Short Term Logging of Mains AC Current

[David_AVD]

Is there an off-the-shelf solution for measuring AC current (mains potential) and logging (graphing) it with reasonable detail over a short period of time?

What's required is to measure the inrush current (in the 10A to 50A range at a guess) of conventional transformers and possibly switch-mode power supplies.

I would expect the measurement period to only a few seconds for each capture.  Perhaps a current transformer and a DSO that can export is all that's required?

[David Hess]

I would expect the measurement period to only a few seconds for each capture.  Perhaps a current transformer and a DSO that can export is all that's required?

This is how I would do it and have done it in the past.  Some multimeters have logging and export capability into the 1000s of measurements per second or higher and would be more suitable.

[denelec]

I bought the "Watt'up Pro" a few years ago  It works well.
Ideal for measuring and logging the consumption of an equipment over a period of time.
https://www.wattsupmeters.com/secure/products.php?pn=0&wai=376&spec=4

There is a 250V version with Australian, UK and European plugs.
The precision seems to be good for a $130 device.
You can log voltage, current, power, duty-cycle, power factor, etc.
You can export the data in Excel.

[denelec]

But if you want to measure inrush current, this won't do.
The fastest it can go is 1 sample per second.

A work we use Hioki Power Analyzers but they cost thousands of $.

I think a Digital Storage Oscilloscope to measure the current  and voltage would be the best and simplest solution.

[David_AVD]

Yeah, I'd need to see the current with much finer resolution than a logger would do.  The thread title doesn't convey that well.

[David_AVD]

I'm looking at getting a Rigol MSO-2072A as the logic channels may be useful to capture analog events.

It seems to have reasonable reviews, but I couldn't find one from Dave?

[David Hess]

I'm looking at getting a Rigol MSO-2072A as the logic channels may be useful to capture analog events.

It seems to have reasonable reviews, but I couldn't find one from Dave?

Look for a review of a DSO in the same series.  Within one series they are identical except for options and bandwidth.

A DSO can certainly do what you want when combined with a current transformer.

[simingx]

I'm looking at getting a Rigol MSO-2072A as the logic channels may be useful to capture analog events.

It seems to have reasonable reviews, but I couldn't find one from Dave?

Look for a review of a DSO in the same series.  Within one series they are identical except for options and bandwidth.

A DSO can certainly do what you want when combined with a current transformer.

If you don't need a DSO($$) + current clamp($$$), then one of these may be cheaper: http://www.fluke.com/fluke/m3en/clamp-meters/Fluke-374.htm?PID=70402

It has an inrush current mode to capture the peaks when they occur.

[tautech]

Yeah, I'd need to see the current with much finer resolution than a logger would do.  The thread title doesn't convey that well.

How fine resolution???
Inrush to what?
Expected average?
There are several easy solutions just depends on precisely what you need.

[David_AVD]

I'd like to see the current draw on a cycle by cycle basis over a few (maybe 10 max) seconds.

By finer resolution I was meaning not sampling as infrequently as a logger would (not better than 1Hz ?).

The steady current (after the inrush) is around 10 Amps max.

[David_AVD]

I'm looking at getting a Rigol MSO-2072A as the logic channels may be useful to capture analog events.

It seems to have reasonable reviews, but I couldn't find one from Dave?

Look for a review of a DSO in the same series.  Within one series they are identical except for options and bandwidth.

A DSO can certainly do what you want when combined with a current transformer.

Thanks for that.  I'll check out the DSO2xxx series for reviews.  I don't currently own a DSO.

I figured I may as well get the DSO version as the logic capability would be useful for other projects.

[tautech]

I'd use a scope current probe and record with a DSO because I have both.
As have been mentioned a CT will also do the job, and often can be acquired from old PCB's in your parts pile.
You could also get what you need with a current shunt and a differential probe setup to a DSO.
If your present gear serves you well enough maybe just a 2 CH USB scope for this job?

