nanochess

For some reason I always wanted a CIB Atari game and I was fortunate to get Asteroids sealed from atari2600.com Believe me, one of the best experiences is to open it and see everything as new Also I casually got a Defender for Colecovision CIB almost new and one of my guilty pleasures is to open it from time in time and play some games

Great! then you almost have it! download a Z80 pinout and Colecovision cartridge pinout and try to follow PCB trace for A14 with a multimeter. Also there is a 74LS138 that could be the culprit as it decodes the upper address range ($c000-$ffff) Check Colecovision schematics at http://atarihq.com/danb/Colecovision.shtml

I've visited several times a flea's market, and there is a guy that sells mostly Nintendo 64 cartridges and he has some Atari cartridges from time on time. When I come and he see me he sometimes says "how are you? now I don't have of the rare Atari cartridges you like"

I'm waiting for this one. It looks amazingly good!

Do it as comic collectors, buy two! one for collecting and one for playing. I'm really upset when I buy things looking as new but in reality them are resealed, sometimes them have scratches not visible at first glance. So now I'm very careful with source vendor.

I have got my e-mail confirmation just after payment.

By some reason the website has the text twice!

Almost I've forgot it, have you tested Defender?, this cartridge uses 24K of addressing space, just to check if cartridge is being accessed right. It could be so simple as a bad address line. I remember there are 32K cartridges, but at this moment I cannot remember one.

Ok, I've opened my Colecovision, it has a green PCB and I had the surprise it uses a NEC PD780 CPU, even older than yours, mine is from 1982. Still yours could be oxydized or slightly damaged (data bus is not protected), but anyway you have a good start for options (voltage measurement, Z80, 74LS04 and 74LS138)

I've some theories about the non-working feature: 1. One of your chips is so oxydized that is eating more energy than necessary (you could check it measuring voltage in both Colecovision) 2. One of the decoder chips 74LS138 / 74LS04 is oxydizing so it became slower. 3. The Colecovision that doesn't work uses a NEC PD780 CPU, it has been recognized that it is a Japanese reimplementation that behaves slightly different than a real Z80A (Z8400A), BTW the Z80 without A is limited to 2 mhz. My Colecovision has a green board and works just fine, I'm going to open it just to make sure what processor it uses, as my theory 3 is just plain interesting and looks promising.

Why not to take detailed pictures of both mainboards? and open another thread to publish them, it could be a PCB variation or simply that there is some different slower chip in your non-working Colecovision.

That's great. So far it looks like every Coleco standard release just runs slower in PAL. For homebrew releases PkK has commented that he checks the 50hz/60hz byte to change speed accordingly, I do same in my games, so music and gameplay are unchanged between PAL/NTSC.

The manual doesn't say anything about taking back a move. I suppose you can edit the board...

No, I'm not feeling bad. Just somewhat surprised that I've touched it! (I couldn't recognize the type of PCB at that time because I never had an Atari!!!) Anyway thanks. We have a proverb that when something isn't for you, you simply cannot get it. I've been watching carefully that flea's market and some others. And maybe with a little of luck...

Yep, it was at Mexico, metropolitan area surrounding the federal district.

Some two or three years ago I saw a little blue box with a handle in top it had a damaged label, I took it for a calculator part. This was in a flea's market and I didn't bought it. So previous year when I finally started buying Atari and learning to recognize carts, well, I got astonished when I saw the pictures. I can only say holy shit! I missed a one very interesting cartridge for mere 8 dollars.

Well, I just have gave a look to Super Cobra and the problem is that Konami made a whole mess in RAM memory access. You know the Colecovision has 1K of memory repeated over 8K $6000-$7FFF, if you access $6000 is exactly the same if you access $7000. But in ADAM, the $6000-$7FFF zone has independent 8K of RAM, so when Super Cobra tries to find data it fails miserably because it uses $6000-$63ff and $7000-$73ff sparely. This would need a complete analysis to relocate every single instruction wrong o.O! Maybe if I work in small blocks by week... I don't promise anything, it's a horribly complicated job with big chances to insert bugs.

Really it also doesn't work? someone could give some light on this?

I'm glad you could test it on your PAL Colecovision. By the way, each of my games Zombie Near, Princess Quest and the future Mecha Eight are ready for PAL consoles so music doesn't slow down.

I was pretty surprised by this. So I've checked your dump and you missed three bytes between $0552-$0554 in a way that put weird sound bytes at GAME OVER. Thanks for putting your dump, otherwise I would have been searching for a ghost But fortunately NIAD dump is right , so everyone enjoy it .

This patch detects automatically if there is a PAL Colecovision and adds delay, but not for NTSC Also added delay to solve bug with controllers in Coleco ADAM (not tested, but it should work) If anyone publishes a patched ROM, please DON'T give me any credit. It is only my little contribution to the community. Address Original byte New byte $000C = $00 -> $3A $000D = $00 -> $69 $000F = $00 -> $FE $0010 = $00 -> $32 $0011 = $00 -> $20 $0012 = $00 -> $08 $0013 = $00 -> $01 $0014 = $00 -> $BA $0015 = $00 -> $01 $0016 = $00 -> $0B $0017 = $00 -> $78 $0018 = $00 -> $B1 $0019 = $00 -> $20 $001A = $00 -> $FB $001B = $00 -> $C3 $001C = $00 -> $75 $001D = $00 -> $DD $0172 = $75 -> $0C $0173 = $DD -> $80 $020F = $3E -> $D3 $0211 = $D3 -> $E3 $0212 = $80 -> $E3 $0551 = $3E -> $D3 $0553 = $D3 -> $E3 $0554 = $80 -> $E3 $074D = $00 -> $E3 $074E = $00 -> $E3 $2FEB = $3E -> $D3 $2FEC = $80 -> $C0 $2FED = $D3 -> $E3 $2FEE = $C0 -> $E3 $3019 = $3E -> $D3 $301A = $80 -> $C0 $301B = $D3 -> $E3 $301C = $C0 -> $E3 $304B = $3E -> $D3 $304C = $80 -> $C0 $304D = $D3 -> $E3 $304E = $C0 -> $E3 $306D = $3E -> $D3 $306E = $80 -> $C0 $306F = $D3 -> $E3 $3070 = $C0 -> $E3 $324F = $3E -> $D3 $3250 = $80 -> $C0 $3251 = $D3 -> $E3 $3252 = $C0 -> $E3 $327F = $3E -> $D3 $3281 = $D3 -> $E3 $3282 = $80 -> $E3 $32B0 = $3E -> $D3 $32B2 = $D3 -> $E3 $32B3 = $80 -> $E3 $32C7 = $3E -> $D3 $32C9 = $D3 -> $E3 $32CA = $80 -> $E3 $32D6 = $3E -> $D3 $32D8 = $D3 -> $E3 $32D9 = $80 -> $E3

This trigered my curiosity and I gave another look. The original ROM doesn't have waiting time between selecting joystick/keyboard ports, so that could be the problem.

And though I'm following every thread as time allows me, please keep me informed if you spot something wrong in the list

Check the updated list (the first post), I've been editing it recently. Several Pixelboy games are being re-released (marked with BATCH) and others are permanently sold-out. Pixelboy has said that he always makes 150 boxes (the minimum required by printer company) for each game and does batchs between 75 and 100, so there is a chance that re-releases can be done, except when there are no more boxes. By the way, I've updated again the list (SEE POST #1) though Princess Quest (that's mine ) and Thexder for SGM are new, the 75 original copies for each one have been sold-out and he is taking note of interest for another batch.

By the way, you know it works because when running in BlueMSX (NTSC) the sprites from the lower part (ships and enemies) will get hidden.