I got the AY-AT version of the tester. It came with the 1.12K firmware version.
I was able to program the 1.13K version using the HEX and EEP files downloaded from the software repository.
Next I wanted to try the 1.27m version but I couldn't compile anything on my windows 10 computer using WinAvr. I tried FlyWheelz instructions but they didn't work. I gave up and used a Windows 7 32 bit computer where I had all the software but hadn't used for a long time. I was able to modify the K code and compile it successfully but it was creating very big code that wouldn't fit on the Atmega 328P.
After more reading and searches on the internet I downloaded Cygwin (64 bit in my case), installed on the windows 10 machine and modified the computer path to include the cygwin64\bin folder. I was able to compile without any problems
and most of the code generated fits on the microcontroller. Finally I was able to compile the 1.27m firmware and today the 1.28m for the AY-AT board.
I found some discrepancies with the different firmware versions:
I like how the K versions show more information when testing components such as transistors. Is it possible to modify the m version to include that information?
The m version code is more organized and easier to follow.
When testing a PN4117A the only firmware that shows correct information is on the m versions. At least from version 1.25m and up ( I haven't tested older versions)
The 1.12K version thinks that is a double diode. The 1.13K version thinks that is a transistor. The 1.28m version thinks that is an N-channel JFET with Idss=75ua. I know that Marcus code can detect JFETs with very low Idss.
I also have a Peak DCA75 and it detect the part as a double diode.
I haven't contacted them to see if they can modify the code to recognize the JFETs with low Idss.
Thank you to Markus and Karl for continuing the work of Markus Frejek and for being so available to help the comunity.