How is this possible:
- Solar: 7.2 kW
- Battery charge: 4.7 kW
- AC loads: 2.9 kW
- Grid disconnected
- No other energy sources or loads connected
1 Like
If the temperature stays around 67°C, you need to take cooling measures to preserve the life and performance of the device, otherwise you will have more chances to see funny data like this.
The temperature reading is from raspbbery pi, not the inverter.
He told you correctly…
Any processor near its temperature limit will start to exhibit strange behaviors, especially on its arithmetic unit.
So your Cerbo like equipment (rPi) will start to improperly compute the info which is coming from the VMs.
1 Like
The typical operating temperature range for the BCM2711 is 0°C to 85°C.
The Raspberry Pi 4, which uses the BCM2711, starts thermal throttling at around 80°C to 85°C.
In normal conditions, operating the BCM2711 at 67°C should not cause errors, and there are safety mechanisms in place (like thermal throttling and standard error detection) to handle potential issues.
So temperature is not a problem.
I know it would be good to lower the temperature for longer lifespan.
The question is: How the critical load readings on the dashboard are calculated?
Take a look at the qml files used for displaying those information.
They are in the /opt/victronenergy/gui/qml folder and there, for the right device, look at what dbus path or variable they are using.
Then, as long as you know now from where it comes, see how is computed.
2 Likes
@Rafal I couldn’t get that video to play.
Critical loads are on AC OUT and metered by the inverter internally.
On a system without a meter and “has AC loads” enabled, it is just called AC loads.
For those who can’t play it, click to download it here and then play it locally on your device / computer by other means / codecs.
Here is what I have discovered so far:
Critical loads = system /Ac/Consumption/L[1,2,3]/Power or system /Ac/ConsumptionOnOutput/L[1,2,3]/Power
PV inverter = VM-3P75CT (PV Metter) /Ac/L[1,2,3]/Power
Battery power = Lynx Shunt
AC Loads = VM-3P75CT (AC Loads) - connected to Multi AC output
Grid = VM-3P75CT (Grid) - grid is disconnected from Multi, but the energy metter is operating beacause grid is available.
Now, the questions are:
- How are com.victronenergy.system /Ac/Consumption/L[1,2,3]/Power values calculated
- Why are the Critical loads more than twice the actual readings from AC loads energy metter?
- Why Critical loads are not equal AC load metter readings?
FWIK:
The power values (for a specific device) are taken directly from the device, that’s publishing them for all others to know.
More of that, they comply with NMEA standard and are also found on PGN 127744.
For example, in my case, the Multi RS is broadcasting them over CANbus and Cerbo is listening for them.
Once Cerbo gets them (in the vecan-dbus driver), it fills the proper dbus paths.
So short answer: they are calculated inside the device’s firmware.
The aggregate values are calculated in the dbus_systemcalc script, for each phase.
See that the /Consumption/ key is deprecated.
The future proof ones are /ConsumptionOnOutput/ and /ConsumptionOnInput/
1 Like
What role have you configured the meter as?
Where is it installed and what are you wanting to measure with it.
People incorrectly set a grid meter as an ac load meter when they have parallel loads.
Roles are set as follows:
Grid - installed between grid and AC input of Multis:
PV inverter - installed on PV Inverter Output which is connected to Multi AC1 output:
AC loads - installed on Multi AC1 Output for measuring AC loads:
Could there be an issue with the Consumption calculation in the Multi’s firmware?
Did you tried, for testing purposes, to have only one, resistive load?
No inductive, no capacitive loads…
Maybe the aggregate value is computed over the instant values provided by the VMs, not the averaged ones, and the way dbus is hammered with (different) values, it could give false results.
Ok, but how do you explain the difference between the energy measured shown by the VRM portal and the energy measured by the AC load meter?
These are the energy readings from the Victron AC load meter over a one-hour interval:
Edit:
Even though the discharged energy is smaller than the consumed energy in the VRM portal (based on energy readings from the Lynx shunt at hourly intervals):
The difference between battery discharge and energy consumption from the AC load meter is expected due to the energy conversion from DC to AC, but the readings from the VM are overestimated.
Consumption is a combination of loads on AC OUT and parallel to AC IN (via the grid meter).
@Rafal would you mind editing the title to make it more useful for someone else to find later, hopefully once the question is answered. Thanks.
Okay, then when the grid is disconnected, it should be equal to the loads on AC OUT—so why is it not equal to the energy meter installed on AC OUT, which actually measures the same?
I changed the title appropriately.
That is a fair assumption. It should only reflect the parallel loads that ran during the interval.
Assuming the meter is installed in the right location.
What Venus OS you running? 3.41?
The key thing to understand is if it is only inaccurate on VRM, or if the GX agrees with VRM - since it is the data source.
I still couldn’t see the full video.
If the GX is reporting incorrectly then why is that the case.
Before I can escalate this, you need to be 100% sure that the wiring is correct.
How are the meters connected to the Pi?
Meter firmware up to date via Victron connect?
Any customisations to Venus OS?