Graphing troubles - resolved

I certainly do need to have a better read of the documentation. Without the Fronius pushing generation data into the same PVOutput system, the IotaWatt records have been (for the past hour) much more complete, but still with gaps (for voltage & consumption) - illustrated.

I have the Fronius pushing to a separate PVOutput system, as it has done for years - the introduction of the battery last year resulted in consumption closely following generation - as expected - shown below - so I now have most of the required data within PVOutput, I just need to figure how to consolidate it within a single system.

There are three data outputs from IoTaWatt to PVoutput. You seem to have consumption defined. You still need voltage and generation. Here’s a complete iotawatt example:

These sites run day-in day-out with no holes.

The battery, along with my need to segregate circuits makes the generation data somewhat confusing - the only accurate data for generation is directly from the Fronius, which is now happily feeding it’s own PVOutput system again.

So, for now I’ll leave the IotaWatt feeding just consumption & voltage to a seperate PVOutput system to confirm if the data is without gaps - recent hours show a few.

I don’t thinkthose are gaps. The voltage is being recorded fine. Don’t have much to go on, but it appears to start happening when the solar would be contributing power. I think you may be sending negative consumption data, which PVoutput will reject. Again, you need to use a max operator to eliminate the negative values.

Do you have a description of how your battery system works or a line drawing?

An overly simplistic explanation of the Victron ESS battery system is that it meters consumption, and direction at the grids point of entry to the property, and via a bidirectional inverter/charger attempts to keep consumption (from the grid) at zero, while supplying loads via the battery.

In this instance the battery comprises of 5 complete Nissan Leaf packs, totalling around 140Kwh. The inverter chargers are Victron MultiPlus II configured for parallel operation.

There is an expected amount of grid inport/export as loads are switched on and off due to inverter under/over supply and their speed of response - illustrated below.

Now adding the battery output to the chart shows it supplying the property loads.

Added to the above there is a MyPV AC:Thor - which diverts to resistive heating, any export - this is a fairly recent addition and is primarily intended to operate when the battery is full.

An amount of fine tuning of the point at which the diversion device starts operating is required, but as can be seen its response to any export is very fast - far faster than the Fronius OhmPilot which it replaced.

As mentioned previously my intention is to consolidate as much data as possible into a single screen, probably eventually via Influx/ Grafina etc - having had the IotaWatt in operation for nearly two years my expectations of what I can do with it have changed - I do appreciate that my application is likely stretching its originally designed for purpose, there’s a fair number of DIY battery system users of your product, some of whom have already achieved what I’m aiming to do in terms of a single point of data display for an energy storage system.

You are spot on regarding the “gaps” - below illustrates before and after the ESS was switched on - as per a previous chart, what PVOutput is now showing as 0w will likely be a brief period of export.

Thank you Bob.

Quite a setup! I’m interested in understanding this better and figuring out how to use IoTaWatt to get the most insight into how it’s working. That said, I have a few questions about the equipment and line drawing, and a couple of comments on IoTaWatt monitoring vs PVoutput vs influxBD.

First the comments:

IoTaWatt records net power and energy at five second intervals. PVoutput does the same at five minute intervals. So looking at the IoTaWatt import/export graph with five-second net would look much different than the same plot with five-minute net. You can open a second graph window and zoom out to five minutes intervals and see just how most of the spikes will be greatly reduced or even eliminated.

That tees up PVoutput, which can digest the five-second data and feed it to grafana or other visualizations.

If the five minute nets serve a useful purpose (and I think they do), then as a donator, you could upload up to six additional data sets to develop a more comprehensive picture.

You don’t need to wait after making changes to the PVoutput driver to see results. You can check the box to re-upload a few days history.

Questions on the line drawing:

I’m assuming your main CT is on the fat blue line between the Grid and the point where the “PV on AC INPUT” comes in.

Do you have a CT on the PV on the yellow PV on AC INPUT line and what is it named?

Do you have a CT on the fat blue line coming out of the top of big inverter/charger and what is it called?

Are there any CTs on the orange Critical Loads line? I’d like to see one between the inverter and the big inverter/charger, and another on either of the other two parts of the “T”.

Where is the diverter in all this? If it’s just part of the loads, that’s fine. If it’s somewhere else, where?

Does the inverter/charger supply the critical loads during a power failure?

This badly edited image should help illustrate the actual - I’m aware that circuit segregation will be required.

The original AC coupled PV was installed some 7MWh ago (I can’t recall how many years ago that was) on it’s own sizeable circuit - a few years ago an assortment of battery systems came and went, along with a number of management systems as part of some evaluation work for a commercial requirement - the system that remains now is somewhat of a legacy from that work, having proved how effective such a system can be.

So all PV, and battery generation is fed via a single circuit back to the point of consumption - thats the physical constraint until I segregate.

To answer your questions:

Mains CT is at position 1
CT at position 2 “Workshop” - which meters all generation PV & battery
CT at position 3 “Diverted” - dedicated circuit to diversion device.
Critical output not used or metered.

