evcc-io
Fast/Now chargin with loadmanagement does not switch to 3p
Describe the bug
I have the following issue
- Defined a curcuit with a limitation to 4000W (as my provider charges high peaks)
- My solar surplus is around 2500W
- I expected evcc to swith to 3 phase and charge my car with 6500w in now/fast mode. But it stays at 1 phase and is therefore limited at around 3500W
No battery is involved. Additional observation: If I raise the circuit to 5000W the switch to 3 phases seems to work.
Steps to reproduce
- Add a circuit with a limit of 4000W
- Make the sun shine so you have some solar surplus :-)
- Enable fast/now mode
Configuration details
circuits:
- name: main
title: 'Main Circuit'
meter: huawei.grid
maxPower: 4000
Log details
[site ] DEBUG 2025/10/06 13:04:17 ----
[lp-1 ] DEBUG 2025/10/06 13:04:17 charge power: 0W
[lp-1 ] DEBUG 2025/10/06 13:04:17 charge currents: [0 0 0]A
[circuit-main] DEBUG 2025/10/06 13:04:17 power: -2409W
[circuit-main] DEBUG 2025/10/06 13:04:17 current: 0A
[site ] DEBUG 2025/10/06 13:04:17 grid power: -2409W
[site ] DEBUG 2025/10/06 13:04:17 pv 1 power: 2971W
[site ] DEBUG 2025/10/06 13:04:17 !! solar production: accumulate add huawei.pv 0.000W power (was: Accumulated: 1093.930kWh updated: 0001-01-01 00:00:00 +0000 UTC)
[site ] DEBUG 2025/10/06 13:04:17 !! solar production: accumulate moved huawei.pv from 1093.930 to 1093.930
[site ] DEBUG 2025/10/06 13:04:17 site power: -2409W
[lp-1 ] DEBUG 2025/10/06 13:04:17 vehicle odometer: 2087km
[lp-1 ] DEBUG 2025/10/06 13:04:17 charge voltages: [230 231 232]V
[lp-1 ] DEBUG 2025/10/06 13:04:17 charge total import: 286.807kWh
[lp-1 ] DEBUG 2025/10/06 13:04:17 charger status: B
[lp-1 ] DEBUG 2025/10/06 13:04:17 charge start energy: 286.807kWh
[lp-1 ] INFO 2025/10/06 13:04:17 car connected
[lp-1 ] DEBUG 2025/10/06 13:04:17 pv timer elapse
[lp-1 ] DEBUG 2025/10/06 13:04:17 pv timer inactive
[lp-1 ] DEBUG 2025/10/06 13:04:17 charge total import: 286.807kWh
[lp-1 ] DEBUG 2025/10/06 13:04:18 vehicle status: B
[lp-1 ] DEBUG 2025/10/06 13:04:18 vehicle soc: 56%
[lp-1 ] DEBUG 2025/10/06 13:04:18 vehicle soc limit: 80%
[lp-1 ] DEBUG 2025/10/06 13:04:18 vehicle range: 286km
[lp-1 ] DEBUG 2025/10/06 13:04:32 set charge mode: now
[lp-1 ] DEBUG 2025/10/06 13:04:32 pv timer reset
[lp-1 ] DEBUG 2025/10/06 13:04:32 pv timer inactive
[site ] DEBUG 2025/10/06 13:04:32 ----
[lp-1 ] DEBUG 2025/10/06 13:04:32 charge power: 0W
[lp-1 ] DEBUG 2025/10/06 13:04:32 charge currents: [0 0 0]A
[circuit-main] DEBUG 2025/10/06 13:04:32 power: -2409W
[circuit-main] DEBUG 2025/10/06 13:04:32 current: 0A
[site ] DEBUG 2025/10/06 13:04:32 grid power: -2409W
[site ] DEBUG 2025/10/06 13:04:32 pv 1 power: 2945W
[site ] DEBUG 2025/10/06 13:04:32 !! solar production: accumulate add huawei.pv 0.000W power (was: Accumulated: 1093.930kWh updated: 2025-10-06 13:04:17 +0200 CEST)
[site ] DEBUG 2025/10/06 13:04:32 !! solar production: accumulate moved huawei.pv from 1093.930 to 1093.943
[site ] DEBUG 2025/10/06 13:04:32 site power: -2409W
