advice on exact measurement with an oscilloscope

[arivel]

Hi everyone .
I recently owned a Rigol 1202z-e oscilloscope.
Unfortunately I don't have full mastery of using all the functions and menus of this tool.
I need to take a measurement but I want to be reasonably sure that the mV value is accurate.
the voltage I want to measure is that of a current sensor, it is a variant of the Rogowski coil.
these are 7 toroidal supports in plastic material wrapped by a coil of enamelled wire. all the coils are connected in series to form a larger one. each toroid is crossed by a conductor through which the current flows. the photograph is attached.
the electrical connection diagram between instruments and sensor is attached.
every time I make a measurement I get significantly different voltage values, I don't want to make mistakes and therefore I ask you what the oscilloscope settings should be.
I set the ch2 x1 probe and the ch1 x10 probe to average mode. the current flows into an 8 ohm rheostat and ammeter. I attach a photograph of the last measurement, I ran a current of 0.5 A.
Hello and thanks .

[tggzzz]

Why not just use a voltmeter?

You need to verify there's no interference from anything else. What does the scope show if there is no current?

Consider using a differential measurement, i.e. Ch1-Ch2.

What's the minimum voltage expected across the coil?

Don't forget that if the voltages have a tolerance of X%, then the tolerance of the differential voltage will not be 10%. Don't forget the consequences of an 8-bit DAC!

[arivel]

since the voltage coming out of the sensor is very small I have to separate its signal from that caused by the mains supply especially if the frequency is the same.
Is a voltmeter suitable for this? Is it suitable for measuring a few mV?
What do you mean with the following words? "What's the minimum voltage expected across the coil?"

[tggzzz]

since the voltage coming out of the sensor is very small I have to separate its signal from that caused by the mains supply especially if the frequency is the same.
Is a voltmeter suitable for this? Is it suitable for measuring a few mV?
What do you mean with the following words? "What's the minimum voltage expected across the coil?"

We do not know what voltages you need to measure. You need to define the minimum and maximum voltages, the frequency, and source impedance.

You need to keep noise (i.e. non-signals) away from the instrument's inputs, since those will affect the measurement.

If the signal frequency is the same as the mains frequency noise, then how do you expect any instrument to distinguish signal from noise?

[EC8010]

Judging by your oscilloscope traces, you have quite reasonable voltages that should be easily measurable to the accuracy of your oscilloscope. First thing to remember is that oscilloscope probes always add attenuation errors. A better way to make your measurements would be to fit some BNC connectors to a metal biscuit tin, connect them to your oscilloscope via short cables instead of the probes, then take wires from the BNCs to your probed points. Solder a lead between tin and its lid, then put the lid on the tin and make your measurements. Doing this, you will have eliminated probe errors, reduced induced interference, and applied a larger more easily measured signal to your oscilloscope.

I have a couple of breadboards with neodymium magnets glued on their back that live inside biscuit tins with BNCs for exactly this purpose. Even with the best efforts, you will always find it difficult to get an accurate measurement at mains frequency. It's best not to try.

[arivel]

since the voltage coming out of the sensor is very small I have to separate its signal from that caused by the mains supply especially if the frequency is the same.
Is a voltmeter suitable for this? Is it suitable for measuring a few mV?
What do you mean with the following words? "What's the minimum voltage expected across the coil?"

We do not know what voltages you need to measure. You need to define the minimum and maximum voltages, the frequency, and source impedance.

You need to keep noise (i.e. non-signals) away from the instrument's inputs, since those will affect the measurement.

If the signal frequency is the same as the mains frequency noise, then how do you expect any instrument to distinguish signal from noise?

it is not child's play to calculate exactly what comes out of the sensor, however it is certainly a matter of a few mV or lower order of micros.
these toroids also function as antennas and receive electromagnetic disturbances. you could enclose the entire sensor in a sheet of mu-metal connected to the ground but this would be a later step.
the frequencies range from 20Hz to 2000 Hz

[RoGeorge]

every time I make a measurement I get significantly different voltage values

Always give numerical examples.  What does significantly different means?  Could be 10% or could be 10 times more. 

My guess is the measurements differ because the wire + the 7 segments of coil are changing their position.  Try moving the wire and/or toroids while measuring a constant current.  If the voltage varies, then make the sensing coil rigid with the wire.

Another reason for the coil voltage to change is the frequency of the signal.  For the same current, the higher the frequency, the more induced voltage in the coil.

[tggzzz]

since the voltage coming out of the sensor is very small I have to separate its signal from that caused by the mains supply especially if the frequency is the same.
Is a voltmeter suitable for this? Is it suitable for measuring a few mV?
What do you mean with the following words? "What's the minimum voltage expected across the coil?"

We do not know what voltages you need to measure. You need to define the minimum and maximum voltages, the frequency, and source impedance.

You need to keep noise (i.e. non-signals) away from the instrument's inputs, since those will affect the measurement.

