"incorrect start/stop bytes" with Voltacon Conversol max 8kW off-grid inverter

Question

"incorrect start/stop bytes" with Voltacon Conversol max 8kW off-grid inverter

Closed this issue 2 months ago · 14 comments

Inverter: https://voltaconsolar.com/off-grid/conversol-max-8kw-off-grid.html
Finally I've got this working and this id output with debug option:

bananapi:skymax-demo:# ./inverter_poller -d
Sun Oct 29 14:33:23 2023 INVERTER: Debug set
Sun Oct 29 14:33:23 2023 INVERTER: Current CRC: 49 C1
Sun Oct 29 14:33:23 2023 INVERTER: QMOD reply size (5 bytes)
Sun Oct 29 14:33:23 2023 INVERTER: QMOD: 5 bytes read: (L
Sun Oct 29 14:33:23 2023 INVERTER: QMOD query finished
Sun Oct 29 14:33:23 2023 INVERTER: Current CRC: B7 A9
Sun Oct 29 14:33:24 2023 INVERTER: QPIGS reply size (110 bytes)
Sun Oct 29 14:33:24 2023 INVERTER: QPIGS: 110 bytes read: (239.5 49.9 239.5 49.9 0742 0631 009 375 38.60 000 000 0040 00.9 101.5 00.00 00000 00010110 00 00 00099 010
Sun Oct 29 14:33:24 2023 INVERTER: QPIGS query finished
Sun Oct 29 14:33:24 2023 INVERTER: Current CRC: F8 54
Sun Oct 29 14:33:24 2023 INVERTER: QPIRI reply size (98 bytes)
Sun Oct 29 14:33:24 2023 INVERTER: QPIRI: incorrect start/stop bytes.  Buffer: (230.0 34.7 230.0 50.0 34.7 8000 8000 48.0 44.0 44.0 56.4 54.0 0 002 020 0 1 3 9 01 0 0 54.0 0 1 400099 010
Sun Oct 29 14:33:24 2023 INVERTER: Current CRC: B4 DA
Sun Oct 29 14:33:24 2023 INVERTER: QPIWS reply size (36 bytes)
Sun Oct 29 14:33:24 2023 INVERTER: QPIWS: incorrect start/stop bytes.  Buffer: (000000000000000001000000000000010000 48.0 44.0 44.0 56.4 54.0 0 002 020 0 1 3 9 01 0 0 54.0 0 1 400099 010

Without parameters nothing is thrown to the console. Different comm protocol?
I've modified inverter.cpp changing

        if (buf[0]!='(' || buf[replysize-1]!=0x0d) {
            lprintf("INVERTER: %s: incorrect start/stop bytes.  Buffer: %s", cmd, buf);
            return false;
        }

to

        if (buf[0]!='(') {
            lprintf("INVERTER: %s: incorrect start byte.  Buffer: %s", cmd, buf);
            return false;
        }
        if  (buf[replysize-1]!=0x0d) {
            lprintf("INVERTER: %s: incorrect stop byte.  Buffer: %s", cmd, buf);
            return false;
        }

with result:

bananapi:skymax-demo:# ./inverter_poller -d
Sun Oct 29 21:31:25 2023 INVERTER: Debug set
Sun Oct 29 21:31:25 2023 INVERTER: Current CRC: 49 C1
Sun Oct 29 21:31:25 2023 INVERTER: QMOD reply size (5 bytes)
Sun Oct 29 21:31:25 2023 INVERTER: QMOD: 5 bytes read: (L
Sun Oct 29 21:31:25 2023 INVERTER: QMOD query finished
Sun Oct 29 21:31:25 2023 INVERTER: Current CRC: B7 A9
Sun Oct 29 21:31:26 2023 INVERTER: QPIGS reply size (110 bytes)
Sun Oct 29 21:31:26 2023 INVERTER: QPIGS: 110 bytes read: (241.9 49.8 241.9 49.8 0676 0558 008 378 44.30 000 000 0039 00.0 000.0 00.00 00000 00010000 00 00 00000 010
Sun Oct 29 21:31:26 2023 INVERTER: QPIGS query finished
Sun Oct 29 21:31:26 2023 INVERTER: Current CRC: F8 54
Sun Oct 29 21:31:26 2023 INVERTER: QPIRI reply size (98 bytes)
Sun Oct 29 21:31:26 2023 INVERTER: QPIRI: incorrect stop byte.  Buffer: (230.0 34.7 230.0 50.0 34.7 8000 8000 48.0 44.0 44.0 56.4 54.0 0 002 020 0 1 3 9 01 0 0 54.0 0 1 400000 010
Sun Oct 29 21:31:26 2023 INVERTER: Current CRC: B4 DA
Sun Oct 29 21:31:27 2023 INVERTER: QPIWS reply size (36 bytes)
Sun Oct 29 21:31:27 2023 INVERTER: QPIWS: incorrect stop byte.  Buffer: (100000000000000001000000000000010000 48.0 44.0 44.0 56.4 54.0 0 002 020 0 1 3 9 01 0 0 54.0 0 1 400000 010

