Subelectro program card info needed

guddler

Busting vectors like it's 1982!
vacBacker
Feedback
10 (100%)
Credits
4,058CR
I dug some of my PCBs out the loft tonight and I remember that one of my favourites is not working.

SO, a Tazzmania bootleg. It runs on a Subelectro board using their program card / cart type thing / system. Problem is the cart has two rows of dipswitches on it that I'm 99% sure configure at least the address space, if not some more config of the card. And I know they've been fiddled with.

Does anyone have ANY information on these dip settings please?

And before anyone thinks about it, it won't be any good going off and looking up "Tazzmania Dip Switches" as that will just bring up the normal dip settings for the original Stern version of the game, not the Subelectro program card. Not that I won't be grateful of people looking, but obviously the normal dip settings I don't THINK will be any good.

Here is the cart in question...

It's the 2 sets of 6 that I need info on. The ones you can just about see at the top of the pic will be the normal ones.

tazz_card.png
 

Macro

Active member
vacBacker
Feedback
4 (100%)
Credits
2,010CR
Hmm, not sure I have any like that

I think mine are the Artic playpacks (which conveniently have stickers on the carts showing you the switch positions needed to play it!)

will have a double check for you at the weekend.
 

guddler

Busting vectors like it's 1982!
vacBacker
Feedback
10 (100%)
Credits
4,058CR
Cheers mike. If the ones you have do have stickers on them then the info could still be useful. In theory I could probably use it to help me trace the tracks and work out what on earth those dips actually do. For sure they do something with the Video rom control and I'm pretty sure they mangle the address lines too. It's almost like it's a primitive encryption key! (very primitive mind)
 

guddler

Busting vectors like it's 1982!
vacBacker
Feedback
10 (100%)
Credits
4,058CR
Don't suppose I can nudge you to see if you can find your board can I?

I manage to work out the following by dumping the roms and comparing them. The sound rom doesn't match at all but by process of deduction I can tell which one it is. The GFX roms match 100% to standard Tazzmania. The program roms match around 30% which is enough for me to tell which is which (and their locations make sense too). I am comparing them to some other random set in my dumped roms folder mind which is labelled as "TazzmanG" that may or may not be the galaxian hardware set from mame, or it could just be some random board I've dumped in the past, I need to check.

So I don't know if the 30% match is because I'm matching against the wrong set or if they are "encrypted". There's certainly no readable text in there so I'm going with crude encryption.

I've also determined that the fault probably lies on this program card. If I put the galaxian test rom in the board it runs the tests fine. It fails object ram and video ram tests but I very much suspect the ram is not where galaxian thinks it is, it's more than likely moved.

I can also see that the lower (right hand in the pic) set of dip switches does indeed affect the addressing of the program board. With my scope on an LS138 in the lower right of the rom board if I change the dips I will get or lose activity on the 138 pins.

So I'm basically screwed until I can work out what the appropriate dip settings are.

Also really interesting and odd is that there is a prom in the bottom right corner of the ROM board right next to the LS138. But I can't see any track going to or from it on either side of the board and it doesn't appear to be getting any power. Most odd indeed!!!

tazz_card_roms.png


From bottom to top in the far right of this picture (bottom left to right as the board is on the bench - I was lazy and didn't rotate the photo) there is a PAL16?8, a 245, SND1, LS138 and mystery prom.guddler2015-04-03 19:15:52
 

Macro

Active member
vacBacker
Feedback
4 (100%)
Credits
2,010CR
I would hazard a guess that it is meant to be the colour prom for the game, and should be fitted in L6

the socket is just a carrier to keep it together with the game

also, why a sound rom - does it have a second Z80 and an AY soundchip to make it more scramble like, which would explain why the ram mapping is different.

The ones I have are not the same as that, I think they are all Artic ones - no sound code, just makes do with standard galaxian noises.
 

guddler

Busting vectors like it's 1982!
vacBacker
Feedback
10 (100%)
Credits
4,058CR
Yeah, the addon board is a two board sandwich. Hiding underneath the one you can see is a slightly bigger one which does indeed hold an additional Z80 and AY chip. The ROM board plugs into that. It's quite a beast of an add-on. It has two ribbon connectors that go all the way to the character roms at the left edge of the board.

Just as well that's just a carrier to keep the prom safe. I too notice it was plugged in the wrong way round earlier
smiley36.gif


It's probably the original galaxian one. I'll have to dump it to be sure.

