Smart Meter Installed

[Siwastaja]

I'm confident they use the import and export registers directly.

OK, which would mean there is no so called "net metering" at all, or in other words, any netting is down to the time resolution internally used by the meter for sampling, so very short. Possibly there can be import and export even within the same AC cycle. Would you agree?

Some years ago I was still under no-netting billing and when the hourly netting was finally mandated by law, the last few reluctant grid companies (such as mine) switched. The difference in bill is quite significant and while I understand the reluctance of going to very long netting periods (allowing people to use the grid like a battery in billing sense), something more than milliseconds would be fair but I guess unless it's mandated by law then there is no incentive for power/grid companies to do that.

This is only important when the import and export prices are significantly different, as they often are, and as, IMHO, they also should be. Paying the same price for exports would be a significant hidden subsidy of a kind. I would rather see net metering + fair prices than no net metering + artificially high export price.

[Someone]

I'm confident they use the import and export registers directly.

OK, which would mean there is no so called "net metering" at all, or in other words, any netting is down to the time resolution internally used by the meter for sampling, so very short. Possibly there can be import and export even within the same AC cycle. Would you agree?

I dislike the term net metering so glad you "quoted" it. Agree that timescale between swapping which accumulator is incrementing may be incredibly short as that is to the benefit of the retailer. It would need some hunting through technical data sheets to find out exactly what (programmable range?) it is set at.

[Siwastaja]

I dislike the term net metering so glad you "quoted" it.

I think while the concept is valid, people carelessly add expectations of some particular netting period on the term itself. For example, even Wikipedia has a definition which veers off to some certain measurement periods:

Net metering (or net energy metering, NEM) is an electricity billing mechanism that allows consumers who generate some or all of their own electricity to use that electricity anytime, instead of when it is generated. This is particularly important with renewable energy sources like wind and solar, which are non-dispatchable (when not coupled to storage). Monthly net metering allows consumers to use solar power generated during the day at night, or wind from a windy day later in the month. Annual net metering rolls over a net kilowatt-hour (kWh) credit to the following month, allowing solar power that was generated in July to be used in December, or wind power from March in August.

Problematic parts emphasized. I mean it's never "anytime". Even if you have the best of the best, annual netting, you won't be able to generate a lot this summer and use it five years from now. And then this quote kinda implies it has to be in the timescale of at least a month to be net metering. Why don't they, for completeness, mention an example of 1 second, why wouldn't that be net metering?

And here "net metering" is used to mean "use within 15 minutes". And in the end, all metering is net metering, so the whole binary discussion of whether one has net metering or not is kinda flawed. Maybe someone has net metering with period of 0.5 milliseconds, as accumulated into the two registers by a certain meter implementation. But it is always there, simply due to laws of physics it has to be, there is no physical concept of simultaneous export and import on a wire within same instant moment.

So I'd prefer to see the number instead of binary yes-no label, then it would be simple to compare the options, because (assuming export and import prices differ) then it's a simple "bigger is better" feature. I made some calculations when transitioned to hourly net from "no netting" (so some millisecond netting) and that increased my PV self use % from 52% to 70%, during the same time frame (calculated using both ways from the same data, both were available from the grid company). Then this also boils down to the question of how you define "PV self use". It has to be associated with some time period as well.

Here a few years ago some companies offered contracts which basically implemented yearly net metering and called them "virtual batteries", but they quickly fell out of fashion. Quite obviously such business cannot be a charity, so average customer of theirs must lose money. Basically what they did is to pay extra for exported energy to match the import price, and then get the same money (+more) back with monthly fees. And of course it comes with complex set of extra terms and conditions. Totally wasted human effort.

