search

patience4711 / read-APSystems-YC600-QS1-DS3

Software for an esp8266 nodemcu to read data from APS inverters.
127 stars 24 forks source link

NEW VERSION 9_13 #114

Closed patience4711 closed 1 year ago

patience4711 commented 1 year ago

The latest version now is v9_13d Via this link you can visit my testenvironment to see how it looks. You can login as "user" and "2222"

I apologise in advance for eventual bugs but i have limited test opportunities, if have only one inverter of type yc600. For the other types i depend on the results of others who want to share their testing results.

The last revision is a security improvement, we cannot access the configuration pages from outside our own network. The menu link is hidden and doesn't work. Also viewing of the pages source is disabled. So we can open it to the world with more peace of mind. Also some bugs where corrected and the pairing process has been fine-tuned.

Please note: If you are currently using a version older than 9_11:

Before updating put off the auto polling. After updating re-pair your inverters or edit the id via the console: eg: if inverter 0 has the id = 0xA1B2 you can change this with 10;edit=0-B2A1. E.g. 0xA1B2 should become B2A1

I have it running for some days now and it seems rockstable. The log shows no zigbee resets. When there are no webclients, the heap is steady over 26000. No memory leakage,

swbouman commented 1 year ago

Did you just reset your device? Or did it crash?

I have difficulties connecting the webserver since 9.12 version.

ESP-ECU

Btw the mqtt is continuously reconnecting my mqtt server. Probably this causes the webserver problems. I do not experience esp reboots, although I can’t check free heap since version 9.12. My mqtt topic is homeassistant/inverter/

ESP-ECU

patience4711 commented 1 year ago

I uploaded a changed software. I already said before, there is something with the mosquitto. I also saw some disconnections this morning. Which is strange after 2 days of working problemless. I'l investigate..

swbouman commented 1 year ago

I’m sorry I missed that message. I already looked into my routers ;) I’ll install the new version.

Op 15 mei 2023, om 20:33 heeft patience4711 @.***> het volgende geschreven:

I uploaded a changed software. I already said before, there is something with the mosquitto. I also saw some disconnections this morning. Which is strange after 2 days of working problemless. I'l investigate..

— Reply to this email directly, view it on GitHub https://github.com/patience4711/read-APSystems-YC600-QS1-DS3/issues/114#issuecomment-1548359880, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIPVFQPRJBOSGKA3JP6BVXDXGJZJNANCNFSM6AAAAAAYB5X4E4. You are receiving this because you commented.

patience4711 commented 1 year ago

@swbouman I meant that I installed something on my esp that caused it to reboot. Anyway i cannot see what is wrong with the mqtt, its all implemented as it should.

When the mqtt connection fails, it tries to reconnect once in 5 seconds, this shouldn't bother the wifi but I changed that period to 10 sec's. I Also updated the pubsub library and changed the keepalive value (from 15 to 150 secs). I cannot think of anything else i could try. It is running now, must wait for results.

If you want to test yourself here it is

swbouman commented 1 year ago

I think it has something to do with the zigbee. It's impossible to pair the inverters anymore. The program changes the inverter type to YC600 and the inverter ID is always 1111. The program is not interupted and hangs on the pairing progress..

sendSB sent FE1824020FFFFFFFFFFFFFFFFF14FFFF140C0201000F060070400030337

analyzing paircmd readZB nothing received no pairing answer pair command = 24020FFFFFFFFFFFFFFFFF14FFFF140F0102000F1100704000303A3D810FFFF80971B01A3D8 sending paircmd 2 sendSB sent FE2324020FFFFFFFFFFFFFFFFF14FFFF140F0102000F1100704000303A3D810FFFF80971B01A3D879 analyzing paircmd readZB nothing received no pairing answer pair command = 24020FFFFFFFFFFFFFFFFF14FFFF14010103000F060080971B01A3D8 sending paircmd 3 sendSB sent FE1A24020FFFFFFFFFFFFFFFFF14FFFF14010103000F060080971B01A3D84F success, inverter got id 1111 sending N.O. cmd sendSB sent FE162401FFFF1414060001000F1E80971B01A3D8FBFB1100000D4F readZB FE0164010064FE034480001401D2 rc=14 sendSB sent FE00270027 readZB FE0E670000FFFF80971B01A3D8000007090011 rc=19 found 0709, running oke

patience4711 commented 1 year ago

@swbouman What version are we talking about now? You get no answers on the pairing requests. This is probably the old range problem. The 1111 is an indication of failure, not success, so there is something wrong indeed. But if you would get answers, the pairing works oke. Why are you pairing, normally no need for? If you know the id you can edit this via the console.