[David_AVD]

Any recommendations on a CT that is suitable for direct connection to the DSO / MSO input ?

Something that can measure up to 100A would be nice.

[RedOctobyr]

I'd use a scope current probe and record with a DSO because I have both.
As have been mentioned a CT will also do the job, and often can be acquired from old PCB's in your parts pile.
You could also get what you need with a current shunt and a differential probe setup to a DSO.
If your present gear serves you well enough maybe just a 2 CH USB scope for this job?

I'm new at this stuff, so apologies if I'm way off-base. But I recently ordered an inexpensive USB scope, and one of the things I have in mind is measuring some brief startup current spikes for small DC motors.

Some Googling brought me to current shunts (shunt resistors), which sound suitable for my needs. From what I found on eBay, they come in a range of current capacities, say 3A to 1000A, and seem to have a max voltage drop of 75mV, regardless of the amperage rating.

So you'll need to be able to accurately measure voltage differences which are quite small. My little USB scope may not have enough voltage resolution when looking at signals (or signal differences) which max out at 75mV. From what I gather, you can use 2 normal probes, just with one on either side of the shunt. I paid about $4 for a 20A shunt on eBay. The 100A versions were maybe $15 or something, but still not that expensive.

I don't know whether shunts lose accuracy if you briefly push them past their rating. Use a 10A/75mV shunt to measure a 20A startup load, and do you simply (and accurately) measure a 150mV drop, instead of 75mV? Or do things get all non-linear and inaccurate? If you measure 150mV, that could be a crude way to provide a larger signal to measure, for use with less-sensitive instruments.

If my shunts could also be used for AC measurements of startup currents, that's good to know.

[jahonen]

I measured inrush current of an amplifier I built some years ago with Agilent MSO6034A and Fluke 80i-110s current probe:



Worked just fine. I suppose that you'll be fine using Rigol DSO for this purpose.

Regards,
Janne

[David Hess]

I'm looking at getting a Rigol MSO-2072A as the logic channels may be useful to capture analog events.

It seems to have reasonable reviews, but I couldn't find one from Dave?

Look for a review of a DSO in the same series.  Within one series they are identical except for options and bandwidth.

A DSO can certainly do what you want when combined with a current transformer.

Thanks for that.  I'll check out the DSO2xxx series for reviews.  I don't currently own a DSO.

I figured I may as well get the DSO version as the logic capability would be useful for other projects.

The DS2000A series is massive overkill for your current logging application but inexpensive even so.  A DS1000Z series unit would still be overkill but cost less and they have 4 channels and an optional logic analyzer as well.

[David_AVD]

The DS2000A series is massive overkill for your current logging application but inexpensive even so.  A DS1000Z series unit would still be overkill but cost less and they have 4 channels and an optional logic analyzer as well.

Yeah, but I will need something with those specs and functionality for other jobs in the pipeline so may as well buy it now and buy it once. 

I found the Fluke i200S switchable 20A/200A probe which seems like a good fit.  Now to find it at a reasonable price.

[David_AVD]

Finding a current probe at a reasonable price (considering the limited use it will get) is proving difficult.

I did see this CT from Ocean Controls which I guess could be adapted  to work with the DSO/MSO ?

[os40la]

http://www.theenergydetective.com/5000c

I bought this one for my house. Works great. 1watt resolution at 1sec. I may have missed it but you didn't say exactly what resolution you were looking for.

They even sell the current probes separately for $47 for a set of 2
http://www.theenergydetective.com/set-of-current-transformers-cts.html

[David_AVD]

Resolution required is cycle by cycle.

[eneuro]

The steady current (after the inrush) is around 10 Amps max.
...
Something that can measure up to 100A would be nice.
...
Resolution required is cycle by cycle.

How long those 100A peak pulses are, while it is at 230VAC 23kW 

In my spot welder project I used for 230VAC mains monitoring HALL effect based current sensor which at 5 turns gives me +/-25Amax range with sensor response about 100mV/1A  @ 5V sensor supply voltage with 0A output at 2.5V.

