Jump to content

Eckhard Stolberg

Members
  • Content Count

    953
  • Joined

  • Last visited

Posts posted by Eckhard Stolberg


  1. Actually memory chips get measured in bits, because they don't nessessearily have to come in byte width. So measuring ROM based games in MegaBits isn't that far off.

     

    Anyway, I think the problem with a 7800 multi-cart is the bankswitching. Sean's 2600 multicart only offers linear games with up to 4KB in size. And I think the 5200 doesn't use bankswitching at all. For those games it's relatively easy to build a multi-cart with dip-switches out of an old game cartridge.

     

    But 7800 games used three different types of bankswitching. Also four games require onboard RAM and two games use an extra POKEY chip. To get all of that handled properly you'd need to design a special board and use a programmable logic chip, like Chad Schell did for the Cuttle Cart. This requires special knowledge and tools, and can get very expensive.

     

    If you knew you could sell maybe a hundret copies of this device, then a small production run a la Cuttle Cart might be possible, but the "build them as the orders come in" approach that Sean is using now is probably out of question.

     

    Also to make the design a little easier, you might have to blow up the size of all the games to the maximum used in 7800 games, which is 144KB. For 65 games that is quite a lot of ROM.

     

    As for a combined 2600/7800 Cuttle Cart, I don't see that either. AFAIK the Cuttle Cart uses almost all of the space on the programmable logic chip, and the next bigger chip is quite a lot more expensive. Also you probably wouldn't want to use an audio loading interface for games that big. A 7800 only RAM cart device, might be possible, but Chad Schell has already said on this board, that he currently doesn't have the time to design one, and I don't see anyone else willing and able to pick up the task.

     

    As for the price, the Cuttle Cart is selling for $100 and Chad isn't probably earning to much at that price. A 7800 cart would probably need about three times the RAM that the 2600 version currently uses. Also you would have to have a serial interface and you'd need to fit the POKEY chip somewhere on the board. I think there is a reason why Atari's RAM cart still costs $150.

     

     

    Ciao, Eckhard Stolberg


  2. In DASM you need to set up your code like this to reach your goal:

     

    SEG.U variables

    ORG $300

    v_variable1 DS.B 1

    v_variable2 DS.B 1

    ...

     

    SEG code

    ORG $4000

    l_StartCode:

    SEI

    ...

     

    The SEG command defines a segment for your code or data. For the 6502 this is usually unnessessary, since this processor doesn't use segments like the 8086 for example. But by defining a segment as undeclared (SEG.U) you can define variable or code space that doesn't get included into the binary, which is usefull for defining variables in the internal RAM.


  3. You are right. The ST based devcart uses games in Motorola S-record format or in MADMAC .O format. And you need to have different files for each bank, if you want to use bankswitched games with it. The advantage of these file formats is that you can use the symbolic debugger in the devcart.

     

    But if one day someone really starts working on a new bankswitched 7800 game, I think we would have to look into the possibilities for creating a bankswitched RAM cart too.

     

     

    Ciao, Eckhard Stolberg


  4. I'm not sure if this has already been brought up here, but the music video for Tom Novy's "music is wonderfull" is basically about a couple of people dressed up as Atari VCS joysticks dancing around some female dancers who are playing some generic arcade machines.

     

     

    Ciao, Eckhard Stolberg


  5. The ROM works nicely on z26 1.50. Just as you described it for your console, including the screen flashes and sounds. Maybe you are just using the wrong emulator.

     

     

    Ciao, Eckhard Stolberg


  6. quote:

    Originally posted by Mitch:

    I think the problem is poor coding. I tried several different guns on this game and none of them worked very well.


     

    So, how many lightguns do you have? I'm asking because I was experimenting with lightgun programming on the 2600 a while back. I wrote a little test program based on the gun detection routines from 2600 Sentinel. But even when I held my lightgun directly on the screen, the hits were detected a little bit away from the gun. So I changed the code to work nicely with my gun, but now I'd like to know if it was my gun that is off, or the one used at Atari for programming the games. Maybe you could try out my little test program on the 2600 or 7800 with all your guns, if I'd send you the binary?