I've dumped the ROMs. I may have to try and work out the memory map and see if I can work out what the dips are doing then if you think it's too different to yours. I was being lazy and trying to avoid removing the board's power cables and I guess also the additional RAM from the main board as it's the usual soldered piggyback stack which is a right nuisance if one of them dies.

I so want to get this going again though - it did used to work!
 

Macro

Active member
vacBacker
Feedback
4 (100%)
Credits
2,010CR
so it is basically "upgrading" a galaxian PCB into a scramble PCB

so I would expect it to have a scramble or Moon Cresta memory map to make transposing the game easier.
 

guddler

Busting vectors like it's 1982!
vacBacker
Feedback
10 (100%)
Credits
4,058CR
My main concern is that pal. I'm sure it's doing a crude job of encryption. When I read it though half of it is all 1's the other half all zeros so either the pal has gone bad or maybe that's what you get when the security bit is set?

I really need to sit down and look at the rom dumps
 

guddler

Busting vectors like it's 1982!
vacBacker
Feedback
10 (100%)
Credits
4,058CR
Plodding along with this. Just sat down and analysed the ROMs. Have confirmed this is an "Encryoted" version of Tazmani2 from Mame. Graphics ROMs are identical, Sound ROM is not encrypted but is also not the same as the one in mame. Not too bothered about that at this stage.

So I've spent a while looking at the way the ROMs are monkey'd about with. Bits 3 & 7 are sometimes flipped if the value is greater than 0 but I can't yet come up with any rhyme or reason to it. Sometimes both bits are flipped, other times one or other of them, other times none of them. And it's not a straight forward lookup table because the same value in the original ROM has multiple values in my version.

So I disassembled the code using mame in the hope that maybe I could see a pattern, something like if M1 then flip bits, etc., etc., but alas, can't see anything still. I'm confident it's not just bit rot because it occurs on ALL cpu ROMs and is always bits 3 or 7.

I think my next step then now I know what the memory map is and what the ROM values should be when presented to the CPU is to get it on the fluke and do some loops on ROM addresses and see if the dip switches affect the values received. If that fails I guess I'll have to map out the schematics of the cart and carrier board see exactly what CPU or ROM pins the DIPs and PAL are linked to.

As much of a pain in the arse this is, it's quite interesting. Even if nobody else is listening to my ramblings
smiley36.gif
 

Macro

Active member
vacBacker
Feedback
4 (100%)
Credits
2,010CR
Several games of that time use one or more of the address lines to indicate whether to xor the byte or not - moon quasar being one that I cracked many years ago has a different byte to xor every 64 bytes (but in a pattern, so once you have the first few sorted, the rest just make sense!)

so map out some of the known bit changes against address and see if a pattern emerges - and see what address lines go to the PAL
 

guddler

Busting vectors like it's 1982!
vacBacker
Feedback
10 (100%)
Credits
4,058CR
I feel a little bit like I'm fighting a losing battle here! I had it on the fluke last night which wasn't easy in itself because the Z80 is orientated the wrong way for the ribbon connector of the pod. Anyway, after much messing about I felt I was managing to get somewhere as I managed to get the CPU to see the correct data values but it still wouldn't run.

I was in the process of looking at memory when the damn sync died! Only video sync, I've still got a picture so not a big deal but that's the second fault in the timing circuit since I've been trying to get it going.

Anyway, what I did determine is that four of the dips control A14 and A15. from left to right, 3&4 are both tied to one and 5&6 are both tied to the other (I forget which way round and my notes are not near me). I'm not sure what exactly they're doing with them but if I loop on a bus test I get A14/15 not drivable with the dips in one position, success in the other.

I haven't yet mapped out what goes to the PAL. I'll check what you said though as what you say sounds like it could easily be what's going on with the bits and they could be related to A14 & A15.

In addition, I'm not convinced everything is correct with the addressing. According to mame, 8000-87FF is meant to be program ram. And that ties up with the fact there are 4 2114's (2x2 soldered on top of each other), yet if I do a me test of that area, sure it passes but the video ram is changing on screen at the time so that doesn't seem right.
 

guddler

Busting vectors like it's 1982!
vacBacker
Feedback
10 (100%)
Credits
4,058CR
Actually thinking about it, not sure that A14 & A15 can be the driving force behind what is flipped and what is not as the flipping occurs in instructions as well as operands and the instructions are obviously all down in ROM where both A14&15 are going to be 0.

Like you say, will map it out and have a look at what is going to the prom. That's not easy in itself though as the prom is on the top card (the cart) and the Z80 is on the bottom card (the carrier). Hopefully I can ping it out without desoldering it all.
 
Top