Which is accuracy you want to measure those high current peaks close to 100A  +/-0.1A or better ?

Cycle by cycle @ 50Hz is 20ms-maybe a few moving average (http://en.wikipedia.org/wiki/Moving_average) I used in my current sensor could do the trick depending on how long are those pulses you want to detect?

I decided custom design it with my own source code for small DIP8 Attiny85 for post processing measurement data, while I have full control how incoming data is taken and this can do exactly what I want depending on project requirements with a few changes in C source code between projects, while main code and functions are ready and tested.

[David_AVD]

OK, "cycle by cycle" was bad wording on my part.  What I meant was "I want to see the actual current waveform" as opposed to getting some type of slow logging.

If I could end up with a measurement rig that allowed for 100A peak reading but resolution down in the 100mA area that would be great.

I have no desire to make something to capture / display the data.  I'll simply buy the DSO/MSO and get on with it. 

Does anyone know any reason the $40 CT I linked to wouldn't be suitable?

[David Hess]

If I could end up with a measurement rig that allowed for 100A peak reading but resolution down in the 100mA area that would be great.

Do you mean measuring peaks to 100A with 100mA resolution or being able to accurately measure peaks to 100A while also delivering useful data at 100mA full scale?  The later is much more difficult.

100A by 0.1A is 1 part in 1000 which is going to require either an oscilloscope in high resolution mode (most do these days) or a datalogger which would could typically sample 1000+ times per second but still deliver 12+ bit accuracy.

Does anyone know any reason the $40 CT I linked to wouldn't be suitable?

It is designed to work into a resistive load of less than 10 ohms which sets the burden voltage so the maximum output at 120 amps will be 400 millivolts which is plenty for an oscilloscope and should allow a full scale range down to 1.5 amps or lower.

There are also inexpensive current clamps designed to interface directly with an oscilloscope.

[eliocor]

I suggest you LEM products: they work flawlessly, have a very wide bandwidth and are not too expensive.
Very easy to connect to a digital scope.
Other than the devices shown at the link, they produce much more current monitoring devices... take a look at their production: http://www.lem.com/hq/en/content/view/13/86/

[eneuro]

I suggest you LEM products: they work flawlessly, have a very wide bandwidth and are not too expensive.

http://www.gmw.com/electric_current/LEM/hall-transducers.html

LEM high accuracy Hall effect based Current Transducers enable an electrically isolated measurement of the total instantaneous electric current in conductors through the Transducer aperture by effectively integrating the magnetic field produced around the conductors.

Unfortunatelly you can not tweak LEM's internals, while when I used 2$ SS495A Hall sensor and $1.5 16Mhz ATTiny85 with its 10bit ADC it is quite easy do specyfic tasks in software and with ADC set to 250kHz close to 20000 current samples per second, which combined with moving average is fine in many applications 
It has advantage we know how those integrations are made since we have MPU's C source code and can monitor for example AC mains and triger external events when ADC conversion takes place, so we know exact timings and its internal processing program flow from analysis of assembler instructions generated in AVR binary.

For example this 25A LEM for about $22
http://www.digikey.com/product-detail/en/LTS%2025-NP/398-1002-ND/409815
has 25mV/A sensitivity @ 5V power supply, so it means that only 0.625V difference for 25A current 
http://www.lem.com/docs/products/lts%2025-np.pdf
Even with my SS495A and about 100mV/A I have close to +/-2.5V voltage difference at the same 5V power supply in th ecase of 25A current...

Nope, I do not like this LEM while yep according to his datasheet linked above:

VOUT Output voltage (Analog) 2.5±0.625V

It looks like that maybe they ... make the same thing for higher currents while for 100A with 25mV/A we get: 2.5+/-2.5V 

Anyway, maybe they are really good-I do not need them