I think the trick will be finding the right way to write data to registers 62-64, you'll probably be able to read them ok.
Indeed by reading them out first (and that is another can of worms - which of the two address spaces, function code 3 or 4 - try which one works), you can easily confirm if they are the current date/time or something else, and if they are, you can verify that what you are going to write to them has the same byte order. Otherwise, it's easy - just write them with function code 0x10 in one go.
Ask Deye for the Modbus register listing. Is this secret? Why?
What I posted is an excerpt of machine translated Chinese document. It isn't a secret per se and should be easy to find with Google, and I see no reason why Deye would not give it upon request given that you find the right person. Which isn't always easy.
One surprisingly big problem with Chinese inverters is that the modbus documents tend to be prefaced with something like "communication protocol for grid-tie inverter" or similar - that's it, sometimes even lacking manufacturer name, and almost never a
list of model numbers that the document applies to. Therefore, for manufacturer X, there tends to be documents A, B and C which are all different and apply to different models, but nothing about to which it applies to, and then for each, older, newer, differently translated versions out of which you can only find some. It's some detective work. And I still prefer that over a committee disaster like Sunspec.
Therefore, your advice is good - be careful before doing random writes. It is
unlikely you accidentally hit some register that causes a disaster like brickage of further write attempts - never happened to me reverse-engineering and poking shit out of dozen of inverters -, but it's
possible.
Home
Help
Search
Login
Register