     

     

    Thanks, Eckahrd Stolberg


  7. It would probably be possible to build a RAM cart from this, but we haven't had access to one of these boards yet, so we haven't a design available for it. You would have to figure it out yourself.

     

    But a friend of mine is currently working on some other designs for better RAM carts. It will definetely still take quite a while before we have anything to release and I think all of his designs require different boards, but you might probably still like to wait before building yourself a RAM cart.

     

    Since the Scrapyard Dog cart is useless for this purpose, you might want ot follow Mitch's advice for building a Sentinel or KLAX proto, though.

     

    BTW Crossbow, did you get the e-mail I sent you some days ago regarding your timeout problem with 7800ctrl and Win2K?

     

     

    Ciao, Eckhard Stolberg


  8. The "limited" in the company name might be a hint that the copany is from somewhere in the British empire. And with a programmer named Andrew Choi, who writes his name with the family-name first, I'd guess that UA is from Hong Kong. That would also explain, how the ROM for an unreleased game could get into a pirated multi-game system.

     

    If the company was from somewhere in south-east Asia, it might also explain why there are so many PAL clones of the Arcadia, even though the Emerson version wasn't much of a success, and why there were several games released for the system, that weren't available from Emerson.

     

    If Funky Fish was licensed properly from Sun Electronics, then someone who speaks Japanese could ask them, if they know anything about UA Ltd. The software devision of Sun Electronics is now known as Sunsoft, and on one of their webpages I think there was a message about how they even had their games on the early Atari consoles, which refers to Kangaroo, which was licensed to Atari. So there might be a chance, that they still have information from that far back available.

     

     

    Ciao, Eckhard Stolberg


  9. The socket only has 24 pins, so it can only hold EPROMs up to 4KB. The switch might allow you to select between two 2K games on a 4K EPROM, it might enable/disable the inverter for the chip select line, so that you can use EPROMs and normal ROMs in this cart, or it might disable address line A11, so that you can play 2K EPROMs with games that are mapped in at $F000 instead of $F800. But as Dan has said, you'll only know, if you open the cart and look how the switch is wired.

     

     

    Ciao, Eckhard Stolberg


  10. The starpath binary only uses 6K of those 32K. A simple file compare wouldn't work in this case. But the starpath version of Rabbit Transit uses the supercharger type of bankswitching and also some of the supercharger RAM for selfmodifying code. If the Atari version is running on a normal EPROM cart, then the binary must have been modified.

     

     

    Ciao, Eckhard Stolberg


  11. quote:

    Originally posted by -^Cro§Bow^-:

    Also notice that the errors in the hi bits are spaced evenly throughout the whole rom image. Interesting. Too bad it would appear this rom isn't working...here I was wanting a Scrapyard dog...oh well.


     

    Very strange indeed. It could be a problem with overspeaking between some of the address lines and one of the data lines. Maybe you should check all the solder joints and the connections.

     

    BTW, what board does this game use and how are the jumpers (labeled W#) set? And this is a used game, that has been working before, right? I'm asking because I once traded a shrikwrapped copy of Crossbow from a guy in Denmark which didn't work. The guy was nice enough to find me another copy of that game, which showed the same problems. So I opened one of the carts up, and it turned out that they soldered the wrong jumper close, so that the bankswitching wouldn't work and I was getting the graphics all messed up. After fixing the jumpers the game now works nicely.

     

     

    Ciao, Eckhard Stolberg


  12. It's not a PAL version, it's a bad dump. Did you try to dump it a couple of times in a row to see, if you would get different results each time?

     

    Anyway, comparing the ROMs shows that the differences are only in the higher bits of the bytes. This means either some dirty contacts or a timing problem when accessing the ROM, I think.

     

     

    Ciao, Eckhard Stolberg


  13. Matt:

    Yes, I believe I have your prototype ROMs. Unfortunately some of the more interesting ones seem to be bad dumps. But if you like you can post screenshots of the others on your website now.

     

    Christian:

    Sorry, but I don't quite understand what you mean with your PC programming question. But the source code for z26 is available at the site mentioned above. Maybe you can find something usefull in there.

     

    Scott:

    The new ROMs are just the Degital Press version of Video Life and some NTSC/PAL variations of already supported games.

