Hybrid inverter - Metering options

With a basic installation, only 1 x EM115 import/export meter is installed with a CT on the grid cable (usually close to the main utility billing meter). This meter connects to the inverter via a Modbus comms cable, and is able to provide live values for power, voltage, current etc along with cumulative kWh energy values.

Given the cloud portal provides energy values for the ac grid import/export, ac inverter import/export, dc PV generation and dc battery import/export, presumably there are other metering devices built-in to the inverter to provide these values.

So the question is, how accurate are the inverter built-in meters, and is there any benefit in adding a 2nd EM115 to monitor the ac inverter import/output and connecting it on the same modbus daisy-chain network? It is shown as an option on one of the install guide schematics, along with an option with 2 x EM418 meters for FIT installations.

My thoughts are that the 2nd EM115 could provide the following benefits:-

1/ It would at least keep recording the cumulative energy readings during periods when the inverter has lost the WiFi/internet connection back to the cloud servers.
2/ Keeps the cumulative import/export values in the event of an inverter failure/reset, that can be read from the EM115 meter display.
3/ Possibly more accurate than the metering built-in to the inverter.

Any thoughts or experiences on this?

This is a good question. In my experience, the accuracy (or possibly timings) of the EM115 and the internal inverter output are quite a way off, so you may indeed benefit in what you are suggesting.

Simon, i don't trust the EM115/CT one little bit if i'm been honest.
In my Loxone home automation setup, i have a meter connected from my incoming mains to my main switch, now this meter is modbus which i monitor through the home automation software, i also have a smaller modbus meter connected to each of my outgoing circuits which are power hungry so i can drill down specific information from.
I have a meter connected right after the MCB which feeds the inverter and i get different readings on all of my meters, and to me it seems that the EM115/CT always seems to draw more than the other meters (granted i know its pulling information for the whole electrical install).

All my modbus meters (i have 9 installed) also have the pulse output, so i've even connected those to the software and set it up to count the pulses, to get a different aspect of information, and it's quite different.
With @hoggy 's help i managed to get the information from the inverter also using GivTCP into my software and even that information is a little bit off, so i guess there is nothing solid to compare against.
I know meters are supposed to comply with MID, but i still feel there is some differences.

Also is the wifi dongle just a data logger also? This will pull the cumulative energy information from the meter for import or export, so even if there is an issue it still reads what you have used, then once back online it will 'catch up' your usage. I wouldn't of thought there would be any 'other meters' stored on the inverter itself (i could be totally wrong like 😄 )

Thanks for the feedback DeanoC75 .
I have 3 spare ways in my mini consumer unit available for the MCB and 2 x EM115, which would keep the install tidy without a multitude of extra boxes for metering etc. So the next question is whether the EM418 meters are more accurate with the direct power connections rather than separately wired CTs. They should be if they are used for FIT installations.
I recently put in a question to GE support about the logging of data in the inverter when the WiFi/internet connection is down, and the answer was no logging in the inverter, so no catching up of the data when the link returns.

Simon_C Any meter you put inline should be an direct indication of consumption, hence why i have a meter on my outgoing supply to the inverter been read by a seperate modbus line into my home automation software.
TBH i don't do anything with the information, i just look and compare. However i don't have any solar at present, just using battery storage.

Mid meters are designed for billing and have to comply with certain specifications. One of my many hats at work is setting up metering systems (anything from a few meters to 2500) and I would never use a em115 unless I was forced. The only meters I like to use are Schneider - mostly because the software I use has them preconfigured so makes life easier. But they are also very good meters. And they have an option to be MID approved. That costs a fair bit more to have (probably more than the cost of a em115).
Not forgetting that the only meter that is used for billing is the suppliers. The em115 I doubt is mid approved.
(The most expensive meter I have had installed and then configured was about £5k. But there are ones that are more expensive.)

The GivEnergy datasheet for the EM115 does not state MID approval or any tolerances of measurement, however there is an EM115 Modbus meter available from Meters UK Ltd that physically looks the same and has a full datasheet that states the following:-

MID approved with module B & D certification
Accuracy class B according to EN50470-3
Accuracy class 1 according to IEC62053-21
Current +/- 0.5% of nominal FS with solid core CT sensor
Current +/- 1.0% of nominal FS with open core CT sensor

The GivEnergy datasheet for the EM418 meter looks more promising:-

Direct metering up to 100 A
MID approved with module B & D certification
Accuracy class B according to EN50470-3
IEC62053-21

The problem with the EM418 is that it is so much bigger, requiring 4 x DIN widths (76 mm), so will require additional enclosures and physical space to install 2 of them for the grid and PV monitoring.
I would stick with the GivEnergy approved meters to ensure compatibility with the inverter, such as the Modbus addresses for each data value. My experience with Schneider Electric meters is that the Modbus addresses vary from model to model.

