Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Is there a practical way to tell whether a noise is CM or DM type when using a scope?

Status
Not open for further replies.

doncarlosalbatros

Full Member level 2
Joined
Nov 26, 2016
Messages
143
Helped
1
Reputation
2
Reaction score
1
Trophy points
18
Activity points
1,758
Hello,

This has been a trouble for me for years.

Sometimes I observe several types of noise on signals or power supply outputs such as spikes, ringing, hum ect.

But I cannot tell whether what I see it is due to common mode interference or differential mode.

Is there a quick way or measurement method to check the nature of the noise in such cases? (Like cutting the earth link from the scope ect.)

Can you give an example of measurement way so one can conclude the noise he sees is CM or DM?

Thanks in advance,
 

Differential mode must have two conductors. Use a common mode choke on the differential pairs (for example wind the differential pairs around a ferrite core for several turns to make a simple common mode choke), then watch if the noise is reduced or not.
 

Use a two channel scope. Measure both signals with respect to ground.
Each signal, or both by adding, them will show common mode noise.
Inverting one signal and adding them will show the differential noise.

Common mode is trace A -or- trace B -or- the result of adding them.
Differential mode is the sum of trace A minus trace B, in other words the difference between them.

Brian.
 

Use a two channel scope. Measure both signals with respect to ground.
Each signal, or both by adding, them will show common mode noise.
Inverting one signal and adding them will show the differential noise.

Common mode is trace A -or- trace B -or- the result of adding them.
Differential mode is the sum of trace A minus trace B, in other words the difference between them.

Brian.

Hello,

Is that the measurement setup you suggest? I tried to draw it. Will the probe leads be floating?

Untitled.png
 

Probably, but 'common mode' refers to a signal on both the + and - wires to the device so it has to be measured against some other reference. If the reference is ground you can connect the probe ground clips to ground but do not connect the clips to the other supply wire.

Brian.
 

To elaborate what Brian said about the common mode measurement:

Both probe's ground clips must be attached to the exact same grounding point.
Both ground clips must be of equal length. And both probes must be the same brand and model.
Adjust the compensation adjustment of both probes prior to making the measurement.
 

Try watching the signals as an XY display or lissajous plot.
When in phase (common mode) the signals form a diagonal line leaning in one direction.
When out of phase (differential mode) the line leans in the opposite direction.
I suppose it is difficult to spot noise if both it and the signals are common-mode. Therefore this method is able to mislead, although it's a useful tool.
 

I dont quite understand where will the ground clips of the probes be connected in my figure?

Untitled (1).png

The test device(powers upply) has only two terminals positive and negative as usual.
Will the ground leads of the scope probes connect to the negative terminals of the test device?
If not where will they be connected in real?
Please could you show me via a diagram?
 

Hi,

if your power supply has only two terminals (like a battery) then it can´t generate commom mode noise.

Common mode noise also needs a reference to which you refer the voltage.
Form your description I assume this is an AC/DC or a DC/DC supply.
If so, then you can reference it to:
* EARTH_GND
* maybe even to the metal (shielding) case. Don´t know if this makes sense.
* also possible: one input voltage signal - mainly for DC/DC converters.
If AC/DC then take care about eventually dangerous voltages. Don´t create short circuit with the connections..

Maybe read through this: https://micro.rohm.com/en/techweb/knowledge/emc/s-emc/01-s-emc/6899

Klaus
 

Hi,

if your power supply has only two terminals (like a battery) then it can´t generate commom mode noise.

Common mode noise also needs a reference to which you refer the voltage.
Form your description I assume this is an AC/DC or a DC/DC supply.
If so, then you can reference it to:
* EARTH_GND
* maybe even to the metal (shielding) case. Don´t know if this makes sense.
* also possible: one input voltage signal - mainly for DC/DC converters.
If AC/DC then take care about eventually dangerous voltages. Don´t create short circuit with the connections..

Maybe read through this: https://micro.rohm.com/en/techweb/knowledge/emc/s-emc/01-s-emc/6899

Klaus

Are you sure?

The SMPS adapter I have has only two terminals and passing through 50Hz common mode noise due to its leakage capacitance. I verified this with voltmeter between + terminal and earth and - terminal and earth. Both gave like 100VAC reading.
Common mode noise can capacitively couple to the power supply.
That is my case and Im trying to see that earth common mode noise on scope.

I need to find a method when I see a noise on a signal how can I experimentally verify what I see on scope is a CM or DM noise?

Please can you show any setup I cannot find any info.

Sometimes the signal have higher freq. noise superimposed in 50Hz CM, I need to extract that high freq. CM noise but lacking a good technique.
 
Last edited:

You are missing the point Doncarlosalbatros. Common mode means the signal is common (the same) on both wires so if you want to quantify it you need some other reference to measure against. Your set-up suggests you have a supply with floating outputs and your reference is Earth. With nothing to tie the output to any fixed potential it is quite normal for an SMPS output to float at any voltage between zero and peak line voltage, including possibly spikes on the line.

However, what you are measuring is not 'common mode noise' as your title says but leakage across the isolation barrier between input and output. At the very least you need to give it a leakage path to conduct some of the output side down to a lower potential. I would suggest using two resistors of equal value between + to Earth and - to Earth. If you use about 47K it should be low enough to drop the residual voltage but high enough that any noise is unaffected.

Brian.
 

Hi,

The SMPS adapter I have has only two terminals
Unlikely.
An SMPS usually has at least two input terminals and two output terminals. thus 4 in total.
Then there may be additionally: EARTH_GND, shielding, remote control, other voltages....

We do not know how your SMPS looks like until you show us. All the SMPS I/Os.

Klaus
 

Hi,


Unlikely.
An SMPS usually has at least two input terminals and two output terminals. thus 4 in total.
Then there may be additionally: EARTH_GND, shielding, remote control, other voltages....

We do not know how your SMPS looks like until you show us. All the SMPS I/Os.

Klaus

Im talking about output DC terminals + and - Of course it has mains input mine doesnt have earth. It has two input AC terminals.
The CM is capacitevely coupling as I said.
 

Hi,

Im talking about output DC terminals + and - Of course it has mains input mine doesnt have earth. It has two input AC terminals.
The CM is capacitevely coupling as I said.

Yes - capacitively coupling - to the AC side.
This is the only way in your case for common mode noise on the DC side .. referenced to the AC side.

Thus if you want to measure CM noise you need to tke the AC side into account. Otherwise it can´t be CM noise.

Did you go through the link I sent in post#9? It clearly tells about CM noise.

Klaus
 

Hi,



Yes - capacitively coupling - to the AC side.
This is the only way in your case for common mode noise on the DC side .. referenced to the AC side.

Thus if you want to measure CM noise you need to tke the AC side into account. Otherwise it can´t be CM noise.

Did you go through the link I sent in post#9? It clearly tells about CM noise.

Klaus

Yes I read that document. Here is my final setup to see the CM voltage on scope:

Untitled (1) (1).png

As you see I left the ground lead of the probe floating. If I earth it nothing changes so i left it floating.
That is the only way I could find.

I want to learn how common mode noise is measured for any given signal transmission.
In my case it is very specific type of CM and Im not sure this setup works for anyother.

As I said Im lookign for a method or maybe an equipment where I can easily see the CM noise.

I found something called current probe but I tried to make myself it didn't show the 50oz common mode interference,
Here is an example from a website´:

electrical-electronic-systems-current-probes-7-266-g003.png

I think those current probes are for very high frequency like 10Meg to 100MegHz
 
Last edited:

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top