[lp-1 ] DEBUG 2025/10/06 13:04:32 vehicle odometer: 2087km
[lp-1 ] DEBUG 2025/10/06 13:04:32 charge voltages: [229 231 232]V
[lp-1 ] DEBUG 2025/10/06 13:04:32 charge total import: 286.807kWh
[lp-1 ] DEBUG 2025/10/06 13:04:32 charger status: B
[lp-1 ] DEBUG 2025/10/06 13:04:32 fast charging: scaled to 1p to match 4000W max circuit power
[lp-1 ] DEBUG 2025/10/06 13:04:33 switched phases: 1p
[lp-1 ] DEBUG 2025/10/06 13:04:33 wake-up timer: start
[lp-1 ] DEBUG 2025/10/06 13:04:33 set charge current limit: 16A
[lp-1 ] DEBUG 2025/10/06 13:04:33 charger enable
[lp-1 ] DEBUG 2025/10/06 13:04:33 wake-up timer: start
[site ] DEBUG 2025/10/06 13:04:38 ----
[lp-1 ] DEBUG 2025/10/06 13:04:38 charge power: 0W
[lp-1 ] DEBUG 2025/10/06 13:04:38 charge currents: [0 0 0]A
[circuit-main] DEBUG 2025/10/06 13:04:38 power: -2409W
[circuit-main] DEBUG 2025/10/06 13:04:38 current: 0A
[site ] DEBUG 2025/10/06 13:04:38 pv 1 power: 2945W
[site ] DEBUG 2025/10/06 13:04:38 !! solar production: accumulate add huawei.pv 0.000W power (was: Accumulated: 1093.943kWh updated: 2025-10-06 13:04:32 +0200 CEST)
[site ] DEBUG 2025/10/06 13:04:38 !! solar production: accumulate moved huawei.pv from 1093.943 to 1093.947
[site ] DEBUG 2025/10/06 13:04:38 grid power: -2409W
[site ] DEBUG 2025/10/06 13:04:38 site power: -2409W
[lp-1 ] DEBUG 2025/10/06 13:04:38 charge voltages: [229 231 232]V
[lp-1 ] DEBUG 2025/10/06 13:04:38 charge total import: 286.807kWh
[lp-1 ] DEBUG 2025/10/06 13:04:38 charger status: B
[lp-1 ] DEBUG 2025/10/06 13:04:38 fast charging: scaled to 1p to match 4000W max circuit power
[site ] DEBUG 2025/10/06 13:04:58 ----
[lp-1 ] DEBUG 2025/10/06 13:04:58 charge power: 3359W
[lp-1 ] DEBUG 2025/10/06 13:04:58 charge currents: [14.8 0 0]A
[circuit-main] DEBUG 2025/10/06 13:04:58 power: -2408W
[circuit-main] DEBUG 2025/10/06 13:04:58 current: 0A
[site ] DEBUG 2025/10/06 13:04:58 grid power: -2408W
[site ] DEBUG 2025/10/06 13:04:58 pv 1 power: 2928W
[site ] DEBUG 2025/10/06 13:04:58 !! solar production: accumulate add huawei.pv 0.000W power (was: Accumulated: 1093.947kWh updated: 2025-10-06 13:04:38 +0200 CEST)
[site ] DEBUG 2025/10/06 13:04:58 !! solar production: accumulate moved huawei.pv from 1093.947 to 1093.964
[site ] DEBUG 2025/10/06 13:04:58 site power: -2408W
[lp-1 ] DEBUG 2025/10/06 13:04:58 charge voltages: [227 232 232]V
[lp-1 ] DEBUG 2025/10/06 13:04:58 charge total import: 286.815kWh
[lp-1 ] DEBUG 2025/10/06 13:04:58 charger status: C
[lp-1 ] DEBUG 2025/10/06 13:04:58 charge start energy: 286.815kWh
[lp-1 ] INFO 2025/10/06 13:04:58 start charging ->
[lp-1 ] DEBUG 2025/10/06 13:04:58 wake-up timer: stop
[lp-1 ] DEBUG 2025/10/06 13:04:58 soc estimated: 56.00% (vehicle: 56.00%)
[lp-1 ] DEBUG 2025/10/06 13:04:58 vehicle soc: 56%
[lp-1 ] DEBUG 2025/10/06 13:04:58 vehicle soc limit: 80%
[lp-1 ] DEBUG 2025/10/06 13:04:58 vehicle range: 286km
[lp-1 ] DEBUG 2025/10/06 13:04:58 fast charging: scaled to 1p to match 4000W max circuit power
What type of operating system or environment does evcc run on?