I have tested the pairing a thousand times and it worked better than ever. The thing is that i cannot reproduce the error because I always get answers on the pairing. But i found the bug, there was a return(50) on a bool function. I corrected that but this only comes into the play if there are no pairing answers. So the real problem is still the lack of answers.

I found another bug that wrote a too long string to the inverter id, this changed the invertertype. Never saw that because my inverter is a yc600.

I corrected that too

swbouman commented 1 year ago

@patience4711 I discovered a problem with 9.13 and the pairing. I needed to refresh the esp due to getting unresponsive behavior. I did get an answer but it was allways the first 4 chars of the ECU ID what was written as correct inverter ID. After editing the correct inverter id, which was displayed in the console output during the pairing, the inverter type changed to YC600 instead of the correct DS3 type. It took me a long time to get them all in line again.

After a day of functioning the ESP-ECU webinterface is unresponsive again :( I saw a free heap of 13000 which should be high enough but still a lot lower than the 27000 at reboot.

So two bugs:

Ill try 9.13b

joffito commented 1 year ago

yeah i can confirm the switch to YC600. but i like the updated webinterface ! still stable after 5 h ontime.

patience4711 commented 1 year ago

@swbouman indeed edit did the same, writing a too long string. This is also solved now in version 9_13c

swbouman commented 1 year ago

Thanks! @patience4711 !

The unresponsive behavior is probably due to a too busy fixed wifi channel. I changed the channel and the esp is responsive again.

I'll test the EDIT string on version 9.13c

swbouman commented 1 year ago

@patience4711 with ver 9.13 this is the output (for as far as I could reproduce it from the console)

`

found invID in MessageToDecode

found invID = A1BA pair command = 24020FFFFFFFFFFFFFFFFF14FFFF140F0102000F1100704000161600A3D810FFFF80971B01A3D8 sending paircmd 2 sendSB sent FE2524020FFFFFFFFFFFFFFFFF14FFFF140F0102000F1100704000161600A3D810FFFF80971B01A3D833 analyzing paircmd readZB FE0164020067FE25448100000F010000141401240014B600000011704000161600A3D810FFFF80971B01A3D87C4B0D6F rc=48 decoding : FE0164020067FE25448100000F010000141401240014B600000011704000161600A3D810FFFF80971B01A3D87C4B0D6F found invID in MessageToDecode found invID = A3D8 pair command = 24020FFFFFFFFFFFFFFFFF14FFFF14010103000F060080971B01A3D8 sending paircmd 3 sendSB sent FE1A24020FFFFFFFFFFFFFFFFF14FFFF14010103000F060080971B01A3D84F success, inverter got id A3D8 sending N.O. cmd sendSB sent FE162401FFFF1414060001000F1E80971B01A3D8FBFB1100000D4F readZB FE0164010064FE034480001401D2 rc=14 sendSB sent FE00270027 readZB FE0E670000FFFF80971B01A3D8000007090011 rc=19 found 0709, running oke

`

the found inverter ID differs from time to time and as you can see, the written ID is the same as last 4 digitsof the ECU ID.

patience4711 commented 1 year ago

in the decoded string, the quoted part is your serial nr?
FE0164020067FE25448100000 F010000141401 240014B600000011 "70 40 00 16 16 00" A3D810FFFF80971B01A3D87C4B0D6F

when i pair a fake inverter with serial 111111111111 i get this: FE0164020067FE25448100000F010000141401 780072A900000011 11 11 11 11 11 11 B3D610FFFF80971B01B3D63A100D67 which evalutes to id B3D6 which is the last 4 digits of my ecu id reversed.

So when i pair a non existant inverter i get the same as you get, So my first question is; could it be that you are also pairing a non existant inverter? Could it be that the serial was corrupted like happened to the inverter id ??

When i pair my real inverter I get a much longer answer that contains the serialnr twice, behind the first occurence the B3D6 belongs to my ecu, behind the second one the sought id. FE0164020067FE25448100000F0100001414017B00F7A900000011 "408000158215" B3D610FFFF80971B01B3D63A100DA3FE1C4481000001013A101414007B0054AD00000008FFFF "408000158215" 3A100E1F rc=81

So the answer isn't correct (or the serial). The strange thing is that it works this way as long as the project exists. I can build an evaluation that rejects the message when too short or when the id equals the ECU short id, but i must study that first.

swbouman commented 1 year ago

My Ecu Id is D8A3011B9780 and the serial of the paired inverter is 704000158937 the working id is A1BA but was 9BDE earlier with the same Ecu id.

patience4711 commented 1 year ago

The ECU-Short id = A3D8, paircommand 1 and 2 should contain that serial but it doesn't.

snprintf(pairCmd[2], sizeof(pairCmd[2]), "24020FFFFFFFFFFFFFFFFF14FFFF140F0102000F1100%s%s10FFFF%s", Inv_Prop[which].invSerial, ecu_short, ecu_id_reverse);

sendSB sent FE2524020FFFFFFFFFFFFFFFFF14FFFF140F0102000F1100 704000161600 A3D8 10FFFF 80971B01A3D8 33

So it looks as if Inv_Prop[which].invSerial has the value 704000161600

swbouman commented 1 year ago

Well, I do have that serial in the inverters. But it was not the inverter that was pairing. For the found serial I have inv Id 9BDE (704000161600) is everything messed up in my config?

patience4711 commented 1 year ago

I don't know what happened but maybe your filesystem is corrupted. Or you where pairing another inverter than you thought.