If the signal frequency is the same as the mains frequency noise, then how do you expect any instrument to distinguish signal from noise?

it is not child's play to calculate exactly what comes out of the sensor, however it is certainly a matter of a few mV or lower order of micros.
these toroids also function as antennas and receive electromagnetic disturbances. you could enclose the entire sensor in a sheet of mu-metal connected to the ground but this would be a later step.
the frequencies range from 20Hz to 2000 Hz

Those levels and frequencies conflict with the scope trace you have shown.

Your should remove noise coupled in from other sources, so that you are measuring signal, not noise. Alternatively define the function/operation a measuring instrument uses to distinguish noise from signal.

A low pass filter will remove much noise, but a scope has a 20MHz bandwidth which will "introduce" noise.

Mumetal will reduce magnetic coupling, but I suspect there will be many sources of electric coupling.

Diagnosing and removing noise sources is difficult to do remotely, since the implementation details are critical.

[nctnico]

The frequency range is certainly within range for using a true-RMS multimeter with a decent frequency range (check the datasheet as many multimeters are accurate for 50Hz only though). Bench multimeters typically have a wider frequency range but you don't have to spend a lot as a 4.5 digit multimeter will be good enough.

Also, I'm not sure a Rogowski coil can be used directly, it needs an integrating amplifier to get a meaningful result / signal from it. For the frequency range, a simple current transformer is likely a more suitable option. A current transformer with a suitable load resistor produces a far more predictable output compared to a plain Rogowski coil.

[arivel]

I don't think the wire suffers vibrations, this happens in transformers.
in fact almost all of the many measurements made are at 50 Hz which in my country corresponds to the mains power supply.
I called it a current sensor because its function is to generate a voltage proportional to the current passing through the conductor but in reality its intended use is different. I don't want to talk about this, I'm interested in getting a reliable measurement. Yes, I will definitely use an opamp integrator.

[RoGeorge]

Not coil vibration. Different position of the discs relative to the wire, and relative to each other.  For example, I expect two tilted rings of coils (relative to the wire) to measure different than 2 perfectly perpendicular and centered coils on the wire.  Because of this, and because the 7 discs are close to each other, I also expect a curved/undulated wire (with perfectly perpendicular discs) to measure differently then when the wire is shaped like a straight line.

Again, ideal-coil's voltage is NOT proportional with the current, but with the speed of the variation of the current.  Which means the 7 discs are not a sensor for current alone, but a sensor for current AND frequency.  If you measure, for example 1V for 1A at 50Hz,  same coil will measure 2V for 1A when the frequency is 100Hz.  This is for an ideal coil, without resonance (which resonance can easily happen in practice, but we will put this aside for now).

If you want something proportional with the current alone, then your sensor has to be a resistor.  You already have an 8ohm rheostat.  The voltage on the rheostat will indicate the current alone (does not depend on frequency).

Again, you didn't say the numbers.  How big are those variations you want to fix?  What do you expect vs what values you measure in practice?  Sorry, can't help you any more by just blindly guessing.  Good luck with the experiment. 

[Overspeed]

Hello

As frequency is low
As voltage ''should be low ''

Why not use a data acquisition board in 16 bits ? whatever that a NI with labview or a ADC board with an Arduino ?
16 bits will deliver highest ''precision'' than a 8 bits oscilloscope , only limitation can be the triggers option and perhaps the need to amplify the signal

Regards
OS

[jonpaul]

very poorly designed Rogowski coils.
https://en.wikipedia.org/wiki/Rogowski_coil
please send URL of mfg, spec.

Suggest that you Read Tektronix "Oscilloscope Probe Circuits" and "Oscilloscpe Probe Measurements"
https://w140.com/tekwiki/images/6/62/062-1146-00.pdf
https://w140.com/tekwiki/images/1/19/062-1120-00.pdf
In fact the entire series  (1980s) is still relivant and worth study.

j

J

[nctnico]

Again, ideal-coil's voltage is NOT proportional with the current, but with the speed of the variation of the current.  Which means the 7 discs are not a sensor for current alone, but a sensor for current AND frequency.  If you measure, for example 1V for 1A at 50Hz,  same coil will measure 2V for 1A when the frequency is 100Hz.  This is for an ideal coil, without resonance (which resonance can easily happen in practice, but we will put this aside for now).

Agreed. A Rogowski coil is more suitable to measure the amount of electric charge (Coulomb) going through a wire rather than measuring the current. In the past I've used Rogowski coils to measure the amount of charge going into a weld.

[arivel]

as I have already specified, it is not the subject of this topic whether this sensor is suitable for measuring current and the fact that the response increases with frequency was known to me.

[arivel]

other measurements are attached, respectively 20,30,40,50,60 Hz  and 0,5 A

[tggzzz]

Technique 1: display c2-ch1, and do your measurements on that.
Technique 2: use a voltmeter connected both ends of the coil. You will have to verify that the meter has sufficient frequency response, and doesn't null out 50Hz.
Technique 3: build a circuit using an instrumentation amplifier with the inputs connected to the coil. That will convert the "differential input" to a "single-ended output".

[nctnico]

Technique 1: display c2-ch1, and do your measurements on that.
Technique 2: use a voltmeter connected both ends of the coil. You will have to verify that the meter has sufficient frequency response, and doesn't null out 50Hz.
Technique 3: build a circuit using an instrumentation amplifier with the inputs connected to the coil. That will convert the "differential input" to a "single-ended output".

