EV charger reads seem low

[PauloGoncalves86]

Hi,

I have an ohme charger and octopus agile. I configured the agile integration a while back and the ohme just a few weeks back and it has been fantastic, it really helps me to monitor it. And also nice to get the clamp value reading, giving another option to estimate the consumption before the actuals arrive from Octopus (no Mini here..).

The Energy dashboard has been working well and I configured it a while ago. I see something that does not look absolutely right. I have a petty stable consumption in the house with some peaks for the use of the oven, except when the EV charger kicks in.

In the attached chart, it is obvious that the untracked consumption goes up whenever the charger is in use. The EV figures look consistent between the detail and total usage, but I suspect they are slightly low, as the 7.7kw would lead to a maximum charge rate of at least 3,5kwh in 30 minutes?

Not an issue for me at all, I am getting some clamp readers anyway, but let me know if I can help

Cheers,

[dan-r]

Sadly this is a limitation of the API. The readings it provides just aren't particularly accurate. They often appear to be around 10% low but seem to vary too much to add a simple correction

[PauloGoncalves86]

Maybe they have some simplification related to voltage variability on their side. Thou checking their app today, I can see they calculate 12.1kwh for the same day.
It has been a hobby of mine to let them know about these holes. In fact, this was how I got into HA!

[dan-r]

The statistics on the Ohme app are correct, but the issue is that to update the energy dashboard we need to provide a realtime incrementing energy number, and these only update after you've unplugged. Because of this, the less accurate 'live' figures need to be used. The API provides an incrementing kWh figure, so there's no calculations on the integration side, and its all we can rely on sadly.

[PauloGoncalves86]

Thanks for clarifying.
I wonder if a workaround could be to add to the less accurate 'live' figures when unplugging happens. A tradeoff, the total EV usage figures on the energy dashboard would look right after a few hours - even if there could be an odd blip of consumption in the period when being unplugged (and even pushing it above the total home consumption).
Or (and I am not aware how the energy dashboard consumes it) if the energy dashboard could instead be fed updated figures for the EV retrospectively, for each 30 min period that was part of the charging session, it could be very close to accurate. So splitting the accuracy difference found (proportionally) between the relevant charging periods.