Fronius Push

OK, after reading a bunch of threads (here and elsewhere) I was able to correct my generation data using the nodayenergy option (my firmware has the Day Energy bug), and now know that my Smart Meter is on the Feed-in side, so fixed the high reading by adding &n=1 to my smart meter push, but now I see odd reading during generation hours, lots of 0W reading and the daily sum is way off. I have tried a few different setting without much success.

My current push settings are;

I have also tried with &v4=PowerReal_P_Sum on the meter but it doesn’t seem to make any difference.

This is my system dwyers 6.440kW

And here is a look at the current issue, note usage line during production vs non-production times.

For anyone playing at home, it seems the only way to get good number from Fronius Push at the moment is to Donate and use the “auto=1” option, I bit the bullet and donated a couple of days ago, Just couldn’t get consistent values from the consumption side of things with any other settings.

I now have the following push settings;
Inverter (using nodayenergy do to firmware bug)

Numbers seem fairly accurate for production and consumption, but oddly my grid import\export values seem to vary, and they should be coming direct from the meter values. I am also using a nodered flow to prob the fronius api and log values into my home automation system, I am “Computing” House Usage and Solar Production as there is no direct meter value for those, my automation system has an option for this, I just feed it current usage every minute.

Yesterdays Values from each system, AC was cranked on a hot day, so consumption is high.

Probably won’t be a big issue on days with lower import\export, but I am surprised by the variation on values that should be coming from meter readings.

To be honest, I can’t see how anybody who sets v3=None & nodayenergy=1 can get anywhere near accurate consumption values given that PVO is merely integrating a “spot” consumption value over the 5 mins. Set the microwave on as that spot sample is taken and you get that spike averaged over 5 mins as representative of your energy used(!). If you have a smart meter then Fronius internally has a class 1 integrating energy counter for both consumption and generation and it just seems ridiculous that Fronius can’t write simple software to reliably push these 2 values into PVO.

The v3 is for the inverter, not the smart meter, I added that to ensure it wasn’t taking consumption figures from the wrong place.

The Inverter is using nodayenergy=1 due to a bug in the firmware that sees its DAY_TOTAL value stop counting, oddly Yearly and Total Production values increase correctly, def some poor programing there.

The smart meter supplies meter reading for both grid consumption and grid production, hence why I am surprised the figures are off. The production values, the ones being calculated are closer than those that are supplied via the meter.

For my home solution I am actually using GetPowerFlowRealtimeData.fcgi as it gives me most of the values I wanted, I only pull the Consumption\Production meter values from GetMeterRealtimeData.cgi and nothing at all from GetInverterRealtimeData.cgi.

I currently have my PVOutput inverter push set to 5min, and the meter set to 1min, could that be causing an issue?

From the Help Docs

Disable consumption energy upload, i.e. calculate from power


Maybe that is causing some issue, but it still should be using my meter push for consumption data, not really sure how the 2 pushes are merged. I’ll remove the v3 and see what happens. But really consumption seems about right, its the import\export that are off.

v3 is energy consumption which has to come from the smart meter and can only be configured on the meter push.

I don’t use the Fronius push at all given the bugs. I use my own RPi application which works 100% accurately and has for all recent Fronius firmware versions regardless of any bugs people have reported with the native push. The JSON API has always worked fine.

From GetPowerFlowRealtimeData.fcgi I grab v1=E_Total.
From GetMeterRealtimeData.cgi I grab v3=EnergyReal_WAC_Sum_Consumed and v6=Voltage_AC_Phase_1.
I push them up as a single 5 min push with c1=1. I also derive v2 and v4 (from the change in v1 and v3) because PVO has a bug that requires you to do that if you want to see standby values.

When auto=1 is enabled, the v3 setting is ignored. Each ‘auto’ push from the meter provides the amount of exported / import since the last push. The gross energy and power is calculated from this and any available interval push generation data.

The 1-minute pushes are rounded to the nearest 5-minutes, if your connection is reliable then a 5-minute push would be better.

Interesting, your avoiding the “bug” by using E_Total, Day_Total seems to be the only value the bug effects, it only dawned on me this afternoon that I could use that as my automation system takes a cumulative total.

Would you be willing to share your furmulas/code, I could easily modify my node flow to push to PVO at the same time as my own system. But still fairly new to solar and still trying to work out what all these V values equate to.

Thanks for the reply and info, I’ll roll back to 5m as my connection is very stable.

How does nodayenery fix work? Is it computing for current power values or does it roll back to using the Total Power figure as a life time Meter?

Can you explain why I am seeing variation in the import/export values? Are they not just a Meter reading?


The DAY_ENERGY values from Fronius are ignored and energy is calculate from PAC or power.

The feed that Fronius receives for solar web is different from the push data that is sent to external parties such as PVOutput. The meter data only contains net import/export information and must be combined with the inverter push to derive gross consumption. PVOutput does not store net data, these are always calculated from gross consumption and generation data.

Firstly excuse my ignorance, I only started playing with this stuff a week ago and I’m still trying to get my head around all the values in these APIs.

So the import/export meter values are not used directly for those feilds, just part of a bigger calculation, and that could cause the difference in values?

The inverter push does include TOTAL_ENERGY, could this not be used in a similar fashion to the import/output meters to derive production rather than using PAC to calculate? Although production values seem close enough anyway, so not sure why I am worried.