I don't see the need to do a differential measurement. In the schematic it shows all signals are ground referenced.

All in all the best way is to compare the voltage across the rheostat (or a current measurement resistor AKA current shunt) with the signal coming from the Rogowski coil. IF the goal is to verify correct measurements coming from the Rogowski coil.

[tggzzz]

Technique 1: display c2-ch1, and do your measurements on that.
Technique 2: use a voltmeter connected both ends of the coil. You will have to verify that the meter has sufficient frequency response, and doesn't null out 50Hz.
Technique 3: build a circuit using an instrumentation amplifier with the inputs connected to the coil. That will convert the "differential input" to a "single-ended output".

I don't see the need to do a differential measurement. In the schematic it shows all signals are ground referenced.

All in all the best way is to compare the voltage across the rheostat (or a current measurement resistor AKA current shunt) with the signal coming from the Rogowski coil. IF the goal is to verify correct measurements coming from the Rogowski coil.

I am guessing the OP wants to measure the voltage across the coil. If not, then what?

But I'm tired of gradually extracting information from the OP.

[nctnico]

Technique 1: display c2-ch1, and do your measurements on that.
Technique 2: use a voltmeter connected both ends of the coil. You will have to verify that the meter has sufficient frequency response, and doesn't null out 50Hz.
Technique 3: build a circuit using an instrumentation amplifier with the inputs connected to the coil. That will convert the "differential input" to a "single-ended output".

I don't see the need to do a differential measurement. In the schematic it shows all signals are ground referenced.

All in all the best way is to compare the voltage across the rheostat (or a current measurement resistor AKA current shunt) with the signal coming from the Rogowski coil. IF the goal is to verify correct measurements coming from the Rogowski coil.

I am guessing the OP wants to measure the voltage across the coil. If not, then what?

If you look at the schematic, you'll see the coil is connected to (circuit) ground at 1 point so measuring between ground and the other end will give the voltage across the coil.

But you are right that any further suggestions will need a much more thourough explaination from the OP about what the OP wants to achieve precisely.

[tggzzz]

Technique 1: display c2-ch1, and do your measurements on that.
Technique 2: use a voltmeter connected both ends of the coil. You will have to verify that the meter has sufficient frequency response, and doesn't null out 50Hz.
Technique 3: build a circuit using an instrumentation amplifier with the inputs connected to the coil. That will convert the "differential input" to a "single-ended output".

I don't see the need to do a differential measurement. In the schematic it shows all signals are ground referenced.

All in all the best way is to compare the voltage across the rheostat (or a current measurement resistor AKA current shunt) with the signal coming from the Rogowski coil. IF the goal is to verify correct measurements coming from the Rogowski coil.

I am guessing the OP wants to measure the voltage across the coil. If not, then what?

If you look at the schematic, you'll see the coil is connected to (circuit) ground at 1 point so measuring between ground and the other end will give the voltage across the coil.

But you are right that any further suggestions will need a much more thourough explaination from the OP about what the OP wants to achieve precisely.

Yeah, you are right.

Strange schematic (especially w.r.t. trace thickness), which I didn't bother to look at sufficiently carefully.

[arivel]

Are the words I wrote unclear?
I don't know how to be clearer than that.
Let's see if I can do it this time. I want to measure what voltage comes out of the coil when I pass a 0.5A current using a range of frequencies. since the signal that comes out of the coil is very small I have to find a way to prevent the noise from overlapping and making reading impossible, especially when I use 50 Hz because this is the mains frequency. To do this I was inspired by the following link: https://www.eevblog.com/forum/projects/oscilloscope-with-trace-averaging-as-a-lock-in-amplifier-(rigol-ds1054z)/
the diagram I attached comes from this link.
is it clearer now?
do you suggest leaving the connection of the ch1 probe floating?.
what other explanation do you want tggzzz to have?

[nfmax]

You could perhaps use a ‘proper’ lock-in amplifier, possibly with a differential preamp. Assuming you have a (frequency &phase) reference source available. Old analog units are usually to be found lurking on eBay at a reasonable price

[tggzzz]

How are you planning on distinguishing the signal (which apparently can be 50,100,150,...Hz) from mains noise of 50Hz+-1% plus harmonics at 100,150,200,250...Hz?

There are many techniques including simple filtering (analogue or digital). Alternatively, if you can control the stimulus you may be able to use that knowledge to correlate the measured response with the stimulus. Such correlation can be equivalent to narrowband adaptive filters. Classic techniques include spread spectrum modulation of the stimulus, or lock-in amplifiers.

It is best to avoid such EMI/EMC from entering the circuit in the first place. There are books on this tricky subject, and all of them provide many techniques which might or might not be relevant in a specific case.

You need to ensure that your measuring instrument does intrinsically have too much noise and also does have sufficient resolution. Both may be a problem with a 100MHz 8-bit scope.

If you haven't already, look in The Art of Electronics, since the authors' job is to support exacting research requirements.

The other avenue is to determine the techniques other people have used with such coils, and reuse their circuits.

[arivel]

thanks everyone for the replies