Dump the ROM of fx-82ES PLUS

If that is true, does the method still work if the last (7×'9D') group is replaced by '1010...'?

About 82ES+A:
The method http://tieba.baidu.com/p/3022526172 can be used to read a 84-byte block of the ROM in the range 0x0000 to 0x8000.
Just replace the two ". Ran#" (0x8C2E) by an even number in 0x0000 - 0x8000. The 84 outputted bytes are part of the ROM starting from that address.
Assume the ROM of 82ES+A is similar to 82ES+ emulator, it is possible to find necessary functions/gadgets for dumping the whole ROM.
For example: If we know that the block starts at 0x1234 on the real calculator is similar to the block starts at 0x1534 on the emulator, and the block starts at 0x2534 on the real calculator is similar to the block starts at 0x2834 on the emulator, it is likely that the inbetween 0x1600-byte block is also similar.
Because convert the screen pixels to bytes manually is very time-consuming, it would be helpful if somebody can make a program that can automatically do that from an image.

can you told me where are you come from

@82espa💧 If all characters can be typed without the unstable character, would it be easier to use math overflow mode instead of linear overflow mode?

@185264646
It is almost sufficient to write a program to get the first half of the segment 0.
Can you try this: (enter as 100 effective characters)
Enter 62 characters
³ : 0 0 (0x763a3030)
0 0 (0x3030)
6 < 0 0 (0x363c3030)
565636 sin 1 sin 1 (0xA031A031) 0 tan 0 tan RanInt# RanInt#
Enter 22 characters
The part "5656 ... RanInt#" is identical to the 16 characters in the method to get a memory block.
If I guessed correctly, the calculator should delay for about a second before printing the memory block start at 0x31A0.
If that is not true, please get the memory block at 0x3a68 ("log :").

@185264646
Expected result: the calculator show some data on the screen, and it changes every ~1 second.
If that is not the case, please get the memory blocks at 0x3A5C ("▫ :") and 0x31A0 ("sin 1") (just to be sure, because in the image above, the number of columns is unknown, and only the 12 bytes in the first line matches).
"6 5"
"( 0 0 0" (0x28303030)
56 characters
"³ : 0 0" (0x763A3030)
"0 0" (0x3030)
"6 < 0 0" (0x363C3030)
"³ : 0 0"
"μp 0" (0x1F30)
"tan : 0 0" (0xA23A3030)
"Σx² J▶kgf·m" (0x10F8)
"A 1 0 0" (0xB8313030 - note A is 0xB8 not 0x41)
"2 2"
"0 1 2 3 4 5"
"⌟ Ans" (0xAE8B)
"8 7"

@82espa💧

@82espa💧

I fixed the errors manually. At least the checksum matches.
However, the methods using r I've tried executes some unrecognized instruction, and the emulator cannot execute them correctly. It would be hard to explain those methods (such as 119)