Open randy- opened 3 years ago
splatch
commented
2 years ago Modbus spec in early days was not clear on that hence -1 and +1 errors are happening. Its all about how end device index registers - if it does it from 0 or from 1, then its matter of client if he assumes any offsets. Its a feature of a protocol.
Switching it might help you with that device but break dozen of others.
syadnom
commented
1 year ago splatch, thank you sir. I didn't come here for that specific question but I couldn't get any data out of a mikrotik knot modbus tcp bridge to a mean well DRS-480-24 unit. I didn't know the +/- 1 issue existed.
For anyone else that ends up searching and finding this, do hex to decimal and add 1 for the register address.
(modbus mikrotik knot meanwell mean well DRS)
the78mole
commented
1 year ago I also see the -1 offset. If you enable debugging output (option -v), you can see, that it requests a different register address than the user requests on the command line. I think that is more confusing than it helps.
git-developer
commented
1 year ago I believe this behavior is defined by the Modbus specification (didn't look it up, though). mbpoll has command line argument -0 to shift the reference adresses from 1-based to 0-based.
cowo78
commented
2 weeks ago IMHO it's worth making this situation clearer in the help message: -r is the OFFSET from register 1 (or register 0 if -0 is specified)
Hello,
Testing mbpoll with Victron Energry Color GX. The Victron Energry Color GX error logs shows mbpoll 1.0-0 (Debian 1.4.11+dfsg-2~bpo10+1) starting reference (-r) is off by -1. Example: ./mbpoll -1 -a 100 -t 3 -r 840 -c 1 192.168.10.180 mbpoll will starts at 839 not 840.
Fix source file mbpoll.c change line 952~ iStartReg = ctx.piStartRef[j] - ctx.iPduOffset; to iStartReg = ctx.piStartRef[j];