HomeAssistant Add-on
External automation
- [x] I have made sure that no external automation like HomeAssistant or Node-RED is active or accessing any of the mentioned devices when this issue occurs.
Nightly build
- [x] I have verified that the issue is reproducible with the latest nightly build
Version
0.209.0
This is due to https://github.com/evcc-io/evcc/pull/23571. We may need to revert or improve this.
Hi Are there any news on when this will be fixed? Currently fast/now charging with a curcuit is rather unpredictable und therefore unusable for me.
Thanks!
Citing from #24955
It is definitly related, although the causes are different:
In my case I have two circuits and Fast mode never goes to 1 phase charging In the other case, there is only one circuit and Fast mode never goes to 3 phase charging For 1), the cause is that the restriction of the parent circuit is not taken into account. For 2), the cause is that the additional power of the PV is not taken into account.
It will affect similar code, the method fastChanrging and surroundings.
I think we need a different approach for fastcharging compliance with load mangagement:
- validate power against all hierarchy levels
- preferably add a timer (can we re-use the PV timer?)
/cc @mfuchs1984
While still waiting for a fix for this issue, I made a rather disturbing experience with evcc, fast charging and load management.
I started a fast charging session with 5000w limit. This is enough to start the charging with 3 phases. That was about 21:00. At 22:00 I had a look at evcc and it was properly charging my car with ~4700W. The next morning I looked at the logs and was quite shocked. At around 2:00 in the morning evcc started to charge my car, ended up in overpowering, stop and started again.
It did this about 200 times in 2-minute intervals!!
No additional consumer was active at that time that could explain this behavior. I am really worried now because this is for sure not very good for the charging hardware of the car. I would really advise not to use loadmanagement in an unattended scenario anymore until this is fixed. I think aside from fixing the behavior itself evcc should implement some kind of safeguard to prevent this kind of stressing the charging hardware.
