Master System Boot (Sega Mega-Tech)

there is only really one main stage of boot on the sega mega tech and that is to get the system operational after reading the bios from the sms and running the top menu, and reading the game from the mega drive system and running that on the second monitor the main focus on master system boot is bios (ic20 m27c256) z80 sms cpu (ic17) z80 sms ram (ic27 ic28 mb8464a) sms 2 vdp (ic14 315 5346) sms vdp (ic16 315 5246) sms vdp ram (ic1 hm65256blsp 10) sms vdp ram (ic2 315 5330) i/o expander (ic7 cxd1095) 315 5348 (ic30 pal16l8bcn) 315 5347 (ic31 pal16l8bcn) all of these and a few more parts are needed for successful boot the most common cause of the failed video out is bios, z80 or io expanders successful boot for detecting boot, you can check if you have the top menu on screen if you do and the insert coin is flashing between that and push select or enter button if it is, you have a fully running sms sub system operating the top menu note by default at boot the time doesn't count up, sits at 00'00" detecting boot (z80 !wait pin) the first part of the board you should look at bringing up / testing is the menu bios there is sometimes a daughterboard containing a z80, ram and some rom it's job is to bring up the game list on the second monitor it handles game selection, coinage, time keeping and so on this daughterboard is based on sega master system the board has the z80 cpu , and everything else appears to be used to simply output diagnostic information over the db port , so is not necessary for operation you can remove the z80 and install it directly in the socket below where the daughterboard connected, to have it directly on the main motherboard, like so the simplest way to test boot is monitor the !wait pin of the z80 (ic17) the !wait pin loading / reading bios pulses low for about 1 second fully booted dc 5v high corrupt ram/bios pulses low constantly crashed dc low note the !wait can also go high but crash later and remain high typically loading of the bios takes just over a second from pulsing !wait pin to dc high and game load / main menu shows z80 activity pins if you are seeing a failed boot, you can observe the rest of the pins to see what pins or traces may be faulty below are the expected behaviours of the pins pin loading fully booted crashed clk 3 58mhz 5v clock 3 58mhz 5v clock 3 58mhz 5v clock address pins pulsing repetitive 3 6v pulsing repetitive 3 6v dc low data pins pulsing repetitive 3 6v pulsing repetitive 3 6v dc 2v !refresh pulsing random 5v pulsing random 5v dc high 3 6v !m1 pulsing repetitive 3 6v pulsing repetitive 3 6v dc high 3 6v !reset dc high 5v dc high 5v dc high 5v !busrq dc high 5v dc high 5v dc high 5v !wait pulsing low every 16 67us dc high 5v dc low !busak dc high 3 6v dc high 3 6v dc high 3 6v !write pulsing random 5v pulsing random 5v dc high 5v !read pulsing repetitive 3 6v pulsing repetitive 3 6v dc low !ioreq pulsing random 3 6v pulsing random 3 6v dc 3 6v !mreq pulsing repetitive 3 6v pulsing repetitive 3 6v dc low !halt dc 3 6v dc 3 6v dc 3 6v !nmi dc 5v dc 5v dc 5v !int pulse low every 59 93hz pulse low every 59 93hz dc low also the z80 sms ram (ic27 ic28 mb8464a), bios (ic20 m27c256) and sms 2 vdp (ic14 315 5346) all have 3 6v pulsing going on while running , and flat dc when crashed checking ics z80 (sms cpu ic17) can be tested using a z80 tester, or perhaps chuck it in a known working console such as a neo geo aes to fully test it bios (m27c256b 12f6 ic20) can be placed into an eeprom reader to read it out and compare its crc value to a known bios file sms ram (mb8464a ic27 / ic28) can be checked by installing it into a known working mega drive 1 sms 2 vdp (pal16l8bcn / 315 5346 ic14) are much harder to test, almost impossible, and the best option is to swap them between a working mega tech to prove them you can find dumps of their code here sms vdp (315 5246 ic16) can be checked by installing it into a working sms sms vdp ram (315 5330 ic2) is pseudo ram that cannot be easily checked except placing it into a working mega tech apparently you can find 315 5330 in some master systems however i have personally never found one in a master system i/o expander (cxd1095 ic7) can only be checked by oscilloscope or swapping with the alternative mega drive i/o expander (ic24) on the same mega tech board presuming the second one works these chips are still available on the market to purchase plas (pal16l8bcn ic30 / ic31) are much harder to test, almost impossible, and the best option is to swap them between a working mega tech to prove them you can find dumps of their code here often a failed ic30/31 leads to black screen with actual csync pulses working, and the mega drive being brought out of reset system clocks make sure you have working clock signals by probing with an oscilloscope the 10 7386mhz and 53 693mhz crystals i/o expanders after this i would check your i/o expanders for general activity (pulsing 5v) the important thing is that you see activity on the pins, not necessarily