Grid meter installed at AC-In (per-phase measurement)
Problem description
Every few weeks, I observe a situation where the system feeds power into the grid very unevenly across phases, although it is configured as a 3-phase ESS system.
Example (from grid meter):
L1: ~ -1000 W
L2: ~ -3300 W
L3: ~ -800 W
Total feed-in is correct, but distribution across phases is clearly wrong.
Important observations
There are no AC-coupled PV inverters
AC-In is directly connected to the grid
All loads are behind AC-Out
During the issue, loads are low and relatively balanced (~500 W total)
It appears that one phase is doing most of the inverter work, while the others are underutilized.
Key point
The issue disappears immediately after:
-> Turning the MultiPlus system OFF and ON again (via software)
After restart, feed-in becomes properly balanced again (~equal per phase).
Conclusion / suspicion
This looks like a temporary VE.Bus / ESS state issue, where:
the 3-phase load sharing is no longer working correctly
but no alarm or error is shown
and the system continues operating
Questions
Has anyone observed similar behavior in a 3-phase ESS setup?
Is this a known issue in recent Venus OS / VE.Bus versions?
Are there recommended firmware versions or settings to avoid this?
Any ideas how to capture useful diagnostics when this happens?
Multiphase / Total of all Phases will make all inverters operate at the same inverterpower (roughly). Consumption and acpv may introduce AC inbalance, and the grid is “punished” for it (yet, it’ll be a saldated 0 as total)
In Individual mode, any phase will zero it’s phase-power (at the grid) individually, it may even move power through the dc bus from one phase to another to try to achieve that.
What it will do, if all phases have a (different) Feedin, I don’t know.
Yet, if a reboot “fixes” that, it is something else happening.
Make sure you configured the ESS Assistant through VEConfigure on all 3 units. If you missed that on one unit, they don’t really operate like expected. One Unit for example may just do whatever it wants based on own measurements, and the other two adapt around it to compensate.
Multiphase / Total of all Phases will make all inverters operate at the same inverterpower (roughly). Consumption and acpv may introduce AC inbalance, and the grid is “punished” for it (yet, it’ll be a saldated 0 as total)
Yes, that’s what I was expecting and this is also how it works 99% of the time. But occasionally this imbalance as shown in my first screenshot happens and I don’t know why.
Make sure you configured the ESS Assistant through VEConfigure on all 3 units. If you missed that on one unit, they don’t really operate like expected. One Unit for example may just do whatever it wants based on own measurements, and the other two adapt around it to compensate.
I actually reconfigured all Multiplus-II devices from scratch two weeks ago and really made sure that the configuration was identical. Also all devices received the same ESS assistant + configuration.
What also concerned me yesterday was that the L2 Multiplus-II got really hot with the fans running at high speed because it was doing most of the work, while the other two Multiplus-II were mostly idling.
Had a quick chat with our DC Expert. When there is dcpv excess, there is actually no rate-based balancing happening. The inverters will use the same voltage setpoint, which should lead to a quite equal discharge to AC then.
Tiny differences based on different AC voltage, but not that huge.
So, what you should check:
Is your L2 AC voltage significant different compared to L1, L3?
are the DC connections of the multipluses “equal”? Length, diameter and overall resistance?
If the multiplus on L2 would have a significant lower resistance to DC, it’ll discharge more than the others.
How old is your installation, was this the case from the beginning, or just started recently? Sometimes fuses and breakers can go bad, but still conducting at a higher resistance (so, should eventually swap them between phases and see if the discharge remains L2 focused or moves to another phase)
Thank you for checking back. Sometimes the AC voltages differ a bit between phases. But only maybe 2-3 volts. This should not lead to such a huge difference in watts, right? And according to my screenshot, L2 even has the highest voltage but additionally has 3-4 times the amount of amps.
I was thinking about the resistance as well. But I think it’s not always L2. I will definitely keep an eye on that.
But again: The issue resolves after restarting the Multipluses. That should rule out any voltage differences and cabling/resistance issues, right?
I will do more tests and will write as soon as I have more information.
Yeah, that’s basically right. Unless it is a difference that not causes it “right away” but more like a continious runaway, chain-reaction of some sort.
And when restarting, the inverters have to jump from 0 to 5000W quicker, so they do that quite equally.