I’m also confused by why my smart Meter has so many meters all with the same value repeated, on my system, assuming 3 phase could complicate it, but do the absolute and sum ever differ?
EnergyReal_WAC_Minus_Absolute = EnergyReal_WAC_Phase_1_Produced = EnergyReal_WAC_Sum_Produced

EnergyReal_WAC_Plus_Absolute = EnergyReal_WAC_Phase_1_Consumed = EnergyReal_WAC_Sum_Consumed

Do you know what the EnergyReactive_VArAC values are?

It would have been nice for fronius to have included a household consumption meter along with the input output values.

EnergyReal_WAC_Plus_Absolute and EnergyReal_WAC_Minus_Absolute are used. These are the total lifetime export/import reported by Fronius meter. The difference in subsequent pushes in these values determines the amount exported/imported over that interval using the push data timestamp.

Consumption = Generation + Import - Export

The inverter only reports DAY_ENERGY, the bug affects this value and it stops increasing at midday. So this value is not reliable.


Sounds like it’s treating those values the same as my home system, I just feed both to a virtual smart meter with usage and return values, the meter also shows instant values as well as the meters. Still not sure why those particular values seem to vary from my local and solarweb values, but not a huge deal.

Oh, I had assumed the push returned the same data as GetInverterRealtimeData.cgi that shows Day, Yearly and Total energy values.

Mine doesn’t stop at midday as such, it seems to count slower most of the day and stop late afternoon, I read in another thread that it may be related to having split arrays, which I have, but both mine seem to produce most of the day.

I’m pretty sure I’ve had, and have now fixed this problem now, it was related to how the installers set up my smart meter - the Fronius install docs allow for two setups with the meter in different locations on your circuit, and if the setting below is wrong, it (ie PVOutput and solar.web) thinks your net consumption is actually generation.

I just had to switch it to ‘Feed-in Point’ in the inverter settings, instead of “Consumption Path” (browse to http://inverter-ip/#/settings/meter and log in with the ‘service’ password for this setting).

This fix, of course, assumes that the installed actually did install the meter this way (apparently my install was the less common way).

1 Like

Hi Sam,

Thanks for the reply, mine is installed on the Feed-in Point, but it was already configured as such in the setting by the installed, you can check it under one of the menu’s. I haven’t been able to find a ‘service’ password, but haven’t looked too hard as it does appear to be return good values now.

I just couldn’t get reliable usage values until I donated and used auto=1, I would be interested in your push settings if you could post them.

That said I am now leaning towards creating my own flow to push to PVOutput, because I am still seeing a 10% variation in the import\export values. Hoping pushing my own consolidated data might help. Just have to get my head around the PVO api values.

I’ve read a few of your posts, I would like to know what push settings you have for your inverter and meter.
How accurate is your pvoutput compared to solarweb with these settings?

cheers mate


I am using the following push settings;
Inverter Push - Interval 5 Min
Upload file name: /service/r2/froniuspost.jsp?sid=&key=&nodayenergy=1

Meter Push - Interval 5 Min
Upload file name: /service/r2/froniuspost.jsp?sid=&key=&v6=Voltage_AC_Phase_1&auto=1

Generation and Consumption Data is pretty good, Export, Import and Self Consumption, not so much. My Numbers for the month of December (in kWh).

I am still waiting on a bill so I can compare some figures with those on a Bill, I have a nodered flow feeding data into my Home Automation software (Domoticz) and configuring that to feed back to pvoutput is on my list, but its a long way down.

@beaud Thanks for the reply…

I’m using very similar settings which have shown to be quite accurate.
I have had 4 ‘test’ IDs going with slightly different variables to see what’s most accurate.
Over the last 2 full days I have been within ~0.1% on Consumption / Imported / Exported / Generated comparing Fronius data to PVO.

5 min pushes
Inverter: &v3=None&nodayenergy=1
Meter: &v6=Voltage_AC_Phase_1&v4=PowerReal_P_Sum&auto=1

Actually found that using &delay on the meter gave me less accurate information. (but still within ~0.01-2%)

I found using &v3=none actually made a (good) difference, even though using auto=1 should not take into account v3 (so I read somewhere)
&v4=PowerReal_P_Sum seems to give more accurate ‘snapshot’ power use data, but at the end of the day doesn’t make much of a difference.

See how it goes over the longer term.

I did have both &v3=none and &v4=PowerReal_P_Sum at points while trying to diagnose, I’ll add those back and see if my accuracy improves, to be honest I have sort of given up of PVOutput giving me quality RIO data, still great for comparing systems and tracking production and consumption.

The Solarweb Premium Trial, is just a free upgrade to premium in Australia, it’s not really documented anywhere but if you go though and sign up for the “Trial” its good for 20 years.

From mine.
You are currently a Solar.web Premium member. Your membership will be renewed on 09.12.2037.

Once a premium member and with the Tariff data entered it give pretty good ROI Data.

So after a few more days of data, it sees those values have had some effect, but still plenty of difference, just on different numbers, maybe a little more spread?

But I have noticed that on a clear day, when output is constant the number are very close, if there is any sort of cloud cover, where production drops, this seems to mess with the numbers, I did read in a thread somewhere that this was an issue with split arrays (which I have) so that might be the reason why I am seeing my variations, I’ll have to look into it more.

The last day (18/01/18) was nice and clear, and the graph is nice and consistent, on the other 2 days the consumption drops to 0 on a number of occasions, which is clearly not right, the drops seem to co-inside with production drops.


Hi just a thought,but have you checked to see all the time stamps line up to the second, thats your computer,fronius push in the inverter and also you second device if using one, i have noticed my figures are off when my inverter time is out by more that a few seconds ,i have seen it up to 3 mins out,and i have found that auto update for time is not good enough on the inverter.:slight_smile: