Using an oscilloscope to measure mains power

[Zensei]

Hi all,

I hope this is in the right forum and has not been asked and answered before. I could not find anything, at least.


I want to use a 2 channel handheld oscilloscope to measure mains power, with one channel for the voltage and the second channel for the current using a current clamp. The math functions on these instruments usually only allows +,-,* and /, so I guess it depends on the probe settings which factor I have to multiply the result with (1mV/A or 10 mV/A or whatever the settings are). I would expect that the scope calculation takes the phase shift into account, right?

I guess that my question is regarding capabilities of the typical oscilloscope menu system.

If I want to capture the data over a period and calculate the total amount of produced power in the logging period, can I use the math function and have that value saved with the channel data?

So if I have the math function operating to produce a third graph, will the oscilloscope save this data along with the channel data in a .csv file?

And if not, how would I go about compensating for the phase shift when calculation the power afterwards?

Maybe what I'm thinking of here is not even possible. What do you think?

[Nerull]

The software features of your oscilloscope are going to depend on exactly which oscilloscope you are using.

[Zensei]

Most likely Tekway (Not handheld), Hantek 1062s or Siglent SHS 1000/800.

But I was hoping that it might be a general feature, that someone else might have tried or knew was/is possible. 

[papo]

If you have a scope that is able to deal with the current clamp and calculate the power properly, then go for it. If you have doubts, I'd suggest the following procedure:

1. Measure voltage on one channel.
2. Measure current on other channel.
3. Export both to csv files or whatever (make sure your scope can do that).
4. Import this into the program of your choice (Excel etc.), resulting in two columns.
5. Correct the current column to be expressed in amps. (Make sure the voltage column is in volts).
6. Multiply current and voltage row-wise.

This gives you instantaneous power in Watts. You can find the energy by integration, but make sure to integrate over one or multiple, full periods. Then divide by the time to get average current. I would not bother about the phase shift as this assumes that your signals are sinusoidal.

Also please make sure that your scope can safely handle mains voltages. You could also consider going for one of these power-meters. Just ensure that these don't get confused by non-resistive loads as this seems to be your concern.

Regards

Matt

[BravoV]

I'd suggest the following procedure:

1. Measure voltage on one channel.
2. Measure current on other channel.

Just be very-very careful there, when using two probes as such way at mains connected lines, especially using a normal probe and a non-isolated scope.

You might kill your scope, or worst injure your self.

[papo]

Just be very-very careful there, when using two probes as such way at mains connected lines, especially using a normal probe and a non-isolated scope.

Yeah I second that. Maybe my warning was not sufficient. Re-reading the original post, maybe the goal is some sort of long time measurement. In that case the scope is probably not able to store the data in a suitable form, or at least you'd have to assume sinusoidal signals... Are you sure a Kill-a-watt-sort of device is not a better alternative?

Regards
Matt

[Zensei]

I was planning on measuring voltage on one channel and current on the other.

About the isolation, the Hantek 1062S is isolated and Siglent has an isolated frontend for their scopes, so I can use a non-isolated scope for mains measurements.

I'm thinking of using the "specrum" analyzer or FFT to scout out harmonic distortion on the mains. Is that a crazy idea?

[oldway]

Why don't you use a low power low voltage (10/1 for example) isolation transformer ?
That's far more secure.

[PA4TIM]

Use diff probes speced for mains use or use an isolation transformer or make an isolation amplifier.
Always look at the specs of the probes you use. It is not wise to use low budget multimeters without the right safety features and even less wise to use a budget scope and probe for mains measurements.

[Zensei]

Dawgie:

The isolated fromtend is an external piece of equipment: http://www.siglent.eu/accessories/isolated-front-end.html

It should solve the isolation problem. A transformer is also a solution, but less practical.

Differential probes are so expensive that they cost more than the scope itself

[Zensei]

I am all ears...

My idea is to use the scope as a harmonics/power quality analyzer. It's a lot cheaper than a real harmonics analyzer...

[PA4TIM]

That isolator can do 600Vpp. We have 230V here. That is 650Vpp.  They say it can be used for 220V.  That is 622Vpp. So I think I do something wrong in the math.

[tinhead]

That isolator can do 600Vpp. We have 230V here. That is 650Vpp.  They say it can be used for 220V.  That is 622Vpp. So I think I do something wrong in the math.

I guess the scope probes you use with it could be set to x10 mode, and so scaling the mains voltage down by a factor of 10 before being fed into the isolator.

no, this is not the module or chinese but an idiot from siglent.eu (which is not siglent itself but an distributor!)

He seems to be not smart enought to read the manual, that isolator is +600Vp to -600Vp, so 1200Vpp (and not 600V p+p!!!)

[Zensei]

Tinhead:

So at 230 VAC level, at 325 Vp it should be fine. Even at 400 Vrms, 650 Vp it should be ok, correct?

Dawgie:

I already bought a Hantek 65 A current clamp to measure the current. On top of that I have a 200 A current clamp as well.

[MarkL]

I was planning on measuring voltage on one channel and current on the other.

[...]

I'm thinking of using the "specrum" analyzer or FFT to scout out harmonic distortion on the mains. Is that a crazy idea?

To measure power, I would do the multiply on the scope to create the power waveform as you've previously suggested.  And also configure an FFT on the current waveform (or whichever waveform interests you for harmonic distortion).

Some scopes will also allow you to compute an average value on a resulting math waveform.  If you can configure it to do an average on the power waveform, it will give you the real power drawn by the load.  You should set it up to compute the average over an integral number of cycles to get the most accurate value.  Depending on the scope's math capabilities, other values could probably also be derived, such as power factor.

You can then move all these waveforms and computed values to a computer for logging and any further number crunching.  Presumably there would be a loop on the computer that would single shot the scope and then get the data.

The overall timing and phase relationship between voltage and current is crucial in power calculations and they need to be captured simultaneously for accurate results.  And depending on the frequency of the harmonics you're interested in, you may also need to deskew your voltage and current probing paths to reduce error.  The typical way to do this is with a test pulse where you can align the edges in the voltage and current waveform display using the scope's "probe skew" adjustment.

As for being self-contained like a power quality logger, I don't know of a scope that will perform long-term logging by itself.  But that doesn't mean it doesn't exist... maybe one of those windows based ones...

[denelec]

Mains voltage can contain a lot of voltage spikes.
These spikes can be harmful to your scope. 
My Rigol is only rated CAT I 300V RMS. You at least need CAT II to measure mains voltage safely.
http://www.ni.com/white-paper/5019/en/

So unless your scope is rated CAT II, you should use some form of isolation.
I don't know what protection a 10:1 probe can offer.

[Zensei]

My plan is to use a Siglent SHS806 non isolated scope aith a Siglent isolation frontend to measure the voltage. I suppose I could use a high voltage x100 probe and set the right settings in the scope settings, if the scope supports that setting.

A section from the data sheet:

Features & Benefits
Dual - input,combine oscilloscope, Multimeter and recorder (including TrendPlot and waveform Recorder) in one unit
Input voltage: input voltage through BNC is up to CATII 300V and CATIII 150V
Standard probe: 10X CAT II 400V
Optional probe: 10X CATII 1000V and 10X CAT III 600V
Oscilloscope and multimeter safety grade is
up to CATII 600V and CATIII 300V

[Terabyte2007]

You might be able to use a variac/auto transformer with isolation to lower the mains to a safer level. Then setup a known ratio for your calculations. Not sure how this would affect your measurements but it's a thought.

[mainman]

My plan is to use a Siglent SHS806 non isolated scope aith a Siglent isolation frontend to measure the voltage. I suppose I could use a high voltage x100 probe and set the right settings in the scope settings, if the scope supports that setting.

A section from the data sheet:

Features & Benefits
Dual - input,combine oscilloscope, Multimeter and recorder (including TrendPlot and waveform Recorder) in one unit
Input voltage: input voltage through BNC is up to CATII 300V and CATIII 150V
Standard probe: 10X CAT II 400V
Optional probe: 10X CATII 1000V and 10X CAT III 600V
Oscilloscope and multimeter safety grade is
up to CATII 600V and CATIII 300V

What country are you in?

I am in  the U.S  and as you know I have the shs810 which is essentially what you just bought except with a slightly bigger band. The residential voltage here is 120V rms, the DSO is rated at CatII 300V rms. Hooking up my scope I should get a 340 pk-pk wave.

Your scope supports up to x1000 probe.

BTW how much did that isolation frontend module cost you, I  imagine less than the cost of you just buying the siglent model with the built in isolation.

What made you not buy the hantek?

[moepower]

I would make sure you have a good understanding of the different types of probes, isolation, grounded inputs vs. isolated inputs before even thinking about probing the mains.   

For instance, you should not use differential probes with isolated scope inputs for probing the mains.  This is because the BNC connections and chassis could float to dangerous levels.  Diff probes should be used with grounded input scopes. 

Improper use of isolation transformers can also create dangerous floating voltages.  Be careful...

For a good explanatory guide for those interested, google and download Probe_Fundamentals_Tektronix.pdf

[Zensei]

moepower: Thanks for the advice, it's on my computer Will read through it later.

mainman:

I'm in Denmark, we have 230 V mains. That's 325 Vp, 650 Vpp.
I haven't completely decided in the Siglent yet, but I think that I will end up buying that anyway.
The extra memory, the nicer design, the trent capture feature, the option with the isolation frontend and the price is what has me interested.

I think the Hantek looks toyish, and I don't really ned 200 Mhz to measure mains

The isolated frontend cost 105 euro.