so apparently there is missing CR in QPIRI and QPIWS response or response is longer. So changed main.cpp

int qpiri = 140; //98;
int qpiws = 60; //36;

and found they are 109 and 40 respectively, with new values JSON response is printed to console but second MPPT input data is missing.

Answer 1 · 2023-10-30T05:11:14.000Z

Hi,
It is normal that the protocol differs slightly.
Thanks to @chadek we now have the parameters now merged (#22).
Isn't it sufficient to set it in the ini ?

Answer 2 · 2023-10-30T08:29:33.000Z

Hi @manio, yes, now buffersize parameters is supported based on what was done on this repo. You have no need to change the init value, they are are just to initialize variables and are currently work with the Axpert VM3. You can define these values into the config file. I'm seeing now that there is no example or information anyway about this. Maybe we could add some comments into the example config file (and why not into the read me)?

# This allows you to modify the buffersize for the qpiri command
qpiri=98

# This allows you to modify the buffersize for the qpiws command
qpiws=36

# This allows you to modify the buffersize for the qmod command
qmod=5

# This allows you to modify the buffersize for the qpigs command
qpigs=110

About your "second mppt data" I think this is proper to your model, the Axpert VM3 inverter has only one. Here there is a manual describing the protocol (at least one old version). As the reply size of your qpiri and qpiws is quite bigger, I guess that the data of your second mppt is stored inside the response you get but not entirely parsed by the main function that only provide a certain quantity of variables. You might want to use the -r option to send a raw qpiri/qpiws command and check out the result to have an idea on the output.

Answer 3 · 2023-10-30T08:42:02.000Z

Here is a slightly more recent version of the protocol and there one with even more information. If compare with the 2014 version, it seams that they did not change the wall thing but only some more information to the protocol.

Answer 4 · 2023-10-30T09:17:46.000Z

Maybe we could add some comments into the example config file (and why not into the read me)?

I am open to PR :)

Answer 5 · 2023-10-30T09:46:01.000Z

@A13xund3r I might have found a doc that could match your model protocol version : this one seams to have a QPIGS2 query that supposedly would output your MPPT 2 value.

Answer 6 · 2023-10-30T11:49:39.000Z

@A13xund3r I might have found a doc that could match your model protocol version : this one seams to have a QPIGS2 query that supposedly would output your MPPT 2 value.

Something is wrong with raw command receiver, it (almost) always read 7 bytes but apparently you are correct, I need QPIGS2 command

bananapi:skymax-demo:# ./inverter_poller -d -r QPIGS
Mon Oct 30 11:40:37 2023 INVERTER: Debug set
Mon Oct 30 11:40:37 2023 INVERTER: Current CRC: B7 A9
Mon Oct 30 11:40:37 2023 INVERTER: QPIGS reply size (7 bytes)
configuration options for the actual inverter polling process...Buffer: (240.7 

bananapi:skymax-demo:# ./inverter_poller -d -r QPIGS2
Mon Oct 30 11:43:34 2023 INVERTER: Debug set
Mon Oct 30 11:43:34 2023 INVERTER: Current CRC: 68 2D
Mon Oct 30 11:43:34 2023 INVERTER: QPIGS2 reply size (7 bytes)
onfiguration options for the actual inverter polling process...  Buffer: (00.0 0-