Meanwhile i implemented something to prevent that an inverter gets an id equal to the ECU short id. Tomorow I'l test it.

patience4711 commented 1 year ago

@swbouman I changed it in such way that pairing answers that are not long enough are not analyzed. I can't pair any non-existant inverters anymore. My own inverter paired without problems in one shot.

v9_13d

swbouman commented 1 year ago

I give it a shot! Thanks for the effort!

swbouman commented 1 year ago

this is my output of polling inverter with id: 704000159303

readZB FE0164020067FE1A448100000C0200001414015000F3A1000000067040001593039FDD0D2E rc=37

decoding : FE0164020067FE1A448100000C0200001414015000F3A1000000067040001593039FDD0D2E no valid pairing code, returning... pair command = 24020FFFFFFFFFFFFFFFFF14FFFF140F0102000F1100704000159303A3D810FFFF80971B01A3D8 sending paircmd 2 sendSB sent FE2524020FFFFFFFFFFFFFFFFF14FFFF140F0102000F1100704000159303A3D810FFFF80971B01A3D8B6 analyzing paircmd readZB FE0164020067FE25448100000F0100001414014E0051BC00000011704000159303A3D810FFFF80971B01A3D89FDD0DBAFE1C448100000101534C141400120074C300000008FFFF704000159303534C0ECF rc=81 decoding : FE0164020067FE25448100000F0100001414014E0051BC00000011704000159303A3D810FFFF80971B01A3D89FDD0DBAFE1C448100000101534C141400120074C300000008FFFF704000159303534C0ECF found invID = 534C pair command = 24020FFFFFFFFFFFFFFFFF14FFFF14010103000F060080971B01A3D8 sending paircmd 3 sendSB sent FE1A24020FFFFFFFFFFFFFFFFF14FFFF14010103000F060080971B01A3D84F success, inverter got id 534C sending N.O. cmd sendSB sent FE162401FFFF1414060001000F1E80971B01A3D8FBFB1100000D4F readZB FE0164010064FE034480001401D2 rc=14 sendSB sent FE00270027 readZB FE0E670000FFFF80971B01A3D8000007090011 rc=19 found 0709, running oke

swbouman commented 1 year ago

and this for the inverter with serial: 704000159286

readZB FE0164020067FE1A448100000C0200001414018F00738F00000006704000159286A1BA0D82 rc=37

decoding : FE0164020067FE1A448100000C0200001414018F00738F00000006704000159286A1BA0D82 no valid pairing code, returning... pair command = 24020FFFFFFFFFFFFFFFFF14FFFF140F0102000F1100704000159286A3D810FFFF80971B01A3D8 sending paircmd 2 sendSB sent FE2524020FFFFFFFFFFFFFFFFF14FFFF140F0102000F1100704000159286A3D810FFFF80971B01A3D832 analyzing paircmd readZB FE0164020067FE25448100000F0100001414015000D4A900000011704000159286A3D810FFFF80971B01A3D82ED00D0CFE1C4481000001019FDD1414005000B6B200000008FFFF7040001592869FDD0EBA rc=81 decoding : FE0164020067FE25448100000F0100001414015000D4A900000011704000159286A3D810FFFF80971B01A3D82ED00D0CFE1C4481000001019FDD1414005000B6B200000008FFFF7040001592869FDD0EBA found invID = 9FDD pair command = 24020FFFFFFFFFFFFFFFFF14FFFF14010103000F060080971B01A3D8 sending paircmd 3 sendSB sent FE1A24020FFFFFFFFFFFFFFFFF14FFFF14010103000F060080971B01A3D84F success, inverter got id 9FDD

patience4711 commented 1 year ago

@swbouman blow the horn!

You can see what's happening here. In the past the answer of the command 1 would evaluate to an id of A1BA (incorrect) but this was overwritten by the answer to command 2, which evaluates to 9FDD. So this worked oke. But if the second answer isn't correct for some reasons, the first (or the incorrect second) value was saved. That can't happen anymore now.

swbouman commented 1 year ago

🥳

patience4711 commented 1 year ago

No news is good news so they say....