the actual data being sent however, activity on the pins can happen even when it is faulty i have seen in on multipe, they always seem to have activity even if dead you could more accurately look at the pinout and probe both inputs and outputs to see if the input pins toggle in and then come out of the associated outputs the cxd1095 is an off the shelf port extender the bios and z80 can access two of them, to handle inputs and drive outputs if these are bad you will get a black screen often it is simplest just to swap them with a new / known working one to test also excessive heat coming from them is another sign of fault, although they also fail without any heat sometimes too (if they have already burned out) the i/o expanders are cxd1095q https //archbee doc uploads s3 amazonaws com/9enoacbhz7zqgrsars4gt/oum5trrefkv6ua zov8qa cxd1095q pdf you should see !reset pin high , and dx and ax pins pulsing with data for example this is a good sign the chips are generally functional if you see no activity , but !reset is high , then the i/o expander could be faulty !write and !read pulse like the z80 and should be pulsing all the time !cs should be high if !reset is low , the chips are held in reset so the reset circuit needs to be fixed first (i suspect its related to the pld ic14 to the top left of the ic7 i/o expander checking for screen output if all the above checks are ok, you can probe the cn9 connector (rgb2) for the top menu screen this is ttl level rgbs (5v logic) when disconnected the sync and rgb pins are just below 5v, and when connected to a receiver drop to around 2 2 2 5v the video output has rgb output, using csync as the vsync/hsync timing the rgb is standard, so is csync csync is basically a combined hsync and vsync signal hsync is small pulses every 16 63us the full frame is 59 9hz refresh rate the vsync is signalled by three long low pulses of 64us (instead of the usual 4 8us) the rgb data is the same (5v ttl logic) sms vdp the video display processor from the sega master system is the 315 5246 it works alongside a custom sega 315 5330 and some sram it is worth checking all your connections between the three chips and the z80 are all good issues / diagnostics below are some common causes and fixes to issues no screen if you get nothing at all on the csync or rgb pins on cn9, check your vdp (315 5246) as this needs nothing else but itself and the system clocks and power rail to generate a csync signal black screen a common issue if you get a valid csync outputting 59 91hz, but the rgb pins are all close to 0v giving a black screen, is a bad io expander (cxd1095 ic7) also a bad io expander on ic24 (to read game cartridges) can cause the system not to boot remove it completely and do not replace it until you get to the menu screen the menu screen will boot without ic24, but it may not boot if you have a bad one installed it can also be any of the items listed at the top of this article as they all need to work to get past black screen however i have noted over 80% of the time it has been the io expanders being the issue white red bars this is a common sign your sms vdp ram (ic1 hm65256blsp 10) is bad less common is the sms vdp (ic16 315 5246) or sms vdp ram (ic2 315 5330) often as a quick test you can sometimes use a hot air station to warm up the ram chip quite a bit and then try and power on if it is faulty sometimes it changes from white and red bars to some corrupt data graphic glitches almost all graphic glitches are caused by sms vdp (ic16 315 5246) sms vdp ram (ic1 hm65256blsp 10) sms vdp ram (ic2 315 5330) traces between any of these or the z80 broken colors wrong color wrong (yellow background on menu for example) and block glitches are issues with the i/o expander (ic7 cxd1095) usually, or shorting of pins causes this game titles junbled / wrong if you get the game titles showing, but wrong with things like sega in or repeating letters or jargon, it is the mega drives z80 or ram address/data pins or bad z80/ram, or traces from there to their destination visual glitches glitches that make the menu screen glitch visually, but do not stop the games running or the flashing text on the menu screen, are almost always caused by the sms 2 vdp, the 315 5330 or its ram system crashes if the system itself crashes including the menu, this is typically the sms bios, z80 or ram dipping 5v rail if you 5v rail is unstable it can cause issues of many kinds, including random system freezing monitor it with an oscilloscope to check it for clean stable 5v with no more than 300mv of ripple sms 2 vdp the master system 2 video display processor is programmed onto a programmable logic array (pla) chip, which is like a baby fpga in a way it is responsible for the processing of the menu screen data and if it is faulty it can cause the menu to crash a simply test to prove this and simulate a fault is shorting pretty much any pin on the 315 5346 temporarily to ground will cause the top menu screen to stay on but crash (go to fixed screen no insert coin motion)