Thanks for the input, ~ Yup that is what did it for interfacing with the LT500. I brought in each change from Dec 1st till I found it
Doesn’t make a lot of sense to me why that impacts the protocol, but in dealing with a bug, the first stage is Detection, 2nd stage is Isolation, 3rd stage is Fixing it.
The Isolation in this case took a couple of days.
The software process is modify and test, and then integration test. In this case what is lacking (for me) is having a defined reference for the SDI-12 sensor. It also seems that working with the LT500, there are retry’s that finally work. So its a general problem with this kind of dispersed protocol testing. USB protocol testing is a lot worse, and they solve it by having plug-fests. But then there is a lot of money in selling USB devices.
I have been keeping my eyes open for a low cost reference sensor, ideally that would be open source as well, but I haven’t found one yet.
I’m thinking that for me, since SDI-12 is so fragile, and its not clear how to verify changes except by running them against reference sensors I should manage any changes happening more closely – fork the libs concerned. Just thinking out a loud.