Is there significant noise in the neutral wire with a sense resistor connected?

[Pagedown]

I have a sense resistor connected in series with the neutral wire.

From the oscilloscope i detect noise of around 500mV - 1V across the sense resistor at high frequencies 60khz- 100khz. I see the noise once the power plug is inserted to the socket, although not switched on(live wire not connected). It seems to have quite an amount of current going through the neutral wire itself.


The source is a 240V 60Hz single phase.

Is this possible?

 

[Averagesupernova]

Schematic please.

 

[Pagedown]

This is schematic I detect lots of noise across the Rsense although the switch is still open.

 

[uart]

Try using both channels of your scope and measure the voltage on each side of the sense resistor wrt ground. You'll probably see the same thing on both channels.

 

[Averagesupernova]

Depending on the type of scope you are using you may be having issues with a slight voltage buildup between conduit ground (which the ground lead on your scope may be directly tied to) and the neutral wire. I see you have a 240 volt source with one wire referred to as neutral so this leads me to believe you are not in North America. I am not all that familiar with wiring practices outside the USA. Do you detect the noise on the secondary side of the transformer with your scope as well? I would check this with NOTHING hooked up to the secondary other than the scope. I would also try running this through an isolation transformer.

 

[Dmytry]

all sorts of hardware put ~100khz noise into mains... your pc's power supply for instance. Also. The noise your oscilloscope sees may not even be across the sense resistor. Each wire works as antenna.

 

[Pagedown]

Yup, I checked the isolation transformer with a known signal across it and the outputs are directly the same.

I dun understand there should be any voltage across the neutral wire as high as 0.5V as there isn't a complete loop across a wire isn't it?

Unless in respect to the oscilloscope ground?

 

[Averagesupernova]

Don't understand what you mean by checking the isolation transformer with a known signal across it.

 

[Jiggy-Ninja]

Just to be clear, you are referring to the 0.1 resistor labelled R_sense down near the op amp?

There's no way your getting 5-10A of noise (the amount of current needed to produce your scope reading). It has to be something interfering with your scope.

Check your channels and make sure they aren't accidentally set for x10 or x100.

 

[uart]

Yup, I checked the isolation transformer with a known signal across it and the outputs are directly the same.

I dun understand there should be any voltage across the neutral wire as high as 0.5V as there isn't a complete loop across a wire isn't it?

Unless in respect to the oscilloscope ground?

Check what I asked you to check in #4 above. You'll probably find the same noise on both channels with respect to ground.

BTW. Just to be specific, what I am questioning is just how good (or otherwise) is the high frequency common mode rejection ratio of your isolation amplifier.

 

[Pagedown]

Uart, I cannot connect the scope across the sense resistor itself because it will trip the breaker in the lab. I can only connect the scope across the 1:1 isolation transformer.

 

[uart]

Uart, I cannot connect the scope across the sense resistor itself because it will trip the breaker in the lab. I can only connect the scope across the 1:1 isolation transformer.

You can if you don't connect the ground clip of the probes to neutral, but instead connect them to the local ground wire. Note that I'm not asking you to try to directly measure the voltage across the sense resistor, I'm asking you to measure the voltage on either side of the sense resistor (or both sides simultaneously if you wish) with respect to earth.

 

[Averagesupernova]

This is starting to get back into what I was referring to in post #5. Power the WHOLE circuit with an isolation transformer. Replace the source you have hooked between the two fuses at the top left of the schematic with an isolation transformer. Make sure that no part of your circuit is connected to what we call 'conduit ground' here in the U.S.A. Then you can hook your scope anywhere you want. I think you will find that there is no current in the sense resistor.

 

[uart]

^ "This is starting to get back into what I was referring to in post #5."

With respect Averagesupernova, what you said in #5 was in essence exactly what I told OP in post #4. That he merely has a common mode noise issue on the neutral wire, something which he still hasn't performed my simple test for.

What will almost certainly turn out to be the case here is that OP's present implantation of isolation amplifier is not up to the job for the HF noise that we suspect is present (wrt ground) on his neutral wire. For example he may need an Earth shielded transformer, and/or to improve the degree of balance of his transform and possibly improve other aspects of his measurement circuits.

 

[Pagedown]

Okay, I will try that. Thus, the purpose of this is to measure the voltage from neutral to earth?

 

[uart]

Thus, the purpose of this is to measure the voltage from neutral to earth?

Are you serious? Read my posts I've told you several times already. Common mode signal. Problems with your measurement circuits common mode rejection.

 

[Pagedown]

Okay, if both signals are same which is the common mode signal, how do I solve it?

 

[uart]

Okay, if both signals are same which is the common mode signal, how do I solve it?

- Improve the balance of your measurements circuits front end.

- Redesign your isolation transformer for better HF isolation and possibly include an Earth shield.

The mistake that many people make is they get an amplifier with some huge CMRR figure like 90 dB or something and think that the problem is solved. But if the front end is imbalanced then some of the common mode signal can become a differential signal before it even reaches the amplifier (at which point the amplifiers CMRR is moot).

There may also be other "grounding" issues with your circuit that make it susceptible to noise. Grounding and noise rejection / noise impunity is a relatively complicated area that can have many different sources of potential problems but let's start with the isolation transformer. Understand that there is capacitive coupling across the windings and if you don't take precautions to minimize it (or at lest make sure it is balanced relative to the two ends of the sense voltage) then you dont have any meaningful HF isolation.

 

[Pagedown]

Can a low pass filter solve this?

 

[Averagesupernova]

Something I've just noticed. Your opamp input stage looks suspicious. Not sure what the power supply is for the opamp. There really is no DC feedback. It looks like it at first, but there isn't because whatever is on the output DC-wise will be the same on both inputs. Remember the old rule of thumb. No current leaves or enters an opamps input so there will be no DC dropped across either resistor. Where the output DC voltage ends up will depend on the input offset and open loop gain.
-
Have you determined if it is noise being coupled in through the transformer? What does the scope tell you with nothing else except the scope hooked to the transformer secondary?
-
Something else I would do is find a way to prevent a catastrophe in the even Rsense opens up.

 

[Pagedown]

The scope tells me the same noise at the transformer secondary. Thus, I have neglected the opamp gain stage at this point.

 

[Averagesupernova]

So with NOTHING hooked to the transformer secondary except the scope you have noise of how much?

 

[Pagedown]

The same amount of noise.

 

[Averagesupernova]

So it appears that the noise is not common mode at least in reference to the output of the transformer. I don't know anything about the physical layout of this setup. Is it on a couple of protoboards? Is it layed out on a PCB? What's the deal? What I would suspect now is that something on your board is oscillating. Disconnect anything that could be oscillating. Try again.

 

[Pagedown]

I have disconnected everything, only with neutral and the 0.1 resistor in series then to the isolation transformer.

At the secondary side of transformer, I connect only my scope, and the same amount of noise exists. It is clearly because of the isolation transformer now. But it doesn't make sense does it, as it is just a neutral wire and the transformer and no circuit loop. Is it interference?

 

[Averagesupernova]

So you have gotten it to the point that the neutral wire comes from the wall socket with no other wires and the .1 ohm is resistor hanging on the end of it with the primary of the transformer across the resistor? This sounds VERY suspicous to me. Are you sure you don't have a broken ground lead on your scope probe? Hook the scope across the secondary so it shows the noise and then short it out. You should never have a signal on a scope with the leads shorted.