I finally managed to finish and test the 3-phase setup and it just works fine. Currently I’m using 3 VTs on one IoTaWatt device. I compared the IoTaWatt outputs with the built-in display of my 3-phase solar inverter several times and it matches almost perfectly (few watts difference at very low power). This could be due to the fact that I use 100A clamps in the low range.
However, for some time, the graph does not come up correctly on my screen. It loads for a very long time (20 sec) and has no feed list and shows an empty graph. Could you please advise?
Assuming you mean the local graph app. There is a known problem with spaces (or other non URL compatible characters) in the names of the inputs and/or outputs that can cause this problem. That problem was due to a fault in the config app that allowed it to happen. Newer versions of the app do a better job of checking the names, but don’t correct existing names except to the extent they won’t allow saving an edit without correcting the name.
You can check that yourself, or post a shot of your status screen and I’ll take a look.
Bob, I checked it and from what I can see there are no special characters in the names, and it used to work for weeks. Or I’m missing out on something.
I put the logs here as well, because I see some self resets, and some NTP sync problems. Yesterday is less relevant (the 26th), because I relocated the device, its power supply and also the VTs, which caused some errors, but the days before and today (the 27th) I see time skew, and update check failures.
I see that you are running 02_02_29. Could you set your auto-update class to MAJOR to receive the 02_02_30 release? I think it will solve the problem.
Regarding the clock skew, IoTaWatt only needs to know about what second it is, so I opted for a Timex instead of a Rolex. The clock is temperature sensitive and runs +/- 1 second/day. With the NTP update , that’s not a problem.
You will need to have decent WiFi and internet to do the update to 02_02_30. The firmware in development has more robust data communications.
Try to get that update done, and let me know if the graph problem persists.
That’s great. Could you try adding it back and see if it fails again? If so, could you upload your config.txt file so I can try to recreate.
This appears to be a problem with a library that I use to build Json responses. It is version controlled and you are running a six-month old version in 02_02_30. I’ll be upgrading to the latest with the next release, and maybe that fixes the problem.
I would consider that fortunate. Deleting your output is not a solution. If I can recreate it, I can fix it (or workaround).
you can download the config.txt file using the “File Manager and Editor” tool. Select it. The left column will be a list of the file system on your SD. You will see CONFIG.TXT among the root files. Right click on it and select download. Retrieve the file from your download folder and either:
Post it, redacting any server credentials, or
Send it to me in a PM (private message - click my icon and select “message”).
I found the problem. Long story short, when the feed list is too long, it doesn’t work. It’s not the number of feeds, but the total size of the list. In your case, the names are long, so it exceeds the limits. You got lucky removing that on output. The gap must have appeared in a non sensitive portion of the Json.
In the short term, you could reduce the size of the names or wait a few more days and take the next ALPHA release which resolves the issue. I was able to recreate the problem and made some changes to the development release. Your configuration is graphing fine here (ignore the data, it’s meaningless)
All right, thanks. Shall I leave the update in major, or switch back to minor?
Once we kinda solved this I’d have another issue:-) I’d like to make a data cleanup, because last year I made some tests and there’s been a lot of junk data stored. How can I delete all data before the 01.01.2018?
Sorry, I missed this inquiry. ALPHA is for users who would not have a problem downloading and installing a new firmware with one of several tools like ESPTOOL.PY should the wheels fall off the wagon. The sequence that I have been following is to run a new version on my three local systems for a day or two. If absolutely no problems, I publish to ALPHA. If no problems with the ALPHA community after a few days, I will push to BETA. MINOR is when I designate it an official release, and MAJOR is after MINOR has proven solid.
So if you think you can download and flash the ESP8266, you can go with ALPHA. I can always use more in that category. But if you would need help with the flashing beyond the instructions available with say ESPTOOL online, then I would recommend you wait a few more days for BETA to reduce risk.
FYI the new release done and is going into local test today.
