Sprinter/mame
2
0
Fork 0
mirror of https://github.com/holub/mame synced 2025-07-04 17:38:08 +03:00
Code Issues Actions 36 Packages Projects Releases Wiki Activity

excal: update notes

Browse Source
This commit is contained in:
hap 2025-01-21 21:05:26 +01:00
parent 6e6a7e1ec5
commit 1aad5b1349
1 changed files with 13 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
src/mame/hegenerglaser/excalibur.cpp
View File

@ -12,7 +12,7 @@ Hardware notes:
Motherboard: Motherboard:
- PCB label: FORCE COMPUTERS - PCB label: FORCE COMPUTERS
- HD68000-8, 16MHz XTAL - HD68000-8 @ 8 MHz (16MHz XTAL)
- 128KB DRAM (16*MB8264-20), half unused - 128KB DRAM (16*MB8264-20), half unused
- 4 EPROM sockets are unpopulated - 4 EPROM sockets are unpopulated
- HD6840 PTM @ 800kHz, HD6821P PIA - HD6840 PTM @ 800kHz, HD6821P PIA
@ -34,13 +34,19 @@ except the interface is more similar to Mephisto Exclusive.
The motherboard is a 1983 Force Computers 68000 kit with lots of patches and wire The motherboard is a 1983 Force Computers 68000 kit with lots of patches and wire
mods. It's an older version of the VME in src/devices/bus/vme/sys68k_cpu1.cpp. mods. It's an older version of the VME in src/devices/bus/vme/sys68k_cpu1.cpp.
The CPU appears to run at 5MHz? This matches both piezo frequency and position The 68000 is clocked at 8MHz, but due to wait states and bus requests, it runs
calculation speed. How it derives this from a 16MHz XTAL is unknown. Or maybe much slower, closer to around 5MHz.
it uses wait states? If so, how and why? The program copies itself to 280ns DRAM,
which should be fast enough to run at 8MHz (500ns per memory cycle). There's a 555 timer (R1=11K, R2=22K, C=1nf: ~26kHz) that clocks 68000 BR. Tests
on real hardware showed that the chess computer is about 30% faster if the timer
interval is reduced by a factor 100. It still worked fine, so it's probably a
leftover for VME comms, not the DRAM refresh circuit.
Other possible sources of wait states are periodic DRAM refresh and user/system
area access time.
See glasgow.cpp on how to verify CPU speed. On the real thing after 6 minutes, See glasgow.cpp on how to verify CPU speed. On the real thing after 6 minutes,
number of positions 2026 for excal, and 2028 for excaltm. number of positions is 2026 for excal, and 2028 for excaltm.
TODO: TODO:
- verify CPU speed, see notes above - verify CPU speed, see notes above
@ -102,7 +108,7 @@ private:
void excal_state::machine_start() void excal_state::machine_start()
{ {
// HACK: slow down CPU to account for suspected wait states // HACK: slow down CPU to account for bus arbiter and wait states
m_maincpu->set_clock_scale(5.0 / 8.0); m_maincpu->set_clock_scale(5.0 / 8.0);
} }