Currently second PV is not connected so 0V is expected.

bananapi:skymax-demo:# ./inverter_poller -d -r QMOD
Mon Oct 30 11:44:27 2023 INVERTER: Debug set
Mon Oct 30 11:44:27 2023 INVERTER: Current CRC: 49 C1
Mon Oct 30 11:44:30 2023 INVERTER: QMOD read timeout
Mon Oct 30 11:44:30 2023 INVERTER: QMOD reply too short (5 bytes)
Reply:

reply line is (almost) always empty.
I'm happy to test but I'm assembler/basic hobbyist programmer so fiddling with C++ is extremely difficult for me.

bananapi:skymax-demo:# ./inverter_poller -d -r QPI
Mon Oct 30 13:39:19 2023 INVERTER: Debug set
Mon Oct 30 13:39:19 2023 INVERTER: Current CRC: BE AC
Mon Oct 30 13:39:19 2023 INVERTER: QPI reply size (8 bytes)
Mon Oct 30 13:39:19 2023 INVERTER: QPI: 8 bytes read: (PI30
Mon Oct 30 13:39:19 2023 INVERTER: QPI query finished
Reply:  PI30

Answer 7 · 2023-10-30T19:45:10.000Z

@A13xund3r you should be able to query your inverter with the code on this PR, using the new -l option to set the length of your buffer. Good luck with the length guessing!

Answer 8 · 2023-10-30T19:48:04.000Z

Maybe we could add some comments into the example config file (and why not into the read me)?

I am open to PR :)

I'll open one to update config and the readme, including the buffer length values I just removed on PR

Answer 9 · 2023-10-30T23:20:30.000Z

@A13xund3r you should be able to query your inverter with the code on this PR, using the new -l option to set the length of your buffer. Good luck with the length guessing!

It works, no need to guess, I can count bytes from PDF you found as my inverter is using this protocol or set something higher and let program tell me message length:

bananapi:skymax-demo:# ./inverter_poller -l 160 -d -r QPIGS2
Mon Oct 30 22:02:39 2023 INVERTER: Debug set
Mon Oct 30 22:02:39 2023 INVERTER: Current CRC: 68 2D
Mon Oct 30 22:02:42 2023 INVERTER: QPIGS2 read timeout
Mon Oct 30 22:02:42 2023 INVERTER: QPIGS2 reply too short (21 bytes)
Reply:
bananapi:skymax-demo:# ./inverter_poller -l 21 -d -r QPIGS2
Mon Oct 30 23:13:19 2023 INVERTER: Debug set
Mon Oct 30 23:13:19 2023 INVERTER: Current CRC: 68 2D
Mon Oct 30 23:13:19 2023 INVERTER: QPIGS2 reply size (21 bytes)
Mon Oct 30 23:13:19 2023 INVERTER: QPIGS2: 21 bytes read: (00.0 000.0 00000
Mon Oct 30 23:13:19 2023 INVERTER: QPIGS2 query finished
Reply:  00.0 000.0 00000

After -l 160 inverter was replying 7 bytes (NAK to every command until i started slsnif sniffer to see what is going on

Answer 10 · 2023-11-01T17:39:37.000Z

After hours of googling it seems my inverter use PI30MAX protocol. So far everything does match with my inverter.

Answer 11 · 2023-11-01T20:58:12.000Z

I don't have access to my hardware to test this out, but it should solve this issue when it is done : #26

Answer 12 · 2023-11-17T19:30:21.000Z

@A13xund3r I just found out that the document you point out describing the PI30MAX protocol, is part of a wonderful repo supporting a lot of inverter protocol, including yours. You should try it, it seams to work quite well :)

Answer 13 · 2023-11-17T21:22:08.000Z

It is in python, not my preferred language to run on SBC https://www.freecodecamp.org/news/python-vs-c-plus-plus-time-complexity-analysis/ but i have it on my list to check. Just day is too short...

Answer 14 · 2024-04-24T18:09:26.000Z

Closing as it is fixed.