Is this a medical application?
Canadian Standards Association has a standard for patient leakage currents.
Since obtaining that standard I have lived in three different places in Canada and three different places in Central America.
I have no idea where my outdated copy is now.
The thresholds for an operating theater are exceptionally low.
It has been many years since I was tasked with designing the electrical for an operating theater but a few of the methods in use then were:
OR room critical circuits were typically fed from a small isolating transformer with a grounded shield between the primary and secondary windings.
Circuits were kept as short as possible to avoid current paths due to conductor capacitance to ground.
Without some specific information there is not a lot more help that we can give you.
What is the device?
Are you measuring current between an isolated ground and true ground?
Please post a single line diagram of the system from the panel to the point of utilization.
Is this for an operating room, a critical care area, a general use device?

Bill
--------------------
Ohm's law
Not just a good idea;
It's the LAW!

Hi Bill,
I have attached a schematics describing the test.
This is a medical device which is emitting a 1 MHZ RF signal via a handpiece.
According to the standard, the leakage current to earth must not exceed 50mA whilst I receive 65mA.

I have tried already to place a Toroid on the GND line inside the device for 1 Mhz signal and it improved it a bit....and also tried to feed the internal circuity directly with an external laboratory power supply which should be greatly isolated....but I cannot detect where the leakage comes from...
Even if I disconnect completley the gnd line to the device result stays the same, how can it be? what closes the circuit to enable this current flowing?

Any idea is appreciative.

• https://files.engineering.com/getfile.aspx?folder=6f78cb28-0c55-4ada-b64f-a

You need to explain how the test is conducted

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert! https://www.youtube.com/watch?v=BKorP55Aqvg
FAQ731-376: Eng-Tips.com Forum Policies forum1529: Translation Assistance for Engineers Entire Forum list http://www.eng-tips.com/forumlist.cfm

Hi
The diagram was attached.
I am emitting an RF signal, approx 100 V rms on a 200 ohm load, in sequence to the load another 200 ohm resistor is connected to the wall gnd, and then I measure on that resistor the voltage to see how much current is flowing to gnd.

> I tried to disconnect the gnd from device and retest and surprisingly it didn't change ..any idea?

Tough to imagine what's going on. Some thoughts fwiw:

Ordinarily we don't expect current flowing in an open circuit (*) so investigate the path for that apparent current flow...

I would try successively lifting leads closer to the measurement on each side until the ground current indication is eliminated in attempt to isolate where the current is coming from.

* High frequency current can flow through capacitance associated with leads or components that are physically close to a grounded component or to a cable shield.

Also if resistor (or connections) in parallel with the voltmeter goes open, then it's easier to induce voltages in the voltmeter circuit (no low impedance path to bleed those voltages off). That would mean no current flow in the resistor but still an induced voltage seen at the voltmeter. Maybe try a clampon measurement of current directly through the resistor.

Another potential troubleshooting step if there are multiple potential sources of intereference (line frequency and radio frequency), use an oscilloscope at the voltmeter to figure out which frequency is causing the measurement.

Also check your voltmeter to make sure it's reading zero when everything is off (sorry if it sounds silly, just trying to cover the bases).


=====================================
(2B)+(2B)' ?

Thanks
Actually I am using a floating o scope measuring ac vrms on the resistor going to gnd.so I believe the measurement is ok.but I cant grasp from where this current is coming from...

Well I'd say you have a puzzle and a mistaken assumption somewhere about what is going on. You have to be systematic and start exploring things with measurements to try to understand it. Some possibilities mentioned:

Measure current through the resistor with clamp-on and see if it matches what you're inferring from the voltage measurement. Make sure the clamp-on does not include a potential return path for current such as a cable shield. I'm picturing you have legs of resistor itself accessible, that's where you'd most like to clamp on. (Actually if you measure a non-zero current while clamping around conductor and shield that could be valuable information, but if you measure zero current while clamping around conductor and shield that doesn't tell you much).

Lift leads at selected points in the circuit (both sides of resistor) to see when the indication disappears. Work you way all the way to resistor itself. Assuming the indication goes away at some point before the resistor is totally isolated form the rest of the circuit then you've got a clue about the path.

Check the frequency of associated with the voltage measurement if not already done (it's not clear to me).

New idea - try an alternate voltmeter.





=====================================
(2B)+(2B)' ?

The answer to your question is probably hidden somewhere in here:

Bill
--------------------
Ohm's law
Not just a good idea;
It's the LAW!

Disregard my previous posts (I'm going to go back and line through them). I made a mistake in thinking you had lifted the ground at the measurement resistor and were still seeing current (which would be really bizarre). Bill's post made me realize you obviously lifted the ground on the DUT. So... Never mind

I assume you have a ground referenced voltage on your power input as a normal condition.

You may have coupling from the input power to the thing you're monitoring for ground current. Either direct connection through a sneaky path, or capacitive coupling.

If there are two or more different frequencies present in your DUT then you might still benefit from checking the frequency of this ground leakage current with your scope to narrow things down.

=====================================
(2B)+(2B)' ?