Smart Meters overreading

[tabbypurr]

From another newsgroup...

Rob

22:16 (34 minutes ago)

- show quoted text - It may matter. When the current is far from sinusoidal, the computed power may be much more when the calculation is not done carefully, or when there are problems with the filtering.

A while ago there was some local coverage here in the Netherlands on a consumer tv programme about the fact that so many customers observe an increased energy consumption when they have a "smart meter" installed. Usually before that, they had an electromechanical meter of the "Ferraris" type, and now it is replaced with an electronic meter with remote readout capabilties (called "smart meter" here).

Of course the difference is not related to the remote readout but to the electronic measurement technology.

I studied the relevant regulations for an electricity meter and also mailed with the responsible person at the authority. It turns out that nowhere in the regulations the non-sinusoidal current problem is discussed. The only thing appearing in the standards is the handling of Cos-Phi. The meter should indicate kWh, not kVAh.

However, in today's households there tends to be a non-sinusoidal current due to small switchmode powersupplies (that are not mandatory to have powerfactor correction), LED lamps, etc. The result is that the electronic meter registers more than the Ferraris meter. How much more, that depends on the exact make and model of meter.

But that is all within spec, because THERE IS NO SPEC. When you claim the meter is wrong, it is being checked with a resistive load and of course that is metered correctly...

The difficult problem of course is: what is reasonable. One could also argue that the Ferraris meter displayed a too-low value, as the electricity companies have to deal with the non-sinusoidal current which causes them all kinds of problems. The customer should pay for that or stop doing it.

The last word has not been said about it, but for now I have deferred the change to a smart meter, and so have many others.

[Andy Burns]

Not related to smart meters as such, but all electronic meters (including smart meters).

Old news we covered it here back in the day

[harry]

Entirely spurious. Which household had 47 LED light bulbs for a start?

What proportion of the load is electronically switched? (Negligable).

So any inccuracies would be negligable too.

ALL electricity meters are inaccurate on very small loads.

Nothing to do with the meters being "smart" anyway.

You only have to look at the setup. The meters are wired with "bell wire". Clearly only been tested on light loads.

[Robin]

Round objects.

I suggest that - as so often - the truth is neither "the sky is falling" nor "we live in the best of all possible worlds". The standards for (and testing of) meters need to change but the risks of big errors in practice are (so far) small.

See eg review paper last year (with one author at NPL):

conclusion

"The Power Quality of the Electrical Grid, due to the effect of diverse emerging technologies and devices, are constantly changing the waveform conditions of voltage and current. These changes are more significant in the low-voltage electrical systems.

Current standards related to the accuracy of Electricity meters for revenue allows a measurement error up to ±2.0%, with a waveform distortion factor less than 3%. Such conditions are not representatives of the real working conditions.

Some Static meters, exposed to unlikely (but still possible) extreme working conditions, have presented high measurement errors in the presence of distorted waveforms.

New accuracy requirements should ensure Electricity meters never exceeds the MPE at any working condition. Regulatory bodies should promptly agree in how standards will include accuracy tests for existing and emerging distorted waveform conditions for an Electricity fair trade. "

[Brian Gaff]

Yes this strange issue came up many years ago when I worked at a factory making colour tvs. Back in those days valves were the main active device and all the sets tended to use autotransformers so the chassis was live in effect. We had a policy of burning in tvs before shipping as this tended to weed out the faulty ones. The CEGB as it was then complained about big hits on one mains supply which made the load 'difficult' The company had to move the racks to different mains phases on different supplies. It cost a packet! Then there were all the tubes of lighting to be considered which of course were the cheapest crappyest they could get, these also had power factor issues as well.

I had not realised that switch mode supplies had similar problems, and heck there are a lot of those in almost everything these days. Brian

[David Wade]

I have a small house in Spain. I think I have 32 LED light bulbs in it not counting the kitchen counter lights. In my house in the UK there are

11 GU12s all now LED in the kitchen alone, 4 LED floods outside plus five garden lights.

I many rooms we have "up lighters" which have multiple "bulbs" in the, So in total we have somewhere around 45-50 LED bulbs in the whole house. We don't put

We don't switch them all on at once of course!.

Also note that the meters were made no later than 2014 and most have been re-designed since.

[Harry Bloomfield]

[e-mail] wrote on 06/01/2019 :

No meter, Smart or dumb, will ever give an absolutely precise consumption figure, so the points made are entirely spurious.

All types are designed so as to give an 'accurate enough' charge on a typical load, rather than the exceptional loads. Which basically means some load types will be undercharged for, some overcharged for. Make an effort to deliberately run the 'over charged for' load types and you will be over charged, no surprise there.

[Andrew Mawson]

That's probably meant to be a 'Ferranti' meter type referred to above. Not aware of a Ferraris meter type

Andrew

[jrwalliker]

Would that have been the Decca factory? I thought they used half-wave rectification for the HT and series connected filaments. Half-wave rectification would certainly have caused problems at the substation if they were all connected up with chassis to neutral. John

[FMurtz]

They go faster.

[PeterC]

Trouble is, if the overcharging is on SMPSUs and the like, I have about 35 -

40 LEDs (only about 125W total load though), TV, PVR, AVR PC etc., microwave oven, fluorescent tubes in places (little used - not worth swapping for LED). Resistive loads are kettle and shower unit; cooker has spiders and moths in it; washing machine; all intermittent for short periods. Motors: pump in boiler. So the bias is towards SMPSUs - difficult to do anything about that. Most LED lamps seem to have a PF similar to CFLs - about 0.5.

[Harry Bloomfield]

PeterC wrote on 07/01/2019 :

At this moment, 19:30 on an evening - there are 5x LED's on, 2x switch mode powered TV's and this laptop, SMPSU x3 for routers, various other small SMPSU's with the CH pump cutting on and off frequently on the stat. + the boilers SMPSU on constantly. No resistive loads at all that I can think of.

A while ago I replaced all of the regular on lamps, with LED versions and they made quite a difference to my bill in winter.

My bills are quite small, so why worry?