When this happens next time, you could try if disabling dc feedin and reenabling it has the same “clearence” effect on the distribution.
For the ac side, that minor difference should not be a problem, i’d say.
Does this phase imbalance only occur at 100% SOC? I am experiencing a similar imbalance, but only when the battery is reaching full SOC. It isn’t restricted to one specific phase, as the issue is intermittent.
Yes, that’s the case. Below 100% everything seems to be fine, but in that case the energy is transferred to the battery anyway and the Mutlipluses just compensate for the current AC-load.
I can follow what you mean and your loads definitely do it symmetric.
it is strange this is happening though in our setups the cable lengths are equal.
Even not pinning on a fixed phase that is showing the behavior.
More strange is that all is working symmetrical but not when battery is getting full.
How strange is that? If there would be some symmetrical problem by cabling or hardware aging it would occur not only with charging and battery almost full.
Perhaps It has with the combination AC PV to do? Hm i have AC PV but Marc has not…
Just to clarify: this behavior isn’t consistent on a single phase?
Just to throw out an idea: it could be the internal AC backfeed relays and/or your MCBs protecting the AC side of the system.
While I didn’t have this exact same issue, I’ve learned those two things can cause these sporadic-type problems.
I had to change from MCBs to fuses to resolve most of my imbalance problems. I’m still dealing with a problem on one of my units, where the internal AC relay “closes” slightly differently on occasion causing my units to go out of balance. This is a parallel config so not exactly the same, but figured I’d mention.
Check this, i did set Grid current limit to 2A and grid keeps smooth symmetrical while batt is charging.
The problem starts as soon there is more PV than needed for charging and there is no feed-in restriction and the battery is slowly charging to full SOC.
No, that is not the case: When the battery is NOT full, then the gridsetpoint is a target the units are working on. Thus they will actively adjust their invertingpower to meet the gridsetpoint.
Invertingpower then is “rate controlled”.
When the battery is full, there will be feedin on any phase - and the setpoint doesn’t matter. Then the multis apply the same dc (dis-)charge voltage and “just let it flow”.
Invertingpower then just is “delta voltage controlled” so to say. If a unit is getting to hot, it will raise it’s dc-voltage a bit to “unload”. If all units do that (cause too much dcpv), then the resulting current / power discharged will just lower, implicit mppt throttling.
And that brings me to an idea: When you have that uneven situation, but a total of 5000 feedin (example value) - what happend, when you change the grid setpoint to -5100? That should make the setpoint algorithm active, and equalize feedin - and if it does, what happens, when you revert to 0 afterwards? Do the units derail immediately or slowly over time?
(I could not post more separate messages so i edited this post)
Narrowed the ‘issue’ a bit more down in the case of charging.
The time frame is not that long at the end of the day and the battery needs to be almost full and than test different settings but again all tests involve charging the battery.
Settings:
Optimized without BatteryLife
DESS turned off
Multiphase regulation Individual phase
Venus OS 3.72 and 3.80~14
After the batteries getting up to 90% then there is les energy going in.
Grid Measurement (in sum) gets lower and lower as the batteries take less energy.
The Imbalance starts to occur at some point where the Gridsetpoint is higher than the Grid Measurement, the MPPT power is reduced too shortly after.
After adjusting the Gridsetpoint down to a lower value than Grid Measurement shows than the MPPT produces more power again en the Grid phases are balancing again.
The behavior is reproducible.
The last test settings (VRM app screenshot 18:19)
Keep batteries charged
Gridsetpoint 0
The MPPT power gets reduced and after a short time Grid Imbalance occurred.
I just tried this and the problem resolves immediately!
Yesterday I had a less significant imbalance but this time L1 was producing more power than the others. I don’t think it’s a cabling/impedance issue.
Edit:
I just made an observation: After disabling DC feed-in and re-enabling it again, the imbalance was gone. But after a couple of minutes I had a high load on L2 which was somehow compensated by the Multiplus on L2 even though it was not really necessary, since the total load on L1+L2+L3 was lower than the excess PV power. BUT the load on L2 was now higher than 1/3 of the total excess PV, moving the flow of power on L2 from feed-in to consumption. After the high load on L2 was gone, the unnecessary compensation on L2 was still there and after that I had the imbalance.
It appears that one phase is doing most of the inverter work, while the others are underutilized.
