Measuring Leakage Currents with IEC 60601 defined Measurement Device (MD)

J

jfdzar

Hi there,

I have been making some research about how to measure the leakage currents in the 60601-1 with the measuring device (MD) described in 8.7.4.4 or Figure 12, unfortunately I did not find yet a reliable answer.

Usually I have been measuring them with a SecuTest SIII+

We went for pre-compliance testing to the test laboratory and the measurements done with the circuit described in figure 12 were all somehow different.

I build the device described in Figure 12 and now I am measuring with a multimeter Fluke 83 V which input impedance is 10 MOhm (> 1MOhm) and 100 pF (< 150 pF) which is in range.
I did not find until know the frequency response, where could I check that?

Or should I measure with an oscilloscope?
The oscilloscope that I have has a common ground that's why I am not confident to attach it to the MD in case it can produce a short cut

Or do you recommend an special multimeter/oscilloscope in order to measure it?

The second think is how should I read the value of the current. I am reading voltage (DC or AC depending on the part I will like to measure) then I am dividing this voltage through 1000 Ohm in order to get the current V/R=I is it that right?

I have read the norm thoroughly but I did not find any clear answers. It would be great if you could give me some advice

Thank you very much in advance. And thank you for the forum it has helped me a lot in the last years!
 

EMengineer

Involved In Discussions
Re: Measuring Leakage Currents with IEC 60601 defined Measuring Mevice (MD)

Hi,

I can't put pics in post yet, so I recommend you to google "rigel medical practical guide to 601", there are measuring circuits of leakage current

Also if you are measuring patient leakage of B, BF or CF applied parts, keep in mind there are different measuring points.

And about current calculation you are right, you need to devide voltage through 1000 Ohm
 
J

jfdzar

Re: Measuring Leakage Currents with IEC 60601 defined Measuring Mevice (MD)

Hey EMengineer!

Thank you very much for your fast response. I knew the document from Rigel, I also read it. The images of the document are taken almost 1:1 out from the IEC 60601-1 norm.

But I am still not confident of the measuring device and the voltage measuring instrument.

- Is the frequency characteristic for the measuring device as a hole specified or only for the voltage measuring instrument.
- How can I check that my voltage measuring instrument has this frequency characteristic?
- Is it right to measure the AC and DC parts with the function of the voltmeter?​

Thank you very much for your responses
 

Peter Selvey

Leader
Super Moderator
Re: Measuring Leakage Currents with IEC 60601 defined Measuring Mevice (MD)

Typical low cost DMMs like the Fluke 87V are only accurate up to around 1kHz, beyond that they degrade pretty quickly. There are some meters out there with up to 100kHz but you need to check the specs carefully.

A scope can let you know if there are any significant higher frequency components. But scopes are often not calibrated at higher frequency and there are a bunch of complex issues to worry about, not only grounding. So, scopes are OK to let you roughly know what is there, but consider any reading as being ±20% and note that strictly any measurements are not legally traceable to national standards.

Another issue is that RC components themselves are probably not OK at higher frequencies. My experience with a home made jig found it started to depart from expected response at around 300kHz. This is likely to be due to stray capacitance and inductance, for example through hole metal film resistors are actually a spiral so they make a wonderful inductor.

That said, if the main source of leakage is mains frequency, high frequency components e.g. due to switching supplies are not going to have any impact. First of all, the standard filters out these components, proportional to frequency in kHz. For example, a SMPS running at 60kHz will have the 60kHz components in the leakage reduced by 1/60.

Secondly, in the rms world, the main item is the main item. The rms calculation is √(V1² + V2²), where V1 and V2 are the different frequency components. This formula works such that if V1>>V2, the influence of V2 is negligible.

For example, if you have earth leakage of 400µA @ 60Hz and 2000µA @ 60kHz, you might be surprised to find the final result is 401.4µA. The influence of the 60kHz part is not even 1%, even though it is 5 times larger.