Simon_C the later modbus meters - Pm5xxx range and pm8xxx (and others ) from memory have the same core registers with additional ones for their better specifications.
The older PM750 and Ion 3000(?) are different.

Simon, i have a slight feeling GE meters are the same from Owen Brothers or they all use the same supplier and just get them to change the artwork design 😄

I've got these Raleigh Meters installed on my higher consumption items, and these Raleigh Meters on my lower consumption items.
There is also a lot of Eastron SDM 230 and SDM120 modbus meters used out there.

You know what really P's me off is when you can't easily get hold of the technical data - like the modbus tables for kit. Neither RI or Owen have them on their websites that I can see, How can you make an informed decision about what data you can get out without seeing the relevant information? The RI ones run at a max of 9.6kbaud - which suggests to me that there is limited data available - but also that you can't change parity or number of stop bits - which these days is rubbish (There is only one system I know of that can support devices on the same network with different parameters and I avoid doing that too)

anglefire It's tempting to throw a bit more money at the system to get out data that we can control and trust but I also think that Givenergy should be providing us (the geeks that is) with a way to get at the data we want. Availability of data from the portal is getting better and more granular but I still don't know why the portal for my plant isn't right. Yesterday, the portal was showing 10.6kW of generation, consumption of 0W, Battery charge of 0W and export of 3.1kW. The consumption and battery figures do change and look like they are right some of the time, but at this point yesterday, the myEnergi app was showing the EV charging at full chuff - 7.2kW for about three hours.

Had I know then what I know now, I would have asked for EM115s to be added to the connection between each inverter and solar consumer unit (before the MCS meter). My MCS meter says I've generated 2,685kWh since commissioning whereas the Givenergy portal gives a figure of 2,898kWh, a difference of about 8%. The portal says I've exported 1,060kWh whereas my DNO meter says it's 1,119kWh a difference of about 5%. While it would be ideal to have much more accurate metering, I don't really mind differences of this magnitude. Of course, since my electricity bill and in-process SEG tariff payments are set by the DNO meter, it is that that is "correct".

The GivEnergy user manual for the direct connected EM418 lists the modbus registers, which will be required for compatibility with the GE inverters.

A web search for Meters UK EM115 reveals a data sheet from meters.co.uk which looks very much like the GE EM115, AND has matching modbus register numbers for the main required values as per the GE published EM418 meter.

Looking up single phase MID+Modbus meters from other manufacturers reveals different Modbus addresses so they will not be compatible with the GE inverter for grid or PV monitoring/control.

Schneider Electric iEM2155 - Direct connect
Meters UK EM408 Modbus - CT
Meters UK EM418 Modbus - Direct
Meters UK EM415 Modbus - Direct

No doubt they will be fine connected to a separate metering system where you can specify the required Modbus registers to poll, but no good for connection to GE inverters.

anglefire The RI tech data is available if you email them and ask for it, that's what I had to do for my stuff.
Also the RI smaller meters, their tech support sent me the software to program them easily, just had to use the USB to RS converter, and specify the com port in the software, then read current data, then send the required values to the meter and that was it done. Only issue i had was that the modbus registers were in hex code on the data sheet, the software I use in my home automation requires actual values, so using a quick hex converter let me do this.

Simon_C If you look at the GE EM418 user manual and the OB's user manual side by side they are pretty much identical.
So i'd probably guess that the GE EM115 will be pretty much the same as OB's OB115's meter.

Thanks @DeanoC75 , they look like a match, and the OB115 has the option of a solid core CT rather than split core. The solid core CT is supposed to be more accurate.

I guess it all boils down to whether the data from the GE cloud portal is meaningful and matches within a reasonable tolerance with data from the billing meter when using the default EM115 for the grid meter alone.

One step up would be to add a 2nd EM115 for the PV Gen meter instead of the inverter built-in metering.

Ultimate step up would be 2 x EM418 for both the grid and PV Gen meters, but at extra cost and difficulty of fitting in additional enclosures.

Tim . It is concerning that you are missing live values for the EV charging. Have you got just 1 x EM115 monitoring the grid import/export that has a shared Modbus network with 2 x inverters? This would indicate that each inverter is using the built-in metering for individual performance.

Simon_C The solid core CT should be more accurate, but check the size of the core as you may not get it round a 25mm meter tail.

Simon_C Have you got just 1 x EM115

No each inverter has one monitoring the cable between the DNO meter and consumer unit. It’s interesting that I can compare the meter values reported by each inverter to see how identical devices measuring the exact same parameter give different values. My inclination (if I was the person at Givenergy responsible for data manipulation) would be to average these values in the summary card. Maybe they do but with the data visible to me I can’t tell what they do.