But, some netting period in range of, say, 10 seconds to 1 hour or so, would be fair to all participants including energy producers, grid operators and customers. Excessively short periods generate all sorts of weird issues with dynamic response of systems, e.g. Dave's issues could be related to too short net period (could be something else too of course). And it's nearly an anti-subsidy that you need to constantly import and export (e.g. when your thermostat is working on/off) while things in grid average out - your neighbor's thermostat isn't in sync with yours, so the burden you are placing for the whole grid and whole existing generation infrastructure does follow some longer than a few seconds average of your net power.

And then again excessively long periods are unsustainable for whoever is paying; it's a hidden subsidy paid to PV owners. Say 24h netting period: now you are getting bargain electricity at the most expensive hours possibly with shortage of production just because you produced 7 hours earlier when nobody really needed your energy. It obviously gets even worse with annual netting.

[Jeroen3]

Net metering is an invoice things. The meter itself should not be a part in that.
What you mean is how exact the meter can replicate the Ferraris energy measurement method.

I found a manual of a landis gyr e350 saying the "measurement time" depends on the current and is somewhere between 20 and 70 seconds. To maintain accuracy in low current the time is longer. Makes technical sense to me. But it has implications.
It is interesting, does this mean that you could net 0 within this time, no counters change?

I have written DSP for power measurement, but not for energy.

[coppice]

Net metering is an invoice things. The meter itself should not be a part in that.

Many meters only collect the net energy. Recent ones are tending to use separate import and export bins, to make things like solar installations integrate more smoothly, but there is a huge installed base of meters which only collect a single net energy bin.

[coppice]

I think perhaps rather than an internet search you should review the standards which set the requirements for metering. Digital meters have been around for decades and nobody is raising concerns about their billing accuracy (some probably did in the early days, of course.. but we didn't have an internet FUD machine to fan the fires of ignorance back then), so what makes you believe the same thing with a radio is worse?

I think if you do an internet search you will find multiple studies which have seen very big errors in the measurement of utility meters with difficult loads. Enough of them from institutions that seem to show something bad is going on. What is missing is enough follow up material to identify what is going wrong in those meters. Well, from having worked in this area I can give two educated guesses:

• The specifications and approvals tests are weak on what happens with loads with a wide harmonic spectrum. If the meter is, say, a 60A maximum one, they want a 60A current signal to exercise the ADC's input range well. However, if they take the sine wave maximum too close to the full range of the ADC they have little headroom for the crest factor of a distorted waveform. You'd think there would be at least a little bit of specification and approval testing in this area, but there isn't. Its quite possible that substantial errors come from the ADCs clipping.
• The meters I have seen which use Rogowski coils digitise the output of the coil, and do the calculus digitally. This is a problem when the waveform is spiky, as the output voltage is proportional to the rate of change of the current being measured. The peak output of the coil can be huge under those circumstances, and will clip hard at the ADC. This is like the first point on steroids, and in the reported testing the really big errors seem to be with Rogowski coil based meters..

[Jeroen3]

Yes. https://ris.utwente.nl/ws/portalfiles/portal/276415600/2021_EMC_Europe_Static_Energy_Meters_Running_Backwards.pdf

I can imagine creating such, "quadrant errors", with incorrectly tuned digital filtering that adds unintended phase shift with those poor current waveforms.
But they never looked into it further I believe.

I don't think domestic meters use Rogowski coils, I believe most of them are shunt based.
As these are the most accurate and not easily influenced by external fields. But I could be wrong.

[coppice]

I don't think domestic meters use Rogowski coils, I believe most of them are shunt based.
As these are the most accurate and not easily influenced by external fields. But I could be wrong.

Very wrong. There are millions of Rogowski coil domestic meters, especially in the US, where split phase power means a simple shunt sensor doesn't cut it. The thing that has held Rogowski coils back is they are HIGHLY influenced by external fields. Considerable research has gone into designing coils where the external influences cancel out across all the turns of the coil, but they are far from perfect. I have seen piles and piles of attempts to build immune coils within multi-layer PCBs, because there is a British company (whose name escapes me) that specialises in doing this which inspires them. Most people abandon the work in failure. Its a problem.