Here is the math:

First, the 2000µA is filtered and reduced to 33.3µA (=2000/60).

Then the rms calculation is:
√(400²+33²) = 401.4µA

And let's assume we have up to 50% error in the 60kHz, so the true value could be anywhere from 1000 - 3000µA. The result changes by less than 0.5% from 400.3 to 403.1µA. <1% variation for a 50% change in the 60kHz component.

So, at the end of the day it is not really worth worrying about. At least in the real world - but a CBTL auditor might still give you pain!
 

QuickSpace

Starting to get Involved
Hello everybody,

Can anyone explain me how to find frequency characteristics of measuring device, like how to choose frequencies and get corresponding magnitude, is the frequencies choosed for the graph is constant or varies? Capture.PNG
 

Benjamin Weber

Trusted Information Resource
Hello everybody,

Can anyone explain me how to find frequency characteristics of measuring device, like how to choose frequencies and get corresponding magnitude, is the frequencies choosed for the graph is constant or varies? View attachment 26058
The frequency response in figure 12 b) relates to the low-pass filter in 12 a), i.e. the MD only. This is not the required frequency response of the voltage measuring instrument! The "b)" at the voltage measuring instrument does not point to figure b) but to the note "b)" below figure 12! It says: "Resistance ≥ 1 MΩ and capacitance ≤ 150 pF".

You have to be careful when choosing the voltmeter. The standard says, AC limits apply to currents having a frequency not less than 0.1Hz, cl. 8.7.3 b). This means that everything above 0.1Hz shall be considered AC, everything below shall be considered DC. When you perform leakage current measurements you have to distinguish between AC and DC components for some leakage current types, e.g. patien leakage current. You could just set the multimeter to AC to measure the AC component and DC for the DC component. But the multimeter has a certain internal cut-off frequency that you can look up in the multimeter specifications.
For the Fluke 87V I cannot really find this value, but the accuracy specifications start only at a frequency of 45Hz. This indicates, that in AC mode only frequencies higher than 45Hz are really measured. Lower frequencies are damped.

This might lead to different results, when different voltmeters are used.

Using a scope is s good idea but grounding might be a killer. Best idea is to use a battery driven scope, thas is isolated from the supply circuit of the measurement setup. But a scope has the same issue with the cut-off frequency when switching to AC coupling (we're using a scope with cut-off frequency of appr. 8Hz). Most scopes have a DC coupling mode, that includes the AC component. So you really see the whole signal. In order to do the standard compliant distinction of AC and DC component, you can either do post-processing with the required cut-off frequency of 0.1Hz. Or the scope has internal signal analysis functions (FFT...) to analyze the frequency components.

Some more point to consider regarding the supply network:

It important to isolate the measurement supply network from the actual mains supply by an isolating transformer. And you need a variable AC-source, because the standard requires to use 110% of the rated maximum supply voltage and the maximum rated supply frequency. IF the isolating transformer is after the AC-source (i.e. the AC source feeds the isolating transformer), you also have to measure the actual output of the isolating transformer separately, because of diffences between the input and the putput voltage of transformers, whoch usually are load dependent.

And another thing is important. You might need to have a second voltage availabe (for touch current with SIP/SOPs, F-type applied parts and if there are metal accessible parts that are not protectively earthed), where you can at least change the polarity independently from the supply voltage of the device under test! This voltage can be derived from the supply voltage, so you don't need a second variable source, but polarity must be changed independently.
 

QuickSpace

Starting to get Involved
The frequency response in figure 12 b) relates to the low-pass filter in 12 a), i.e. the MD only. This is not the required frequency response of the voltage measuring instrument! The "b)" at the voltage measuring instrument does not point to figure b) but to the note "b)" below figure 12! It says: "Resistance ≥ 1 MΩ and capacitance ≤ 150 pF".

