NivagSwerdna
Active member
OK so next challenge...
and
and
Konami GX319 - Untested and Incomplete
- Thread starter NivagSwerdna
- Start date
NivagSwerdna
Active member
Most obvious problem is the splayed/destroyed socket and missing device.
NivagSwerdna
Active member
Read the ROMs... E02 and L04 read differently each time so definitely dodgy...
Comparing to my local copy of MAME...
The two Dodgy ones are dodgy so no point really considering them. I see a mismatch for my 319_L03 at 6H.
My 319_G12_C8.BIN matches MAME faad...863e but interestingly gives a different answer at http://romident.coinopflorida.com/uploadrom.php where it says...
I don't think that is correct. Pretty sure that is the correct image for megazone.
So the $1M question... do I erase and re-flash just the dodgy and mismatching ones, or replace, or burn 'em all?
Code:
64f54621487291fbfe827fb4cecca299fd0db781 *319_E01_A3.BIN
2ba28b03b4310aecf54b80ef60b0df0dabed6805 *319_E02_D5_DODGY.BIN
fc8c0ecd3b7f03d63b6c3143143986883345fa38 *319_E08_D2.BIN
4c038e80d575988407252897a1f1bc6b76af597c *319_E09_E2.BIN
be2dd46cd0121cedad6dab90a22643798a3176ab *319_E10_D3.BIN
210aae91a3838e5f7c78747d9b7419d266538ffc *319_E11_E3.BIN
faadcb20ee8b26b9ab0692df6a81e5423514863e *319_G12_C8.BIN
1f6fbf804dacfa44cd11b4cf41d0bedb7f2ff6b6 *319_G13_C10.BIN
44cad3c07aa9fb36be5f5b9d0ce0828495c92619 *319_L03_H6.BIN
26c513c2d78f6c0fca07323a73d322a6d05496e3 *319_L04_H7_DODGY.BIN
d6373f5be06cd61b6d3ffbe36c160167ba9852f3 *319_L05_H8.BIN
4c8688b7bde5ee1adfe6a0db7178cca046eca7f4 *319_L06_H9.BIN
ecdcdd085020f8ffe87a574832f7cedb9bcacef9 *319_L07_H11.BINComparing to my local copy of MAME...
Code:
ROM_START( megazone )
ROM_REGION( 2*0x10000, "maincpu", 0 )
ROM_LOAD( "319_l07.11h", 0x6000, 0x2000, CRC(73b616ca) SHA1(ecdcdd085020f8ffe87a574832f7cedb9bcacef9) ) OK
ROM_LOAD( "319_l06.9h", 0x8000, 0x2000, CRC(0ced03f9) SHA1(4c8688b7bde5ee1adfe6a0db7178cca046eca7f4) ) OK
ROM_LOAD( "319_l05.8h", 0xa000, 0x2000, CRC(9dc3b5a1) SHA1(d6373f5be06cd61b6d3ffbe36c160167ba9852f3) ) OK
ROM_LOAD( "319_l04.7h", 0xc000, 0x2000, CRC(785b983d) SHA1(389e2f5494284089d39249e91293f3998c2b22c0) ) DODGY
ROM_LOAD( "319_l03.6h", 0xe000, 0x2000, CRC(a5318686) SHA1(8c8fbb76a36108f4a0b3a50e8a9b1781508490f1) ) MISMATCH
ROM_REGION( 0x10000, "audiocpu", 0 )
ROM_LOAD( "319e02.6d", 0x0000, 0x2000, CRC(d5d45edb) SHA1(3808d1b58fe152f8f5b49bf0aa40c53e9c9dd4bd) ) DODGY
ROM_REGION( 0x1000, "daccpu", 0 ) /* 4k for the 8039 DAC CPU */
ROM_LOAD( "319e01.3a", 0x0000, 0x1000, CRC(ed5725a0) SHA1(64f54621487291fbfe827fb4cecca299fd0db781) ) OK
ROM_REGION( 0x08000, "gfx1", 0 )
ROM_LOAD( "319e11.3e", 0x0000, 0x2000, CRC(965a7ff6) SHA1(210aae91a3838e5f7c78747d9b7419d266538ffc) ) OK
ROM_LOAD( "319e09.2e", 0x2000, 0x2000, CRC(5eaa7f3e) SHA1(4c038e80d575988407252897a1f1bc6b76af597c) ) OK
ROM_LOAD( "319e10.3d", 0x4000, 0x2000, CRC(7bb1aeee) SHA1(be2dd46cd0121cedad6dab90a22643798a3176ab) ) OK
ROM_LOAD( "319e08.2d", 0x6000, 0x2000, CRC(6add71b1) SHA1(fc8c0ecd3b7f03d63b6c3143143986883345fa38) ) OK
ROM_REGION( 0x04000, "gfx2", 0 )
ROM_LOAD( "319_g12.8c", 0x0000, 0x2000, CRC(07b8b24b) SHA1(faadcb20ee8b26b9ab0692df6a81e5423514863e) ) OK
ROM_LOAD( "319_g13.10c", 0x2000, 0x2000, CRC(3d8f3743) SHA1(1f6fbf804dacfa44cd11b4cf41d0bedb7f2ff6b6) ) OK // same as e13
ROM_ENDMy 319_G12_C8.BIN matches MAME faad...863e but interestingly gives a different answer at http://romident.coinopflorida.com/uploadrom.php where it says...