The diversion device is modbus tcp/ip connected to the Victron Venus device - there is also a Batrium BMS for the battery, which dictates to the Venus device charge and discharge parameters etc … this is all functioning rather well, other than my wanting to consolidate the data, rather than having it spread across Victrons VRM, Fronius SolarWeb, and PVOutput - added to which is IotaWatts ability to log per circuit.

If I could aggregate just the Fronius pushed solar generation with the IotaWatt data into a single PVOutput system I would be happy, but I’ve tried that and it stomps on the Iotawatt data (leaving no actual consumption records)

Did any of the above make sense ?

Just some additional info on the diversion device for anyone interested - this Victron illustration differs slightly to my actual installation only in that the diversion is taken from the main load, rather than the critical load side.

1 Like

It’s different. How does the inverter/charger know when to generate/export to the load, and when to import from the PV to charge?

There is an energy meter connected immediately after the suppliers entry to the property - any PV surplus (after property consumption) is throttled to near zero by charging the battery - once battery is full and assuming property consumption is satisfied the potential export is diverted to resistive heating.

During no/insufficient sun hours the Venus device, using the metering device provides inverted power attempting to keep import at zero.

Ok, that completes the AC picture. Is there any measurement of the PV production on the DC side, and is that somehow included in the production data that the Fronius is uploading to PVoutput?

All of the DC side is either via Victrons VRM, or locally via the BMS - so all of the data you are ever likely to need, but scattered.

Some smart folks are building some very tidy all in one displays.

I hope Crhis doesnt mind me using his image, this is his Schneider charger/inverter system, Batrium BMS, IotaWatt, Influx and Grafana combo display.

1 Like

Nice. Now I see where you’re going with this. I’ll digest this information and give you some of my insights later.

We all enjoy your insights Bob, thank you.

The image I attached shows how Tesla often installs a Powerwall partial Home Backup design for residential users. I measure via CTs (red circles) my PV and the Powerwall (battery) at the sub-panel. Unlike the example, my mains terminate in the sub panel as well turning my old main panel into a sub panel. I also have a few other high load circuits in the sub-panel such as my Range/AC/Dryer which in my case is to prevent the battery from being drained from the heavy load items, and they are not able to be used during an outage from the grid. So, unlike with this diagram, my mains (excluding the sub-panel circuits) are backed up by the battery during an outage. While I have solar and grid, all items are run off the battery unless there is not enough solar/battery to supply the home’s needs.

Once night falls, my entire home runs off the battery until the designated percentage I set for reserves. The one Powerwall often gets me through the night and begins charging once the sun comes out. When it does not, the home simply takes from the grid what is needed to run the home. The battery will only charge in this setup from the PV and not the Grid. Seems quite different from your setup, but thought I would share if it helps.

Here is a view of my physical setup with the components labeled. A bit more helpful seeing a picture I thought. I had a basement window that was blowing out the image in the top right which I tried to tone down some. As you can see, I have the IoTaWatt with ties into each of the three panels (Main, Grid, and PV/Battery). I measure all without issue. Note, this is a two-phase US install if you were not aware.

he Non-Backup are the high load circuits that will not function during an outage (Dryer, Oven, AC) while all the circuits in my main panel will (heat, well pump, and all other power needs). I have CTs measuring the Battery. Solar, Grid, and a few of the high load items as well.


My system operates in a very similar manner - it simply sits parallel to the house supply, feeding just enough inverted power such that the amount imported from the grid is near zero - upto the limit of the inverter, any consumption above the inverters rating is provided by the grid.

In the event of a grid supply failure the system effectively becomes a UPS via the critical load output, which I have physically wired, but not well distributed - so could be used if required.

Timed grid charging is possible, to take advantage of ToU tariffs, there’s node red integration … etc etc and it’s all open.

1 Like

Seeing as you’ve shown me yours I feel obliged to show you mine, although your image is rather more illustrative and well planned.

1 Like

Impressive if I do say so myself.

With my monitoring using the IoTaWatt, I can measure Total PV, Home Consumption, along with Grid Usage, Battery Import (discharging or drawing) and Battery Export (charging or sending excessive solar to battery). All of this I send to my local EMonCMS running on a PI. The top box portion is relative info. So the Grid is what I imported and what was exported in that one number. The bottom section are actuals of what I have done in the 24 hour period. Here you see I imported 5.46kws from the Grid. But with a good solar day so far, I exported 24kws to the grid and 9.2kws to the battery for a new full charge.


As @overeasy said, I also use the extended data in PVOutput to send my battery status around charging and discharging. It works great.

I can say that much of this was a simple process of working with others here, testing, and luck in some ways. Once things are working, this is a great set-it-and-forget-it solution.

Hey Sean… I though u had said above that your system allows 50w to be fed back to the grid before it will divert to hot water element?