[Siwastaja]

Note though that all reported incidents of serious accuracy problems that I can find are from very unusual (I mean it in the sense of: either unintended, or malicious) loads. Like the water pump example discussed on this forum, and the above "let's use a triac dimmer to supply short 30-amp pulses to off-the-shelf SMPS". Such loads are either illegally on the market (probably the water pump), or the products are used against their user manuals (e.g. triac dimmer manuals normally list types of loads they can be used with). But in latter case the customer did not really understand they were operating a triac dimmer. So it will be hard to say who is to blame.

In ideal world, the meters should not misbehave even with such unintended loads, but as long as standards do not specify clear test requirements, nothing's going to change.

[coppice]

Note though that all reported incidents of serious accuracy problems that I can find are from very unusual (I mean it in the sense of: either unintended, or malicious) loads. Like the water pump example discussed on this forum, and the above "let's use a triac dimmer to supply short 30-amp pulses to off-the-shelf SMPS". Such loads are either illegally on the market (probably the water pump), or the products are used against their user manuals (e.g. triac dimmer manuals normally list types of loads they can be used with). But in latter case the customer did not really understand they were operating a triac dimmer. So it will be hard to say who is to blame.

In ideal world, the meters should not misbehave even with such unintended loads, but as long as standards do not specify clear test requirements, nothing's going to change.

That pump example was a very odd one. I wouldn't focus too much on that. I think it represents more of an interesting puzzle than a broad problem. Some fairly large studies seem to have observed very poor accuracy with some common place loads. A huge number of things, even products with PFC, present a horrible mix of harmonics to the grid.

[Electrodynamic]

Siwastaja

In ideal world, the meters should not misbehave even with such unintended loads, but as long as standards do not specify clear test requirements, nothing's going to change.

There is no such thing as an "ideal world" and reality gets messy.

For example, the smart meter shown below was designed and built by competent professionals following all the applicable standards and regulations. It is a marvel of modern technology which just so happened to burst into flames almost taking the house it was attached to with it. Were not sure why the smart meter had a melt down but we can say as a fact it did. I'm not implying smart meters are unsafe only that we do not live in an ideal world and stuff happens.

Note though that all reported incidents of serious accuracy problems that I can find are from very unusual (I mean it in the sense of: either unintended, or malicious) loads. Like the water pump example discussed on this forum, and the above "let's use a triac dimmer to supply short 30-amp pulses to off-the-shelf SMPS". Such loads are either illegally on the market (probably the water pump), or the products are used against their user manuals (e.g. triac dimmer manuals normally list types of loads they can be used with). But in latter case the customer did not really understand they were operating a triac dimmer. So it will be hard to say who is to blame.

I used to do a lot of R&D with transient effects and any time we have abrupt switching, a large inductance and small capacitance were asking for trouble. The reasons why this happens are known and are usually a result of localized LC coupling causing transient effects. So a water pump start switch partially sticks due to pitting or carbon build up, starts acting like a spark gap coupling one or more windings to the start or run cap. Now we have a LC oscillator generating HV HF oscillations within the system.

What many fail to understand is that it only takes one HV impulse in a local system to break down or degrade a component. Transistors or mosfets can keep working but we get leakage, variation of a threshold/negative resistance or abnormal fast or slow switching. In capacitors the dielectric can be damaged and can act like a resistance or continually break down abruptly at a different voltage.

So much of my testing was not under ideal conditions but to find the worst case scenario and it's really easy to find.

[coppice]

Siwastaja

In ideal world, the meters should not misbehave even with such unintended loads, but as long as standards do not specify clear test requirements, nothing's going to change.

There is no such thing as an "ideal world" and reality gets messy.

For example, the smart meter shown below was designed and built by competent professionals following all the applicable standards and regulations. It is a marvel of modern technology which just so happened to burst into flames almost taking the house it was attached to with it. Were not sure why the smart meter had a melt down but we can say as a fact it did. I'm not implying smart meters are unsafe only that we do not live in an ideal world and stuff happens.

