Weird spikes in Enphase power production

[friso82]

Hi,

I connected my Raspberry Pi Domoticz install to my local Enphase Envoy gateway. P1 meter was automatically added and results are registered. Sometimes I see some weird spikes.

Any tips to solve / debug?

Thanks!

[waltervl]

What method do you use to read the values? Plugin, script, MQTT etc?
Try to check the forum topic about this Enphase method. Without the method it will be difficult to help you.

[friso82]

I am using the built in functionality called “Enphase Envoy with LAN (HTTP) interface”.

[waltervl]

It seems to be a communication issue with the server that is being compensated a little bit later. Do you have an Enphase website to check the values?

[friso82]

Yes, those results are fine. I think it has to do with the fact that Domoticz saves data using 5 minute intervals and the Enphase Envoy gateway polls the inverters every 5 minutes. For most of the times, this works out just fine, but as soon as the Enphase gateway has a bit of a delay in a polling period, there will be a moment when the 5 minute interval of Domoticz is over and no update has been received from the Enphase gateway, resulting in a production of 0. However, in the next Domoticz interval, the gateway reports twice.

Not sure how to fix this by the way... I cannot change the polling / reporting settings from the Enphase gateway and cannot change the interval frequency of the Domoticz data logging...

[FireWizard]

Hi @friso82,

You wrote,

I think it has to do with the fact that Domoticz saves data using 5 minute intervals and the Enphase Envoy gateway polls the inverters every 5 minutes.

Domoticz writes the values from sensors to its log every 5 minutes, but I'm not sure about your statement that the Envoy S or Envoy S metered polls every 5 minutes. Yes, it is documented to be 5 minutes, but I have seen faster updates, so I do not know, if the update frequency is time-driven or event-driven.
See the Technical Brief, I posted at: viewtopic.php?p=224674#p224674 (September 25, 2019 at 22:01)

For most of the times, this works out just fine, but as soon as the Enphase gateway has a bit of a delay in a polling period, there will be a moment when the 5 minute interval of Domoticz is over and no update has been received from the Enphase gateway, resulting in a production of 0.

The Enphase "hardware" polls every 30 seconds and not every 5 minutes. I never saw a value of 0 (except at night) between two consecutive values.
If it has not been updated with a new value the previous value is used.
The Enphase "Hardware" uses as URL: http://<Envoy IP Address>/production.json.
You can check this at: https://github.com/domoticz/domoticz/bl ... aseAPI.cpp at line 159.

This URL is an unofficial endpoint. The official endpoint, as documented by Enphase, is http://<Envoy IP Address>/api/v1/production.
See also the above mentioned Technical Brief

You can check the endpoints in a browser of your choice

The result between the two different API's do not differ much, as expected. From time to time I see a spike as well and I see these both on the unoffical API (Enphase hardware) and on the official API. which I use also to send the data by MQTT to Domoticz.
I see that spike even on the logging of each individual inverter. So I do not think it is related to the Enphase hardware in Domoticz.

Could it be that you have some interference with other equipment, such as switching of the Q-relay?

[Eddiever]

You do have a single phase Envoy system? Are there any ferrite rings installed in the switchboard?

[FireWizard]

Hello @Eddiever,

I do not know to whom you raised your questions, but anyhow I will answer.

You do have a single phase Envoy system?

No, I have a 3 phase system, where the solar panels are equally divided over the 3 phases.

Are there any ferrite rings installed in the switchboard?

No, in the switchboard are no ferrite rings installed.
However the network cable, connected to the Envoy-S, is protected by a ferrite bead.
The power cord is standard.

Regards

[friso82]

I noticed that a spike happens every day at 10:20.

I have a single phase system. As far as I know there are no ferrite rings installed.

[FireWizard]

Hi, @friso82.

If it happens every day consequently at 10:20 a.m. I suggest that you look to all your log-files and check, what is happening also at 10:20 a.m.
As far as I know there is nothing in the Enphase hardware, that may invoke such a spike.

Regards

[Eddiever]

@friso82,

I can't explain why the spikes occur on 10:20 am. But I do know that those ferrite rings prevent unusual behaviour of the envoy - s. They are to be placed between the circuit breaker and the envoy-s.


But you can always call Enphase, they are very helpfull (to me so far). Try to call them on 073-3035859

[friso82]

I decided to apply an alternative approach using the method explained by @waaren in this topic. This results in much nicer charts. In the image below, the first chart is using the built-in method and the second chart is using the alternative method. I use NodeRed to read out the values from the Enphase Envoy and send them to Domoticz using Mosquitto.

Works fine for me this way, but leaves me with two questions:
- Why does the built-in method use the 'unofficial' API url and not the URL /api/v1/production url?
- Why is the data in the built-in method presented as P1 meter and not as 'Electric (Instant+Counter)' which can be set to type 'Return' and then even displays PV-icon.

builtin-vs-nodered.jpg (32.8 KiB) Viewed 1245 times