Here an extract of the log:
[lp-1 ] INFO 2025/12/02 04:12:06 start charging ->
[circuit-main] WARN 2025/12/02 04:12:46 over power detected: 7993W > 5000W
[circuit-main] WARN 2025/12/02 04:13:06 over power detected: 7993W > 5000W
[lp-1 ] INFO 2025/12/02 04:13:26 stop charging <-
[lp-1 ] INFO 2025/12/02 04:14:06 start charging ->
[circuit-main] WARN 2025/12/02 04:14:46 over power detected: 7998W > 5000W
[circuit-main] WARN 2025/12/02 04:15:06 over power detected: 7998W > 5000W
[lp-1 ] INFO 2025/12/02 04:15:26 stop charging <-
[lp-1 ] INFO 2025/12/02 04:16:06 start charging ->
[circuit-main] WARN 2025/12/02 04:16:46 over power detected: 8008W > 5000W
[circuit-main] WARN 2025/12/02 04:17:06 over power detected: 8008W > 5000W
[lp-1 ] INFO 2025/12/02 04:17:26 stop charging <-
[lp-1 ] INFO 2025/12/02 04:18:06 start charging ->
[circuit-main] WARN 2025/12/02 04:18:46 over power detected: 8007W > 5000W
[circuit-main] WARN 2025/12/02 04:19:06 over power detected: 8007W > 5000W
[lp-1 ] INFO 2025/12/02 04:19:26 stop charging <-
[lp-1 ] INFO 2025/12/02 04:20:06 start charging ->
[circuit-main] WARN 2025/12/02 04:20:46 over power detected: 8007W > 5000W
[circuit-main] WARN 2025/12/02 04:21:06 over power detected: 8007W > 5000W
[lp-1 ] INFO 2025/12/02 04:21:26 stop charging <-
[lp-1 ] INFO 2025/12/02 05:42:06 start charging ->
[circuit-main] WARN 2025/12/02 05:42:46 over power detected: 8012W > 5000W
[circuit-main] WARN 2025/12/02 05:43:06 over power detected: 8012W > 5000W
[lp-1 ] INFO 2025/12/02 05:43:26 stop charging <-
[lp-1 ] INFO 2025/12/02 05:44:06 start charging ->
[circuit-main] WARN 2025/12/02 05:44:46 over power detected: 8006W > 5000W
[circuit-main] WARN 2025/12/02 05:45:06 over power detected: 8006W > 5000W
[lp-1 ] INFO 2025/12/02 05:45:26 stop charging <-
[lp-1 ] INFO 2025/12/02 05:46:06 start charging ->
[lp-1 ] INFO 2025/12/02 05:46:30 stop charging <-
[lp-1 ] INFO 2025/12/02 05:47:06 start charging ->
[circuit-main] WARN 2025/12/02 05:47:46 over power detected: 7981W > 5000W
[circuit-main] WARN 2025/12/02 05:48:06 over power detected: 7981W > 5000W
[lp-1 ] INFO 2025/12/02 05:48:26 stop charging <-
@McFSR this needs a more verbose log, debug level. And your configuration details.
@mfuchs1984 Thanks for your post. Regarding verbose log: I could only create such a log by repeating the process - which is something I obviously do not want to do :-)
No log, no cure 😟. There‘s absolutely no point in posting „bad things happening“ without details about the bad things. It‘s just not actionable.
There is no other way to find out what was going on than to reproduce with proper logs. My guess is that the charge process was not even started by evcc, it just reacted. And even when, I am sure it did not instruct the charger to charge with more that the allowed power. But in any case, only a debug log will tell.
I understand of course. I will try to reproduce this behavior in an attended scenario so I can stop the charging after a few cycles. This will probably take some time...
@mfuchs1984 The charging process was started by evcc and worked properly for some hours until it started to show this behavior. But I know, no logs -> no proof :-)
Tried to generate a suitable log today, but could not reproduce the special "On/Off" behavior. I will give it another try tomorrow.
Meanwhile I looked at my config: What i realized is that my pv/grid meters that I fetch from home assistant via API are updated every 30 seconds in HA. My evcc instance is configured to do a refresh every 20 seconds. This means that in some situations evcc could get the same value for pv/grid twice from HA. Could that cause such a behavior? Is it recommended to increase my evcc refresh interval to e.g. 35 seconds so evcc gets a "fresh" value every time?
Using a slower interval in evcc in this case is a good idea, but I doubt that this caused the observed behavior.
Hi I was not able to reproduce the issue until now during the day. I don't dare to run it overnight though.
The only thing I found out: Shortly before the strange behavior started, the log revealed a single communication issue with the go-e charger (timeout). But just once. I check if after that communiation worked again and it seems all "start chargin" and "stop chargin" commmands properly arrived at the go-e charger as I could see it in the power consumption of my grid meter.
So for now lets pause this until I can find out more.
Are there any news regarding the initial topic of this thread: Loading management and phase shifting?