PROGRAM: ; This program *doesn't work* on the emulator. The emulator takes shortcuts when synchronizing ; WAIT and KEY, and the bad timing that this program uses does not trigger. ; ; If you are interested in details: ; ; The program calls WAIT D12 (opcode 0x0AC0) instead of the canonical WAIT D1 before SET IDLE. ; This shifts the IDLE loop 11 digit positions out of phase with the display strobe cycle. ; ; On real hardware the display is multiplexed in real time: at each digit time Dn, the SCOM chip ; asserts the LED strobe for digit n while the CPU must simultaneously drive the segment data for ; that position. When SET IDLE is phase-shifted, CPU data and strobe get misaligned — the B ; register's DPT nibble (which the 7-segment decoder maps to a hex glyph 0x0–0xF) ends up driving ; an LED position it normally never would. That exposes all 16 hex segment combinations, which is ; exactly the test's goal. ; The emulator's display on the other hand is a phase-independent snapshot taken at digit == 0 ; This copies the whole A/B register arrays unconditionally at digit 0. The phase shift from WAIT D12 ; has zero influence on which index lands in which slot. The display always shows expected content; ; the DPT-as-digit trick is invisible. ; ; The same is true for keyboard scans. Phase shifted, keys are detected wrongly, or do not even exist. ; Since the ROM often dispatches key presses via the scan code using the PREG mechanism, the handler for ; inexistent keys is just some arbitrary ROM location, and can have any effect. ; ; The emulator again does not phase shift, and reads the keys correctly. ; ; Without some significant rework, the emulator cannot handle this particular ROM program. So far ; it's unknown if other tricks discovered in the 80ies do not work due to the shortcut taken by ; the emulator. It's possible that this needs to be changed in the future. ; Original comment follows: ; ; 7-segment decoder test ; This test is also little bit tricky. It uses "bad" digit synchronization to display DPT on LED. As DPT is ; hexadecimal, this allows to display all combinations available in 7-segment decoder. Value displayed is ; BB.AAX, where BB is value in B.DPT, AA is value in A.DPT and X is resulting 7-segment digit. ; Because of illegal synchronization, some keys behave strange! (RCL makes CPU reset, LRN row ; doesn't work, BST row doesn't detect key press but increment is done until keys are held – this can be ; useful to test higher values...) ; Note: B.DPT can't be used to increment using mask value because B is on the same ALU input as ; mask constant, so B (or D) is or-ed with this constant instead of adding it. 1800: 0A01 CLR IDL 01D0 MOV A.ALL,#0 ;prepare digit mask to B register: 99990 01D3 MOV B.ALL,#0 07DB MOV B.EXP,#-1 0153 SHL B.ALL,B 0153 SHL B.ALL,B 07DB MOV B.EXP,#-1 01D4 MOV C.ALL,#0 ;set IDLE but shifted so DPT is visible 0AC0 WAIT D12 0A09 SET IDL ;main loop ; wait for key press 180A: 0AE0 WAIT D14 0820 KEY 20 180C BRA1 +6 ;1812 0A45 TST KR[4] 0A55 TST KR[5] 0A65 TST KR[6] 0A75 TST KR[7] 100F BRA0 -7 ;180A ;check key press = debounce key 1812: 0AE0 WAIT D14 0820 KEY 20 1805 BRA1 -2 ;1812 ;increment A.DPT 0300 ADD A.DPT,A,#1 1806 BRA1 +3 ;1819 ;increment C.DPT if carry 0324 ADD C.DPT,C,#1 ;and copy C.DPT to B.DPT 0223 ADD B.DPT,C,#0 ;copy A.DPT to display the number 1819: 01D6 MOV D.ALL,#0 0206 ADD D.DPT,A,#0 0176 SHL D.ALL,D 0630 ADD A.EXP,#0,D ;copy C.DPT to display the number 01D6 MOV D.ALL,#0 0226 ADD D.DPT,C,#0 0176 SHL D.ALL,D 0176 SHL D.ALL,D 0176 SHL D.ALL,D 0930 ADD A.MANT,#0,D ;clear COND 1002 BRA0 +1 ;1824 ;set DPT position 1824: 0A37 MOV R5,#3 1837 BRA1 -27 ;180A HEX: 0A01 01D0 01D3 07DB 0153 0153 07DB 01D4 0AC0 0A09 0AE0 0820 180C 0A45 0A55 0A65 0A75 100F 0AE0 0820 1805 0300 1806 0324 0223 01D6 0206 0176 0630 01D6 0226 0176 0176 0176 0930 1002 0A37 1837