I don't think that is correct. Pretty sure that is the correct image for megazone.
So the $1M question... do I erase and re-flash just the dodgy and mismatching ones, or replace, or burn 'em all?
Last edited:
NivagSwerdna
Active member
I put three of the UVEPROMs (the two DODGY and one MISMATCH) in my Big Moon cheapo UV Eraser and after two blasts of 60mins they still aren't completely blank... I'll give them another blast and if not will give up on them then.
The broken socket is shown as "Konami-2" on the schematic.
Have managed to get the broken socket out after much trouble with pins 14 and 28; assuming those pins are connected to copper pours.
I think the missing device is a Konami 083 which is a shift register thing...
Basically 16 bits in... and then shifted out 4 bits at a time... Once I have the ROMs sorted will move on to that.
You can sort of guess how it works from the schematic... 2x8bits IN, loaded by a signal on pin 2. Pin 1 must be for clocking it out and pin 27 probably direction. That's my theory anyway.
The broken socket is shown as "Konami-2" on the schematic.
Have managed to get the broken socket out after much trouble with pins 14 and 28; assuming those pins are connected to copper pours.
I think the missing device is a Konami 083 which is a shift register thing...
Basically 16 bits in... and then shifted out 4 bits at a time... Once I have the ROMs sorted will move on to that.
You can sort of guess how it works from the schematic... 2x8bits IN, loaded by a signal on pin 2. Pin 1 must be for clocking it out and pin 27 probably direction. That's my theory anyway.
cliff_poole
Active member
If you can wait until the weekend, I think I have a bootleg Time Pilot with a PCB replacement for the 083. I'll take some pictures of it if that is the case.
NivagSwerdna
Active member
That would be interesting. I'm hoping it is 4-off 74194. I'm not in a hurry.
NivagSwerdna
Active member
This is weird... I took the 3 UV EPROMs, removed the labels and erased them for 50mins... not blank.... another 50mins... not blank... mostly blank not not blank... another 50mins... not blank!
Mostly blank... but not blank...
and...
etc.
I've never actually seen this before. Could they be all that bad? The UV Eraser seems to be glowing merrily so I assume that is OK.
I guess I will just bin them and use some others.
Mostly blank... but not blank...
and...
etc.
I've never actually seen this before. Could they be all that bad? The UV Eraser seems to be glowing merrily so I assume that is OK.
I guess I will just bin them and use some others.
NivagSwerdna
Active member
OK... so fear, uncertainty and doubt got the better of me... so I took out all the UV EPROMs, removed the labels and put them for a 50min spin in the eraser... They all blank check fine except the three which were misbehaving... so the naughty three are destined for the bin and now I have rather a lot of programming to do once I get my programmer out of the loft! I'll use three 'new' ones in place of the naughty ones... I have some NMC27C64Q-200 which should be OK.
cliff_poole
Active member
Interesting that I've replaced a chip in the past. I started logging all my repairs about twenty years ago and I have no notes on this, so must have done it a very long time ago.
NivagSwerdna
Active member
That's interesting. At a high level it does what I expected... ie. 4x 74LS194... no idea what the other bits do.
NivagSwerdna
Active member
OK... well that was a nightmare... of the three intermittent ROMs I went around a few iterations before I convinced myself they were unreliable and then with a few issues with my programmer I managed somehow to kill a couple of extras and in the end I had to replace a few more.... Anyway after several days of tearing my hair out I think I have a complete set that reads reliably!
NivagSwerdna
Active member
So this weeks job has been to try and create a verilog version of a Konami 083...
I think I'm getting close...
This is a real one in the test bench.... (The black line is GND... i.e. Pin 14)
and this is my verilog equivalent...
obviously it has no pin 14 but otherwise it's looking pretty good.
I'll try a hardware version next week when I have had time to build a JTAG adapter (and fixed my bicycle).
I think I'm getting close...
This is a real one in the test bench.... (The black line is GND... i.e. Pin 14)
and this is my verilog equivalent...
obviously it has no pin 14 but otherwise it's looking pretty good.
I'll try a hardware version next week when I have had time to build a JTAG adapter (and fixed my bicycle).
NivagSwerdna
Active member
Burnt into CPLD...