I think its clear why there was a suddenly spate of e-meter fires, although I have no idea why any specific incident happened. If you are going to run a program of rapid meter replacements, you are going to have a lot of people doing the work who don't have a lot of experience, and probably a lot who don't have the sense of commitment wanting to sustain a reasonable reputation until retirement brings. So, you are going to see a lot of sloppy poor quality connexions which will overheat. Its inevitable in the real world. Massive demand surges in any job activity bring these kinds of problem.

I expect there are also some lousy designs which were either inherently a bit weak on contact quality or cooling, or made forming nice clean connexions during installation unreasonably hard, exacerbating the problem with inexperienced fitters. There is a big difference between getting pristine samples through approvals tests, and mass production and deployment.

[Siwastaja]

I would agree the most probable explanation for the fire is incorrect wiring: insulation getting under screw terminal, wrong torque, completely loose screw (I have seen that happen, like, the sparky just pushes the wire in thinking "I'll tighten the screw later" then forgets about it). But can, of course, be a mistake in meter design or manufacturing quality control.

[coppice]

I would agree the most probable explanation for the fire is incorrect wiring: insulation getting under screw terminal, wrong torque, completely loose screw (I have seen that happen, like, the sparky just pushes the wire in thinking "I'll tighten the screw later" then forgets about it). But can, of course, be a mistake in meter design or manufacturing quality control.

I expect a lot of people who never wired anything bigger than a kettle have been doing these installs, and not really grasping the huge difference between what you can get away with in a low duty 10A circuit, and the connection quality you need for a stable 60A or 100A circuit that might run at high load for hours.

[tszaboo]

I don't think domestic meters use Rogowski coils, I believe most of them are shunt based.
As these are the most accurate and not easily influenced by external fields. But I could be wrong.

Very wrong. There are millions of Rogowski coil domestic meters, especially in the US, where split phase power means a simple shunt sensor doesn't cut it. The thing that has held Rogowski coils back is they are HIGHLY influenced by external fields. Considerable research has gone into designing coils where the external influences cancel out across all the turns of the coil, but they are far from perfect. I have seen piles and piles of attempts to build immune coils within multi-layer PCBs, because there is a British company (whose name escapes me) that specialises in doing this which inspires them. Most people abandon the work in failure. Its a problem.

EU directives have made shunt based meters outdated, because they were inefficient. The meters were adding a bit of consumption, but if you add that together for 100 Million homes, it's a lot. So they are not going to install new meters with shunts.
The EDMI smart meters will be Current transformer based. At least all of them that I've seen were like that.

[coppice]

EU directives have made shunt based meters outdated, because they were inefficient. The meters were adding a bit of consumption, but if you add that together for 100 Million homes, it's a lot. So they are not going to install new meters with shunts.
The EDMI smart meters will be Current transformer based. At least all of them that I've seen were like that.

The last time I looked they didn't ban shunts. They just set consumption limits for the whole meter and the sensor. If you use a 150 micro-ohm shunt, at 60A the shunt is only consuming 0.54W, and the average consumption is much lower. The guts of the meter is consuming all the time, and these days those guts have got quite complex. How much is the typical EU smart meter now consuming continuously, with all they have forced into them?

[Someone]

I dislike the term net metering so glad you "quoted" it.

I think while the concept is valid, people carelessly add expectations of some particular netting period on the term itself.

Net weight: the weight of the underlying part contrasted with gross weight the sum of the whole.
Net price: inclusive of costs, taxes, discounts etc

Some places used "net metering" as in net weight, 100% of solar sold out at the export price, 100% of consumption bought at import price. Zero self consumption behind the meter, all energy accounted for through the market pricing. Which is a dumb way to meter but it adds massive confusion when using the "net" term as they mean completely the opposite in the two uses.

