That’s the reason/difference. You got a TOTAL - for ONE phase. I got 3 totals, for 3 phases. But not no sum of all three. 
I did not know about that data “dump”, so thanks for that but is there really a COMBINED total? To me it looks like it’s just the total for one phase? Or?
It’s 3 seperate totals. Not one sum
That’s the output:
{"wifi_sta":{"connected":true,"ssid":"xxxxx","ip":"192.168.1.140","rssi":-60},"cloud":{"enabled":true,"connected":true},"mqtt":{"connected":false},"time":"14:31","unixtime":1630931518,"serial":xxxxx,"has_update":false,"mac":"DC4F22xxxxx","cfg_changed_cnt":0,"actions_stats":{"skipped":0},"relays":[{"ison":false,"has_timer":false,"timer_started":0,"timer_duration":0,"timer_remaining":0,"overpower":false,"is_valid":true,"source":"input"}],"emeters":[{"power":-779.61,"pf":-0.97,"current":3.49,"voltage":228.74,"is_valid":true,"total":2488676.3,"total_returned":56383.7},{"power":-539.79,"pf":-0.92,"current":2.56,"voltage":229.09,"is_valid":true,"total":5615129.3,"total_returned":30759.3},{"power":-920.07,"pf":-0.99,"current":4.00,"voltage":231.92,"is_valid":true,"total":1754564.7,"total_returned":50914.4}],"total_power":-2239.47,"fs_mounted":true,"update":{"status":"idle","has_update":false,"new_version":"20210720-190729/v1.11.0-g6abd92e","old_version":"20210720-190729/v1.11.0-g6abd92e","beta_version":"20210901-204310/v1.11.4-rc3-gf11bc53"},"ram_total":49464,"ram_free":30896,"fs_size":233681,"fs_free":154867,"uptime":1216198}
Yeah. You’re right. I missed the {index} when reading the API docs:
That will be 0,1, or 2 for the three separate phases.
Sorry for misleading you.
@Spitfire : If what you say is true then it has massive impact on the ENERGY panel as well, as it will give completly wrong numbers if done how I did it …
1 Like
The ShellyForHass integration has also a total, however I never got it to work very well so I went for the official integration. Do not know if they just added the total or made the SQRT(3) division.
However I think the official intergration should calculate the total at least for people using it measuring a 3 phase consumer/producer because pretty much everyone seems to have got it wrong (until someone proves otherwise)
Yes, that is how I did it also until I noticed that the output was totally wrong.
Yup, not something we can use directly.
I posted both (old & new method’s) values above. I think I’ll check my meter now and in one week and then we’ll see which method is the accurate one. Other ideas?
Yes, please do that. 
Hang on, there is a total of the three phases:
Without reading more than what you posted…Isn’t that just total POWER?
Right. I’m going to bed before I embarrass myself any further.
NP, 
It’s good to have some extra eyes. I might be very wrong myself. Good if someone could confirm what I think.
Any update on your measurements?
No. I get my new meter on Thursday. Until then all measurements are mixed up between consuming and feeding back. So I will start new comparison on Friday. We should have results this weekend.
Willl be really interesting to see your results. I’m now far from sure that you should divide with SQRT(3) to get the right consumption. Got an hourly .csv file from my electric company and checking the values during nighttime (no incoming solar power) it showed actually that just summing should be the way to go.
In my case I’m a little bit worried that my Solar inverter might mess something up since I got very bad power factor values according to 3EM (as low as 25% on one phase).
Have stumbled into this post which addresses the exact same issues I have on 3EM
- Peaks when HA restarts that send my total template up sky
- How to measure the energy returned to the grid
Would be great to have an update on your test
1 Like