Subject: Cinematronics Masked ROM part numbers From: Zonn Moore Date: Tue Jun 10 1997 - 13:28:00 CDT ------------------------------------ Hi guys! I've been on vacation for the last 10 days roaming in and around the canyon lands in Utah. Great country! So sorry I didn't get back with you Al! I've seen a couple of different boards use the "solder wire to the other socket" approach. There mostly on the older games (pre- Star Castle), and seem to have been done that way simply because the board contained no sockets in the place where the mask parts should have been placed. The memory mapping of the C-CPU is a little strange. The 2nd and 4th sockets from left are the low addresses when using normal EPROM / PROM devices. The memory map then continues to the 1st and 3rd sockets. When Vectorbeam re-merged with Cinematronics, Cinematronics got a "shitload" of old Vectorbeam boards -- according to a technician that worked there at the time. I'm guessing that the Vectorbeam boards where all Space War or Speedfreak boards and were wave soldered with only the 1st and 3rd sockets intact. When you have a PROM masked for you, you can choose to have the select lines inverted from the normal EPROM usage, and this is what vectorbeam did. Therefore since the only difference between the 1st and 3rd sockets and the 2nd and 4th sockets are the inverted select lines, Vectorbeam masked parts plug into the 1st and 3rd socket even though they are addressed lower in memory. Most of the Cinematronics masked parts seemed to have been masked using a normal EPROM select line, and plugged into sockets 2 and 4. So Cinematronics dilema was that they "inherited" a bunch of Vectorbeam boards with sockets only in the 1st and 3rd position, yet their mask parts were designed for the 2nd and 4th position. The choice is to solderwick all the plated through holes to allow the addition of a socket (a pain), or simply to use the supplied sockets, bending up the selection pin on their masked parts and solder a jumper to the proper pin on the adjacent socket (much easier). The Sundance mask ROMs were designed to plug into the 1st and 3rd sockets and were 4k parts that were setup to be plugged into a standard 2k socket CPU card which you can do by rearranging a couple of Address/Selection lines and inverting their logic levels -- something you can do with mask parts. That's why the Sundance PROMs required a special convertor to read them. The boards I've seen to date that used masked parts are: Space War Tailgunner II Sundance Starhawk Boards I've heard of using them, but haven't on the ones I've seen: Tailgunner Space Wars Warrior I don't have the numbers for any of these here at work, and won't be able to get them until this weekend. So none of this answer the question of whether your board is a Vectorbeam Scramble. If the board says "Vectorbeam" and the socketed parts are in the 1st and 3rd locations chances are *real* good it's a Space War though there's a slim chance it's a Speedfreak. If the parts are soldered in you can be 99.9% sure it's a Space War. Have you plugged it into a Star Castle or something (do you have a Star Castle or something to plug it into?) Since they plug into the 1st and 3rd sockets with jumpers to the 2nd and 4th they should look like normal parts (if they're 2k parts), you could try reading them and running them with CINEMU. If there 4k parts they're sometimes a little harder to read, though you should be able to read 2k of them and compare them to what's on the net. -Zonn -------------------------------------------------------------------------------- Subject: Cinematronics Masked ROM part numbers From: Gregg Woodcock (woodcock@nortel.ca) Date: Tue Jun 10 1997 - 12:58:00 CDT Some game boards use 2 PROMs instead of the normal 4 PROM setup. These 4K masked PROMs were mistakenly manufatured with the 2K blocks reversed (either that or they put the ROM sockets in the wrong pair of spots)!!! To fix this, they lifted the 2K address leg of each PROM so that it didn't go in the socket and jumpered it to the ROM spot to the left which uses an inverted signal for that line. This jumper easily comes loose so the best solution is to pull the PROMs, read them in, swap the 2K blocks and burn them into a new EPROM and do away with the address hack altogether.