter (although the actual calibration certificate is a bit out of date!) and measuring each power supply individually shows much better results for the Yamaha PA 150As.
Using a 100k 1% resistor again, the AC current readings are pretty close to the measured voltage readings, as the much lower loading would suggest they should be.

And the leakage for each of my two PA150s show a leakage current of 157uA for one and 175uA for the other. Much more like the figures you have suggested they would be.

Of the other two supplies, one is a 5V USB ‘charger’ type supply, which has even lower leakage.

The final supply is a much older 24V unit, and the figures for that are significantly worse, and I suspect this is the main culprit. That device needs to be replaced.

So, as you suggested, the actual leakage currents are much lower than I originally posted. Again my apologies for the misleading information.

Having said that, there is still enough ‘leakage’ to cause damage in the scenario as I described, and you will still feel a minor electric shock if you happen to be in contact with the either power supply output connection and a ‘real’ earth.
So my (and your) suggestion to not ‘hot swap’ even audio connections to and from the MOX still apply....

And I’m still leaving a permanent earth connection, even if it does hum a bit!

Meanwhile, I have a damaged mixer to go and fix.....

rogs - 15 August 2014 11:09 AM

[...] So, as you suggested, the actual leakage currents are much lower than I originally posted. Again my apologies for the misleading information.

Having said that, there is still enough ‘leakage’ to cause damage in the scenario as I described, and you will still feel a minor electric shock if you happen to be in contact with the either power supply output connection and a ‘real’ earth.
So my (and your) suggestion to not ‘hot swap’ even audio connections to and from the MOX still apply....

And I’m still leaving a permanent earth connection, even if it does hum a bit!

Meanwhile, I have a damaged mixer to go and fix.....

Thanks for the additional information and clarification.

Although making/breaking connections with equipment unpowered is a good practice, a well-designed output stage should be protected from transient over-voltage conditions. (That includes static discharge as well as the tip of a phone plug momentarily touching a jack shield connection.) Behringer equipment is often a good value, but sometimes their designs leave something to be desired. I could discuss design measures to protect the mixer’s headphone output stage against externally applied over-voltage, but this isn’t the place for that.

5pinDIN - 17 August 2014 09:35 AM

Although making/breaking connections with equipment unpowered is a good practice, a well-designed output stage should be protected from transient over-voltage conditions. (That includes static discharge as well as the tip of a phone plug momentarily touching a jack shield connection.) Behringer equipment is often a good value, but sometimes their designs leave something to be desired.

Finding Behringer schematics online is not easy, but I have located one of a model from the same series as mine, and that shows that I have been unlucky in my choice of connection.

Most of their outputs are capacitively coupled, with both series output resistors and ground shunt resistors fitted. I made the mistake of using the headphone outputs (the others were already in use, or I wouldn’t normally have bothered!)
In order to keep costs down, Behringer have used the fact that the mixer uses a floating ‘split’ supply to allow them to omit the output capacitors on the headphone outputs.
In addition, they use a standard op-amp as a headphone amp (presumably again to keep costs down?) and only fit 22R series resistors to the outputs.

Shunting leakage current to ground through 22R, in series with the op-amp output?… explains why I now have no headphone amp!

Hopefully though my observations may help prevent anyone in a similar position from damaging their kit.... and I did find the ‘tingle’ shock when touching a MOX connector ‘ground’ unpleasant… so my permanent earth connection helps there.....

...although I’m guessing that ‘chasing’ USB ground loops in the future maybe a whole new ball game now!!....

Thanks for your guidance on this thread, and for querying my current figures.... I should have realised that 3mA was way too high to be considered ‘safe’, and checked before I posted.  Embarrassing!.....

rogs - 17 August 2014 10:28 AM

[...] they use a standard op-amp as a headphone amp (presumably again to keep costs down?) and only fit 22R series resistors to the outputs.

Shunting leakage current to ground through 22R, in series with the op-amp output?… explains why I now have no headphone amp![...]

From one perspective, the less series resistance from the op-amp output to the headphones, the better - it minimizes frequency response variations caused by headphone impedance irregularities, even with relatively low impedance headphones. For that reason, Behringer’s choice of 22 ohms isn’t unreasonable, but of course (as you pointed out) it isn’t very protective. Some sort of voltage-clamping arrangement might have prevented the output stage damage.

You experienced an unfortunate circumstance, but it’s good that you posted about it so that others are aware of the possibility. Don’t be concerned about the initial error concerning the leakage current - it doesn’t detract from the value of having warned about the potential problem. (Yes, my choice of the word “potential” was intentional. :-) )