Measuring AC low power consumption with multimeter/oscilloscope combo?

[nightfire]

Situation: I am currently thinking of doing some measurements of (standby) power consumption on some small devices like USB power supplies or some LED lightbulbs, that run off 230VAC.

The question I am currently brooding over is:
Is it (theoretically) possible to do some high quality measurements of AC power without a dedicated energy meter?
I take it that in the measurement setup all parts can be connected firmly via 4mm plugs (multimeter), and an oscilloscope via BNC->4mm adapter.

Would it theoretically work if I put a DMM with a significantly burden voltage (like, 1V in the used range) in series and put another DMM in parallel at that point to measure the voltage drop?

For the modern plague of electronic power suppllies that generate circuit feedback or operate in a pulse fashion like PWM modulation, probably a DMM is too slow to check for things- would a standard 2 channel oscilloscope be able to capture enough data so that calculations of the actual consumption could be generated later offline in other tools?
(Yes, I know about the reasons to use things like isolation transformer and differential probes in such moments)

[jonpaul]

$10..20 Mains digital wattmeter

https://www.amazon.com/Electricity-Electrical-Consumption-Backlight-Protection/dp/B09BQNYMMM

Due to PF and non sine currents, using meters and scopes is not a solution

j

[Kleinstein]

A simple multimeter would not work, as the current and volage are not in phase and with low power devices (usually without much PFC) the current waveform is also usually not sine. There may be a few exceptions of DMMs to actually offer AC power measurement.  1 or 2 very fast DMMs may replace the scope, but that would be the exception and the scope would be easier with synchronous measurements.

In theory it would work with a digital scope. However this needs suitable transducers / protection (e.g. isolation transformer, shunt and suitable divider/probe for the voltage). Just 4 mm cables and a shunt would not be safe.
For the math needed I don't know if there are suitable functions available - possible, but 2 channel math is somewhat rare. Worst case one could export the data and do the math with a spread sheet program.

It may be possible to modify one of the cheap power meters to get a smaller current range. This would still be tricky with the protection (add fuse and diode clamping).

There is also an indirect thermal method:  put the DUT together with a sensitive temperature sensor and known power heater inside an thermally insulated box and record the temperature rise from its power consumpten. The extra heater could be used in a 2nd run with a known DC power for calibration.

[inse]

For the same reason, I bought myself a power meter with only 5A current range.
The readout goes down to single mA and Power measurement 10mW.

[nightfire]

I already have this for consumption up to 0.1 Watts: https://de.elv.com/elv-energy-master-basic-2-energiekosten-messgeraet-130412

For me, the question was how to detect and measure lower power draw, and I have several multimeters and adapters on my bench ;-)

I was thinking of the following setup:

Isolation transformer to float the DUT. On the L' I would place a handheld multimeter for Amps measurement. After the multimeter, I would place the socket for mounting the DUT, and where the N' from the isolation transformer is connected.
Measuring of current by the DMM itself, and then I thought of connecting the mass from the scope to the COM port of the DMM as a common denominator.
(And therefore earthing the whole stuff again, unless a differential probe would be used- this for reference for later readers that look for suggestions, as this will blow up when accidently wiring this in the wrong way)

The scope would measure the voltage drop above the shunt of the DMM, and the voltage after the DMM and N', therefore being able to have some idea of the difference between voltage and current.
Depending how well I would get the measured data out of my scope, I can piece them together later in some external application with ease.

The role of the DMM here is to provide a good quality shunt including protection mechanisms around it, and to give me an idea of the current currently flowing, so to get some rough idea to check if the values the scope measures in relation to the measured resistance of that shunt+DMM fuse are plausible.
Of course I will have to use some 100:1 probe for the 230V portion of that setup to get that signal conditioned...

[bdunham7]

would a standard 2 channel oscilloscope be able to capture enough data so that calculations of the actual consumption could be generated later offline in other tools?

Yes, and some will be able to do your power analysis on-board in real time.

Here is a Milwaukee Tool SMPS battery charger on standby (no battery inserted) set up using a 1R shunt and a 10X probe on a Siglent 2354X+.  An isolation transformer was used to power the DUT, the 1R shunt was in series with the DUT and across CH1 of the scope, the 10X probe was on the other side of the DUT and connected to CH2 of the scope.  You can see the reactive power, presumably due to X/Y caps, and then the small real power spikes. 

[inse]

As bdunham shows, you will need a scope with the required math/analysis capabilities, also a higher resolution than 8 bits may be favourable if you want more than just visualization.
Above that, a multimeter is not the best solution for a shunt as the fuse is in the wrong place.
You include the unknown fuse resistance to your measurement and once it should blow, you have the mains voltage on your current sense line.
This is the meter I use in the 5A version:
https://www.ebay.de/itm/263951200018?mkcid=16&mkevt=1&mkrid=707-127634-2357-0&ssspo=qbAJ5icsSjG&sssrc=2349624&ssuid=UJEfguXdQ76&var=&widget_ver=artemis&media=COPY

[Kleinstein]

The 8 bit resolution of most scope should not such a limitation. There are quite a lot of data points and thus oversampling to keep the noise down. Different from the cheap ready made power meters one can choose a suitable range and thus looses less of the dynamic range. The cheap power meters have a fixed shunt and thus a range up to some 20 A even for a low power divice that may take less than 100mA. So they loose quite a lot of the dynamic range for unused headroom. A 16 bit ADC with 8 bits lost is not really better than a 8 bits ADC with nothing lost.

The 5 A power meter may be slightly better than the existing meter that is likely made for a 16 A range, but not very much. To get a much better resolution it would need a change in the shunt / current sensor.

