Cause the inverter is starting up. That only happens if there is “enough” light. And in the beginning there are some points of “just enough” (or not) light.
This is normal behaviour for an intverter.
This is not normal behaviour for an inverter. The integration needs to be fixed or the inverter settings need to be changed.
e.g. SMA inverter:
It is though. An inverter starts with light… and in the shade it sometimes starts but then stops and starts again. Same at the end of the day, it stops and starts a few times.
The pv panels just so not give enough to start and remain started.
Ik have at least 2 friends with other pv inverters (solis and solaredge) who have exactly this startup behavior.
Nevertheless, is there an option to keep the energy dashboard clean with this behavior?
Why is the inverter reporting yesterdays full energy total during these times it starts up intermittently?
Once it resets to 0 it can go on and off all it likes but it should still be reporting 0, or a very small amount at least. Not yesterdays total.
Nope. Garbage data in → garbage data out.
I suppose you could filter it with a template sensor but I’m not sure how.
Where to start to ask to “change this” in the esphome section for growatt modbus?
I used to use the custom component but now I use this code I need to aks the maintainer?
Same problem with and inverter:
Now I will try to reset the value to 0 every night in esp
time:
- platform: homeassistant
id: homeassistant_time
on_time:
# Every morning on weekdays reset ID todaygen
- seconds: 0
minutes: 30
hours: 1
days_of_week: MON-SUN
then:
- lambda: id(todaygen).publish_state(0);
This did not work. the inverter spits out bogus data in the startup phase. Alternately 0 and previous day gen until the inverter is completely running at 50Hz…
Ok so now I have another problem… the “spikes” are gone:
But I keep getting “spikes” in the energy dashboard. Now I also reset the value of the daily gen on 0:00 to 0 it shows me a negative spike:
What can I do to repair this?
Any help available here? Issue still persistent 
Hi Tom,
I tried your 2 tips:
-
the integration needs to be fixed:
I asked for help here:
Spikes with integration growatt_solar · Issue #3965 · esphome/issues · GitHub -
inverters settings need to be changed:
I contacted Growatt but I get no reply.
Can you help me on the way please? Is there a way to “filter out” via esphome the startup values?
Most likely in your sensor setup for esphome.
Indeed, but how?
well, start by posting your configuration for that esp device
I have now added some code I found on github. this “skips” the first 20 readings after ESP boot. But I doubt this will work because the ESP is already running when the inverter is starting up in the morning.
the current code:
esphome:
name: growatt_modbus
platform: ESP8266
board: d1_mini
on_boot:
# set a default value for the skip counter, otherwise it's nan
- sensor.template.publish:
id: ${my_sensor}_skips
state: 0
substitutions:
esp_name: Growattmodbus
my_sensor: my_sensor
wifi:
ssid: "SSID"
password: !secret wifi_key
use_address: 192.168.6.42
domain: !secret domain
# Enable fallback hotspot (captive portal) in case wifi connection fails
ap:
ssid: "${esp_name} fb Hotspot"
password: !secret Fallback_Hotspot
captive_portal:
# Enable logging
logger:
# level: VERBOSE
# baud_rate: 0
# Enable Home Assistant API
api:
ota:
time:
- platform: homeassistant
id: homeassistant_time
on_time:
# Every morning on weekdays reset ID todaygen
- seconds: 0
minutes: 0
hours: 0
#days_of_week: MON-SUN
then:
- lambda: id(todaygen).publish_state(0);
uart:
- id: uart1
baud_rate: 9600
tx_pin: D1
rx_pin: D2
# stop_bits: 1
modbus:
uart_id: uart1
web_server:
port: 80
sensor:
- platform: wifi_signal
name: "${esp_name} - ESP WiFi Signal"
update_interval: 60s
- platform: uptime
name: "${esp_name} - ESP Uptime"
icon: mdi:clock-outline
update_interval: 60s
- platform: growatt_solar
update_interval: 3s
protocol_version: RTU2
inverter_status:
name: "${esp_name} - Status Code"
id: inverter_status
phase_a:
voltage:
name: "${esp_name} - AC Voltage"
current:
name: "${esp_name} - AC Current"
active_power:
name: "${esp_name} - AC Power"
filters:
- filter_out: NaN
pv1:
voltage:
name: "${esp_name} - PV1 Voltage"
current:
name: "${esp_name} - PV1 Current"
active_power:
name: "${esp_name} - PV1 Power"
pv2:
voltage:
name: "${esp_name} - PV2 Voltage"
current:
name: "${esp_name} - PV2 Current"
active_power:
name: "${esp_name} - PV2 Power"