[Jeroen3]

I don't think domestic meters use Rogowski coils, I believe most of them are shunt based.
As these are the most accurate and not easily influenced by external fields. But I could be wrong.

Very wrong. There are millions of Rogowski coil domestic meters, especially in the US, where split phase power means a simple shunt sensor doesn't cut it. The thing that has held Rogowski coils back is they are HIGHLY influenced by external fields. Considerable research has gone into designing coils where the external influences cancel out across all the turns of the coil, but they are far from perfect. I have seen piles and piles of attempts to build immune coils within multi-layer PCBs, because there is a British company (whose name escapes me) that specialises in doing this which inspires them. Most people abandon the work in failure. Its a problem.

EU directives have made shunt based meters outdated, because they were inefficient. The meters were adding a bit of consumption, but if you add that together for 100 Million homes, it's a lot. So they are not going to install new meters with shunts.
The EDMI smart meters will be Current transformer based. At least all of them that I've seen were like that.

That's odd. I know from someone developing meters it's very hard to get the accuracy requirements at low load with CT's.
I also don't think a CT is lower resistance than a copper bar as shunt.
But maybe EU meters are different from US based.

50470-3:2006 has a table with as low as 2.5VA for class A meters or 1 VA for transformer meters.
And 2 W/ 10 VA for the voltage circuit. Did they make this even lower?

[nali]

EDMI use both shunts & CTs across the range. A problem with shunts is self-heating and maintaining accuracy across the environmental range, they also do suffer a bit from external magnetic fields as the busbar/shunt + Kelvin connections make a loop which has to be nulled by counter-looping the PCB tracking. The objection to CTs is probably cost.

[tszaboo]

EU directives have made shunt based meters outdated, because they were inefficient. The meters were adding a bit of consumption, but if you add that together for 100 Million homes, it's a lot. So they are not going to install new meters with shunts.
The EDMI smart meters will be Current transformer based. At least all of them that I've seen were like that.

The last time I looked they didn't ban shunts. They just set consumption limits for the whole meter and the sensor. If you use a 150 micro-ohm shunt, at 60A the shunt is only consuming 0.54W, and the average consumption is much lower. The guts of the meter is consuming all the time, and these days those guts have got quite complex. How much is the typical EU smart meter now consuming continuously, with all they have forced into them?

I don't have the numbers or the EU directive in front of me, but a quick search suggests that the limit is 0.6W per smart meter, which seems reasonable, even though they could've gone lower. So you will hardly pass that requirements with shunts and 3 phase systems. And no, they didn't ban shunts, they just made it very difficult to achieve the legal requirements with shunts.

[IanJ]

All,

In my system I am reading import and export power, but not from my EDMI Smart Meter, I fitted my own Owen Brothers modbus capable OB418 meter.
As I have stated before I also notice a residual import wattage when solar is making more than enough to power the house.......it varies between 1W and 20W, sometimes it is 0W.

Here's the spec, which may offer some insight as the EDMI Smart Meters are probably something similar:

From the OB418 manual:
-lref = 10A
-ltr = 1A
-lmax = 100A
-lmin = 0.5A
-lst = 40mA

Not from the manual, I believe this is the terms:
Iref (Reference current) = 10A: This is the reference current used for calibration. It’s typically the current at which the meter operates with its highest accuracy.

Itr (Transition current) = 1A: This is the current at which the meter transitions from low-current measurement mode to normal current mode. It’s a threshold below which the meter might use a different mechanism or mode for accurate measurement.

Imax (Maximum current) = 100A: This is the maximum current that the meter can safely handle and measure without damage or loss of accuracy.

Imin (Minimum current) = 0.5A: This is the minimum current that the meter can reliably measure. Below this current, the meter may not give accurate readings.

Ist (Starting current) = 40mA: This is the starting current, which is the lowest current the meter can detect and start measuring.

Ian.

[uer166]

Transition current has absolutely nothing to do with internal meter "modes", it's only a meter accuracy term.