|Re: 15C LE Synthetics|
Message #3 Posted by Eric Smith on 21 Sept 2011, 2:51 p.m.,
in response to message #2 by Gerson W. Barbosa
It does NOT work on the 15C LE. I had strong reason to suspect that it wouldn't, and just verified it on an actual calculator.
The way synthetics on the 15C worked depended on an obscure hardware feature of the chip that is totally unrelated to the Nut instruction set, so most microcode-level simulators will NOT support it. If it works on Michael's DM-15C, it is because he went out of his way to support it.
The ON-y^x key combination of the original 15C (and the same combination of physical keys on the other original Voyagers) forces a hardware reset. The firmware has no control over this. It is mentioned in the original manual as something to try if the calculator is unresponsive, with the note that the X register will be altered and should be cleared afterward. The specific change to the X register is that it is rotated 22 bit positions. This can result in an NNN ("non-normalized number"), which on the 15C can be an invalid matrix descriptor.
I have reason to believe that other portions of the calculator internal state are also modified, but have not verified this. Probably the A,B,C,M, and N registers are all rotated by 22 bits, and the F, G, and S registers are probably also rotated by some amount. The P register might or might not be changed.
For example, press the key sequence 1 0 1 7, then the ON-y^x combo. On an original 15C this gives an invalid matrix descriptor, while on the 15C LE it just turns the calculator off, and turning it on again shows an unmodified value 1017 in the X register.