active_power:
state_class: "measurement"
name: "${esp_name} - Output Power"
filters:
- lambda: |-
if (x <= 5500 && x >= 0) return x;
else return 0;
#- lambda: |-
# if (x <= 5500) return x;
# else return {};
pv_active_power:
name: "${esp_name} - Input Power"
frequency:
name: "${esp_name} - Grid Frequency"
energy_production_day:
name: "${esp_name} - Today Gen"
id: todaygen
filters:
# increment a skip counter, this skips the first two measurements
- lambda: auto s = id(${my_sensor}_skips); if(s->state > 20) { return x; } else { s->publish_state(s->state+1); return {}; }
#filters:
# - lambda: |-
# if (x >= 0) return x;
# else return 0;
#lambda: |-
# if (id(inverter_status).state != 0) {
# return x;
# }
# return {};
# - lambda: "if (x - id(todaygen).state) > 0.5) return {}; return x;"
# https://github.com/esphome/feature-requests/issues/607
#- lambda: auto s = id(${todaygen}_skips); if(s->state > 100) { return x; } else { s->publish_state(s->state+1); return {}; }
total_energy_production:
name: "${esp_name} - Total Gen"
#state_class: "measurement"
#device_class: energy
accuracy_decimals: 1
inverter_module_temp:
name: "${esp_name} - Temperature"
# create a skip counter
- platform: template
id: ${my_sensor}_skips
internal: true
accuracy_decimals: 0
This is the original “problematic” code:
esphome:
name: growatt_modbus
platform: ESP8266
board: d1_mini
substitutions:
esp_name: Growattmodbus
wifi:
ssid: "SSID"
password: !secret wifi_key
use_address: 192.168.6.42
domain: !secret domain
# Enable fallback hotspot (captive portal) in case wifi connection fails
ap:
ssid: "${esp_name} fb Hotspot"
password: !secret Fallback_Hotspot
captive_portal:
# Enable logging
logger:
# level: VERBOSE
# baud_rate: 0
# Enable Home Assistant API
api:
ota:
time:
- platform: homeassistant
id: homeassistant_time
on_time:
# Every morning on weekdays reset ID todaygen
- seconds: 0
minutes: 0
hours: 0
#days_of_week: MON-SUN
then:
- lambda: id(todaygen).publish_state(0);
uart:
- id: uart1
baud_rate: 9600
tx_pin: D1
rx_pin: D2
# stop_bits: 1
modbus:
uart_id: uart1
web_server:
port: 80
sensor:
- platform: wifi_signal
name: "${esp_name} - ESP WiFi Signal"
update_interval: 60s
- platform: uptime
name: "${esp_name} - ESP Uptime"
icon: mdi:clock-outline
update_interval: 60s
- platform: growatt_solar
update_interval: 3s
protocol_version: RTU2
inverter_status:
name: "${esp_name} - Status Code"
id: inverter_status
phase_a:
voltage:
name: "${esp_name} - AC Voltage"
current:
name: "${esp_name} - AC Current"
active_power:
name: "${esp_name} - AC Power"
filters:
- filter_out: NaN
pv1:
voltage:
name: "${esp_name} - PV1 Voltage"
current:
name: "${esp_name} - PV1 Current"
active_power:
name: "${esp_name} - PV1 Power"
pv2:
voltage:
name: "${esp_name} - PV2 Voltage"
current:
name: "${esp_name} - PV2 Current"
active_power:
name: "${esp_name} - PV2 Power"
active_power:
state_class: "measurement"
name: "${esp_name} - Output Power"
filters:
- lambda: |-
if (x <= 5500 && x >= 0) return x;
else return 0;
#- lambda: |-
# if (x <= 5500) return x;
# else return {};
pv_active_power:
name: "${esp_name} - Input Power"
frequency:
name: "${esp_name} - Grid Frequency"
energy_production_day:
name: "${esp_name} - Today Gen"
id: todaygen
filters:
- lambda: |-
if (x >= 0) return x;
else return 0;
#lambda: |-
# if (id(inverter_status).state != 0) {
# return x;
# }
# return {};
# - lambda: "if (x - id(todaygen).state) > 0.5) return {}; return x;"
# https://github.com/esphome/feature-requests/issues/607
#- lambda: auto s = id(${todaygen}_skips); if(s->state > 100) { return x; } else { s->publish_state(s->state+1); return {}; }
#state_class: total
#state_class: total_increasing # was empty
# https://developers.home-assistant.io/docs/core/entity/sensor/#how-to-choose-state_class-and-last_reset
total_energy_production:
name: "${esp_name} - Total Gen"
#state_class: "measurement"
#device_class: energy
accuracy_decimals: 1
inverter_module_temp:
name: "${esp_name} - Temperature"
which sensor is outputting 0 or NaN on startup that’s causing the issues?
But its outputing:
You’re most likely running into a timing issue. The last calculation of the day is slightly past midnight by the time it reaches HA, so HA adds it to the next days values. Your only option is to control the timing of your data points, by forcing the values to specific intervals.
Or have HA calculate the energy from your power entities.
Ok, but how can I solve that? Get the first bugus readings out every day?