You have to be careful when choosing the voltmeter. The standard says, AC limits apply to currents having a frequency not less than 0.1Hz, cl. 8.7.3 b). This means that everything above 0.1Hz shall be considered AC, everything below shall be considered DC. When you perform leakage current measurements you have to distinguish between AC and DC components for some leakage current types, e.g. patien leakage current. You could just set the multimeter to AC to measure the AC component and DC for the DC component. But the multimeter has a certain internal cut-off frequency that you can look up in the multimeter specifications.
For the Fluke 87V I cannot really find this value, but the accuracy specifications start only at a frequency of 45Hz. This indicates, that in AC mode only frequencies higher than 45Hz are really measured. Lower frequencies are damped.

This might lead to different results, when different voltmeters are used.

Using a scope is s good idea but grounding might be a killer. Best idea is to use a battery driven scope, thas is isolated from the supply circuit of the measurement setup. But a scope has the same issue with the cut-off frequency when switching to AC coupling (we're using a scope with cut-off frequency of appr. 8Hz). Most scopes have a DC coupling mode, that includes the AC component. So you really see the whole signal. In order to do the standard compliant distinction of AC and DC component, you can either do post-processing with the required cut-off frequency of 0.1Hz. Or the scope has internal signal analysis functions (FFT...) to analyze the frequency components.

Some more point to consider regarding the supply network:

It important to isolate the measurement supply network from the actual mains supply by an isolating transformer. And you need a variable AC-source, because the standard requires to use 110% of the rated maximum supply voltage and the maximum rated supply frequency. IF the isolating transformer is after the AC-source (i.e. the AC source feeds the isolating transformer), you also have to measure the actual output of the isolating transformer separately, because of diffences between the input and the putput voltage of transformers, whoch usually are load dependent.

And another thing is important. You might need to have a second voltage availabe (for touch current with SIP/SOPs, F-type applied parts and if there are metal accessible parts that are not protectively earthed), where you can at least change the polarity independently from the supply voltage of the device under test! This voltage can be derived from the supply voltage, so you don't need a second variable source, but polarity must be changed independently.



Thank you, But my exact question is how to find Frequency characteristics of MEASURING DEVICE(MD), i know it is frequency characteristics of low pass filter, I need the exact procedure to find the same, like how to choose frequencies, how to find corresponding gain and plot the same in the graph. Is the frequency in the x-axis is constant all situations or it changes?? any calibration procedures??
 

Benjamin Weber

Trusted Information Resource
OK, I see. You want to check, wether your selfmade MD matches the required frequency response?

We use the follwing calibration specification for our MD:

Using 6V inoput voltage, at the following frequencies the corresponfing output voltage shall be measured:
20Hz => 6V,
50Hz => 6V,
60Hz => 6V,
1KHz => 4.37V,
10kHz => 633mV,
100kHz => 63.7mV,
1MHz => 6.37mV

Regarding calibration you will need to give tolerances to the expected values, we use 3.5% to 60Hz, 5% to 1kHz, 10% above 1kHz

Additional we perform a DC impedance measurement: Expected value 1kOhm witg a tolerance of 1%.

Does this help?
 

QuickSpace

Starting to get Involved
OK, I see. You want to check, wether your selfmade MD matches the required frequency response?

We use the follwing calibration specification for our MD:

Using 6V inoput voltage, at the following frequencies the corresponfing output voltage shall be measured:
20Hz => 6V,
50Hz => 6V,
60Hz => 6V,
1KHz => 4.37V,
10kHz => 633mV,
100kHz => 63.7mV,
1MHz => 6.37mV

Regarding calibration you will need to give tolerances to the expected values, we use 3.5% to 60Hz, 5% to 1kHz, 10% above 1kHz

Additional we perform a DC impedance measurement: Expected value 1kOhm witg a tolerance of 1%.

Does this help?
Extreme thanks to you, One more question, Regarding tolerances to calibration, on what standard basis you are taking these in consideration (If any)?
 
Top Bottom