Thanks

I just decompiled the important part:
http://nopaste.info/18a647495d.html
http://nopaste.info/ea058d8046.html
http://nopaste.info/145e617042.html

hf playing

the second attempt was to seen why this software behaves so strange.
it needs at least a pentIII with 800MHz to run smoothly. for a task that a 8bit embedded processor is needed.
they use a DCOM interface to communicate via windows to the rs232.
maybe there lies the problem. and the timeing gets too slow.
but you have to see the code for yourself.

at least the inverters and their software work without problems.

even without decompliling you can get a lot of information by reflection. you can debug the software.
there are some configuration dialogs. a loop to get the values from the inverters and to write them to a file. also a dialog to display the current values.
and you dont realy have to decompile the files just look at them with a hex editor.
i belive that you will find all the strings you need to complete you communication.

i also like challanges. but their software(hopefully not the one in the inverters) just does not deserve to be treated by respect.
the secrect password for changing the inverter settings was given to us as a PDF document. but over the password was a black box. so nobody could see the password.
but viewed with a slow computer the adobe reader would print the password and put the black box over it. so you had time enough to see ist. and printed with an old laserprinter the box was not printed black at all and you could see the password under the box.

after that i dont see a need to be challanged. if you use java or some .net language and you dont want your code to be reverse engineered then us an obfuscator or write your software in C/C++. then it would run on a 386 processor.
and why use java in the first place. when you use DCOM runs just on windows machines.

the solarmax people may have good intentions. but somehow they failed to impress me.
the expection are there inverters. they work without problems. (only one was replaced and this one only lost its time an configuration settings. we would not have noticed it without the monitoring.)

für RS232:
RJ45 Pin 4 —– GND —– subd9 Pin 5
RJ45 Pin 5 —– TXD —– subd9 Pin 3
RJ45 Pin 6 —– RXD —– subd9 Pin 2

für RS485
RJ45 Pin 1 —– V+ —–
RJ45 Pin 2 —– V+ —–
RJ45 Pin 3 —– GND —–
RJ45 Pin 4 —– GND —–
RJ45 Pin 5 —– NC —–
RJ45 Pin 6 —– NC —–
RJ45 Pin 7 —– A —–
RJ45 Pin 8 —– B —–

i think the V+ is used to power the tranceiver chips.
the internal inverter and the external interface part are opto isolated.
that’s why the power has to be provided externaly.
but i’m not copletely sure.
i bought a cheap RS485 interface and installed it long ago.
should have taken more notes then.

You don’t have to reverse engineer the protokoll. you only have to download the software from solarmax. locate the jar files and decompile the java code. in the java there is all the information you need to see how everything works.

There is no CRC involved, it’s a simple checksum: just sum up all the bytes and keep the lowest 16 bits of the sum.

J.