     

    The problem with Desert Falcon seems to have somthing to do with the automatic PAL detection. It seems that the game does some frames longer than normal, which makes the emulator believe that it is a PAL game. Untill we fix this problem, you can enforce the right TV type with the -c0 command line switch. When you use it, z26 will not try to autodetect the game.

     

    As for the other requested features, I'm not sure we'll try this anytime soon. We are currently still doing too many internal changes between the versions, so a state save feature might suffer from incompatible file formats. And the chating feature would probably require a lot of new code for the user inputs, and I'm not sure if John or I have enough time for that right now.

     

     

    Ciao, Eckhard Stolberg


  14. OK, I guess you should have a look at Nick Bensema's site and download his "How to draw a Playfield" example. It is very well commented and demonstrates nicely how to get the TIA to display anything and still do some game game calculations. That would already answer a lot of your questions in part one.

     

    Also you should join the Stella mailing list, as this is where most of the VCS homebrew programmers hang out. You would probably get much better responses there than here.

     

    But now to answering your other questions:

     

    On 2: Yes you can make the moving objects as large as the screen. But you can also reposition them during the screen and get one sprite for top part of the screen and another sprite for the bottom part of the screen, that are essentially the same object. Kaboom uses that for the bombs for example.

     

    On 3: Yes, that is exactly how it is done. You can change any of the registers at anytime. That way you can change the shape or the colour of an object during the scanline.

     

    On 4: All registers are 8 bit wide. But some of them don't evaluate all of the bits. Writing to the unused bits has no affect.

     

    On 5: You can read the state of the joystick buttons directly from bit 7 of INPT4 and INPT5. However you can also enable the latching feature for these registers. If this is turned on and someone pushes the button and releases it, then INTP4/5 will report a pushed button as long as you don't disable the latching feature through the VBLANK register. Ususally it is not nessessary to use the latch feature, so you can just ignore this part of the manual.

     

    On 6: The only other devices that use the output on the joystick port are the Compumate Keyboard and the KidVid tape drive, AFAIK. But you probably won't be able to use either of them in your homebrew games, so no need to worry about that.

     

    On 7: There is no special manual on the driving controller. But this device just uses a quadrature encoder on the pins for left and right. On these pins you will see a signal like 00/01/11/10/00 etc. Depending on which way you turn the dial, it will cycle forward or backward through this list. A full turn of the dial has 16 state changes.

     

     

    Ciao, Eckhard Stolberg


  15. I'm not sure if anyone here feels like continueing with this type of discussion today, but if any of you prototype owners feel like sharing some screenshots of your rare gems with us, I'd like you to know that we have improved the screenshot support in z26 for you now. You can get the newest version of z26 here.

     

     

    Ciao, Eckhard Stolberg


  16. On the subject of performance issues:

     

    Browsing AtariAge during the time when America is still sleeping is nice and fast from over here in Germany, but once people on your continent start getting up, the reaction time of your site slows down consideranbly. Of course at this time it's evening over here when more people have time to surf the net, so I can't definately blame the speed reduction on your server, but on the other hand other American hosted sites still seem to work at normal speed. So maybe moving AtariAge helps to improve this situation.

     

     

    Ciao, Eckhard Stolberg


  17. On the 7800 board Chad Schell asked about a Cuttle Cart like development cartridge for the 7800. It would involve getting the hardware done to support all the different 7800 bankswitching formats. Also you would have to write the communication software for the 7800 and the PC. The 7800 side would probably have to be done in assembler. I'm not sure if that would be big enough a project for your thesis, or if you would consider doing something for the 7800 at all, but it might be worth some thoughts.

     

     

    Ciao, Eckhard Stolberg


  18. Video Chess does the colour flashing on z26 and on Stella. It doesn't do it on PCAE. I think this is because the game doesn't generate a sync signal while it's thinking, so PCAE doesn't know when to update the screen buffer.

     

    And I think Albert is right about why the flashing is there in the first place. The game uses all the cycles for the game calculations, so there is nothing left to generate the board display. To signal that the game hasn't crashed, they it just keeps changing the background colour.

     

     

    Ciao, Eckhard Stolberg

×
×
  • Create New...