> configuring the MODBUS plug-in is a bit tricky. The reason: there are some devices out there that work with wrong

> byte order (according to MODBUS specification) and some that use wrong address offsets (and some that do both

> wrong). So when you configure a plug-in for your device you in most cases will not have exact specification of this

> devices behaviour, so you only can try all variants: invert/non-invert address/data byteorder, start with an address

> counting from 0/1.

I’ll go along with that but it appears as if un-checking the box gives small-endian byte order, and checking it gives big-endien. The box is checked by default so I suppose it makes a sort of sense.

In version 3.3.2 the return value appeared to be in the wrong order as well though, I had to divide by 256 to correct it.

Version 3.6 doesn’t appear to receive at all, So far I’ve tried effectively the same device program but on two different microcontrollers, and a example device program that outputs a continuous count value on register 999.

A raw com-port dumper would be really useful at this point.

Incidentally the “DLL hell” issue is that if you uninstall OpenAPC 3.6 and reinstall 3.3.2 then you get an odd error and it won’t run. Windows XP and 7. It means I can’t readily switch between them unless I do something messy like putting OpenAPC in a virtual machine. I’ve got the error message saved somewhere, can’t find it now. At work I managed to roll the system back enough to clear it but this home machine’s had 3.6 on too long, system restore would muck up too many other things.