Testing gives expected output...
Time to move back to the game PCB!
PS
Bicycle fixed
NivagSwerdna
Active member
A bit of testing in my recently fixed Kicker...
083 Custom at A9 removed...
Original Konami 083 in A9...
New Konami Replacement 083 in A9...
I think we are good! Maybe the definition on the background is not as defined... maybe that is the screen capture?
Back to the MegaZone...
083 Custom at A9 removed...
Original Konami 083 in A9...
New Konami Replacement 083 in A9...
I think we are good! Maybe the definition on the background is not as defined... maybe that is the screen capture?
Back to the MegaZone...
NivagSwerdna
Active member
OK... checked again... renders fine! Phew!
Now I can move forward!
Now I can move forward!
Last edited:
NivagSwerdna
Active member
Put a replacement socket, put in my new custom, replaced ROMs...
Wired it up...
OMG...
Fantastic!
£50 and a bit of time and I have a Megazone... how cool is that!
Wired it up...
OMG...
Fantastic!
£50 and a bit of time and I have a Megazone... how cool is that!
cliff_poole
Active member
This sort of thing amazes me, Well done !. Any chance you could do a tutorial ?
. Are you using one of those Saleae logic analysers ?.
. Are you using one of those Saleae logic analysers ?.NivagSwerdna
Active member
Cliff,
This one is relatively straightforward as Konami produced daughterboards for some of these chips as seen on the TimePilot. The daughter boards are often incomplete in that they only drive a subset of functionality but they do do the core function. Also in the case of a 083 it is pretty obvious what it does from the schematic... it's 16 bits in and 4 bits out in various orders. I have a homebrew chip tester that allows me to tickle the pins to try out theories with a real device (which is how I produced the 28 channel diagrams earlier); it works at slow speed but does manage to tickle some of the functionality.
I'm also looking at the 085 which is pretty much just an adder to calculate scroll for backgrounds... From a functional point of view it is pretty obvious but it has many pins that I haven't worked out yet and some pretty weird behaviour which is possibly most irrelevant. I haven't seen a diagram for a 085 daughter board so I am interested in that.
Basically if the device has state (i.e. registers) then it is very hard to work out the internals; if it is just combinatorial then its easy as you try all the combinations and then reduce to a set of equations.
Although its a bit silly my ultimate goal is to understand the 084 and it has always annoyed me how MAME doesn't do the Tutankham star field correctly! It's been on my to do this for a number of years and I am getting closer to having the time and tools to look at it. The 084 should really just be a LFSR of some kind since that is how most random number generators and star fields are done; again... anything stateful is hard but not impossible if you have some clues.
I knew the 083 was do-able so that's why I started there!
Regards
Nivag
This one is relatively straightforward as Konami produced daughterboards for some of these chips as seen on the TimePilot. The daughter boards are often incomplete in that they only drive a subset of functionality but they do do the core function. Also in the case of a 083 it is pretty obvious what it does from the schematic... it's 16 bits in and 4 bits out in various orders. I have a homebrew chip tester that allows me to tickle the pins to try out theories with a real device (which is how I produced the 28 channel diagrams earlier); it works at slow speed but does manage to tickle some of the functionality.
I'm also looking at the 085 which is pretty much just an adder to calculate scroll for backgrounds... From a functional point of view it is pretty obvious but it has many pins that I haven't worked out yet and some pretty weird behaviour which is possibly most irrelevant. I haven't seen a diagram for a 085 daughter board so I am interested in that.
Basically if the device has state (i.e. registers) then it is very hard to work out the internals; if it is just combinatorial then its easy as you try all the combinations and then reduce to a set of equations.
Although its a bit silly my ultimate goal is to understand the 084 and it has always annoyed me how MAME doesn't do the Tutankham star field correctly! It's been on my to do this for a number of years and I am getting closer to having the time and tools to look at it. The 084 should really just be a LFSR of some kind since that is how most random number generators and star fields are done; again... anything stateful is hard but not impossible if you have some clues.
I knew the 083 was do-able so that's why I started there!
Regards
Nivag
cliff_poole
Active member
I don't have any bootlegs with an 085 I'm afraid. I've seen bootlegs of Tutankham but never owned one, not sure if it uses a daughterboard.
I've never seen a bootleg of any game with the 504 which is a shame because I could do with a couple of those. It divides up the clock for the main processor and provides all the video timing signals. My ABI chip tester is one of the older ones with the 24pin socket, but I guess if the 504 has enough unused pins I could bodge together an adapter. Maybe one day.
I've never seen a bootleg of any game with the 504 which is a shame because I could do with a couple of those. It divides up the clock for the main processor and provides all the video timing signals. My ABI chip tester is one of the older ones with the 24pin socket, but I guess if the 504 has enough unused pins I could bodge together an adapter. Maybe one day.