[nightfire]

Thinking a bit deeper about it, one of my main goals here would be that I have some visualisation of whats going on, like @bdunham7 described- alone to get some grip on whats exactly going on in the DUT.
It would also allow me to use already existing tools on my bench to get to some results without buying new stuff.
As my existing power meter goes down to 0.1 Watts, I can measure lots of things quite correctly, but would want to cross-verify its readings and to be able to go a bit lower for devices that only use very small amounts of power. Here detection is the first priority, very tight precision measurements up to 0.00x percent accuracy not so much, at least not in the moment.

Good point with @inse, that a fuse might blow due to some unexpected circumstances and therefore I would have to take this scenario with safety considerations into account whilst planning the setup.

Otherwise I might be tempted to order some 1R precision resistor to use as a shunt, and measure this with a good bench DMM (Have a Fluke 45 available that was calibrated)
Aaand I just was aware while browsing the manual that the Fluke 45 might do with its dual display a 3-wire reading of AC current and AC voltage simultaneously, and probably do some output via its serial interface.
I might get a bit excited about this, albeit the question is if the sample times are sufficient for the things I want to do, AND the readings are made in the exactly same timespan, and not with a certain small delay.

I see, I have some homework to do...

[Kleinstein]

The Fluke 45 is to slow to replace the scope.  Most of the measurements are done for the average of a whole power line cycle or even more. To calculate the true power one needs a time resolution way better than that, more like 100 µs or better. So more like the speed of a sound card or better. It is still a very bad idea to use a sound card for measuring mains !

The resistor for the shunt does not have to be that accurate - the sopes are usually only in the 1-3% range and the temperature is usually stable enough for the mains period. The DMM is OK to measure the resistance and maybe check the scales of the scope. Just make sure that the shunt is not failing open (e.g. more like sufficient power rating and maybe 2 or 3 in parallel). 

The slight difficulty come from the point that power is flowing in and out. So the final power consumption is a sum with positive and negative contributions. In the curve shown by bdunham7 the more sine like part of the current is likely from an EMI capacitor and thus resulting in near 0 net power. The actual power used is mainly from the short extra current peak.
The good thing is that the same scale factor applies to the positive and negative parts - so here errors from the shunt or scope gain don't enter for the compensating parts.

[nightfire]

Regarding dimensioning of the whole setup: As I already have a decent power meter, I only have to measure a power below 2-3 W. This means, even if there is a cosPhi of 0.3 involved, something in the region of 30 mA- which gives together with decent voltage drop of some volts probably less that 0,5 Watts power dissipation over that resistor.
So some ordinary 5W resistor should be enough to withstand that kind of energy, or the big gun like Vishay RH005 ;-)

What makes me curious: In the datasheet to the RH005 types, Vishay states:
"Models of equivalent physical and electrical specifications are available with non-inductive (Aryton-Perry) winding."

Question for the more experienced folks: To which extentz do I have to worry about inductances from those wirewound resistors?

[mariush]

Here's a question ... as most hardware will have some kind of full wave or half wave rectification inside to DC, why not just  power the device you want to measure with DC voltage ?

Get a bridge rectifier and maybe 1-2000uF worth of capacitance to get the voltage relatively smooth and then add a current shunt in series with the output.
You can then measure the dc voltage after the shunt, and the voltage drop on the shunt to get the current.

The downsides .. I imagine the rectifier diodes inside products would be more "stressed" as now only half of them are always conducting, and maybe some products rely on undersized bulk capacitors after the rectifier to basically assure the input voltage never gets too high (as the capacitor discharges fast before it's charged again by next ac cycle)
And I guess you have to be careful with capacitive dropper style power supplies (cheap led bulbs)

But just the feeding of DC to a device should not be an issue, am I right?

If you use enough input capacitance, your input voltage will be quite steady, when you try to measure a usb charger that consumes less than 1w at idle so that would make measurements quite easy. 

[Kleinstein]

The power consuption from DC may be different from the power consumption from AC.  This is especially the EMI protection part that can add more loss if AC powered  (the capacitors are not perfectly loss-free and the current can cause losses in fuses, inductors or resistors (e.g. used as a fuse or to limit peak current). There is also the question which voltage will be reached after the rectifier, especially if there is relatively little capacitance or series impedance.

Ideally the data-sheet, maybe for a similar type should get a hint on how much parasitic inductance to expect.  Resistors in the 10 ohm range would also be available as film type, if nothing else helps use a few in parallel.
Most of the relevant power should be below some 1 kHz and there is usually little of these higher frequency in the mains voltage. With nothing in the voltage there is no phase shift to worry about for the higher frequencies. So inductance should not be that bad.

[bdunham7]

Question for the more experienced folks: To which extentz do I have to worry about inductances from those wirewound resistors?

Probably not a lot for this type of application.  I simply have an 1R 1W WW 5% resistor with a 'shim' resistor to get it as close to 1R as possible and then a 1A fuse on the high side so that if it blows there's no at the resistor.  You could order a 1R 1W NI 1% resistor for a few bucks and it would be more then good enough.  You might get better (less noisy) results with a 10R.  I just picked up my ready-made small current shunt that I use with a DMM and rigged it up as stated with the scope.  I have nice Welwyn 0R1 1W shunts that I use for actual power measurements up to a few amps, you might just want to get a set of 0.1/1.0/10 ohm resistors to do this sort of stuff.

[rs38]

@nightfire did you succeed?

I'd be very interested if there are some agreed outcomes how to get low power AC standby consumption measured with some confidence and accuracy without spending >1000$.
could be starting with a scope meter a nice option especially if you can import data to PC and use more advanced OSS scope tools?