Jump to content

Bruce Tomlin

  • Content Count

  • Joined

  • Last visited

Community Reputation

117 Excellent

1 Follower

About Bruce Tomlin

  • Rank
    River Patroller

Contact / Social Media

Profile Information

  • Custom Status
    CD C9 01
  • Gender
  • Location
    San Antonio, TX

Recent Profile Visitors

25,038 profile views
  1. This. A long, long time ago (2005-2007) I tried to see just how much I could optimize out of the ROM, so I checked the disassembly of every game I could find to see what parts of the ROM were called. I made a list of those addresses, and marked those in a certain way in my .asm of the ROM, so that I could see in the symbol table dump if they changed. I actually got over 1K of contiguous free space, you could put a small game in there. I may have even made a version with a controller test built in before I got tired of doing Colecovision stuff. So basically the "skip startup screen BIOS" is just someone's sloppy hack that stepped on the wrong bit of code that broke the even sloppier code in a single released game. Here are the notes I put in the source code for my optimized ROM: ; LOCKED ADDRESSES -- THESE *MUST* BE CORRECT OR SOME CARTRIDGES WILL NOT RUN! ; ; In order to allow easy verification, these addresses have labels ; starting with an 'A' and containing the locked address. ; ; These are the important addresses (like vectors), plus a few used by the ; small-time offenders. ; ; 0000 0066 080B 18A3 ; 0008 0069 08C0 18D4. ; 0010 006A 143B 18F7* ; 0018 006C 144D 1968 ; 0020 01B1* 1463 196B ; 0028 01D5 14B4 1CCA. ; 0030 023B. 14C1 1D43 ; 0038 02EE 14C3 1D47 ; 003B. 07E8* 15A3 1F61 ; 158B* ; ; ...and these are the addresses used by three of the four Interphase ; games, most of which have vectors available. ; ; 0213. 114A*. 1BAA* 1D3E*. ; 025E. 18E9. 1C27. 1D57. ; 027F. 1979*. 1C4F* 1D5A. ; 0300. 1B1D*. 1C82*. 1D60*. ; 1105* 1BA3. 1D01*. 1D66*. ; ; These two were called from Boulder Dash, but the call instructions ; themselves may be unreachable from the rest of the game. ; ; 01C1* 026A* ; ; Any address with a '*' is only referenced by ONE game that I can find. ; Note that Interphase games account for 11 out of 15 of these. Any address ; with a '.' has a vector available. Note also that all but two Interphase- ; only addresses (18 of 20) have vectors! Oh the humanity! ; ; These are in the "OS7SYM" file which was provided to Adam programmers: ; ;FREQ_SWEEP EQU 000FCH ! ;ATN_SWEEP EQU 0012FH ! ;DECLSN EQU 00190H ! ;DECMSN EQU 0019BH ! ;MSNTOLSN EQU 001A6H ! Also, f*** Interphase. coleco29.asm ctltest.asm coleco.h
  2. I wrote my own a long time ago because I was unimpressed by all the binary-only DOS-only releases. It's not the same syntax as CUPL, but it's more than good enough for getting a chip working. I have some example files in the test sub-directory which I had used for 7800 and Colecovision bank switching. http://xi6.com/files/palwiz-1.2.2.zip
  3. Here's some more info on 315-5342: https://www.smspower.org/forums/9277-AddingFMSoundToMasterSystemModeOnAGenMD https://web.archive.org/web/20190403192608/http://www.smspower.org/forums/14084-PowerBaseConverterInfo It seems to be a PLA chip that uses "one weird trick" to become an 8-byte ROM that vanishes after the first memory write. It forces opcodes that set the stack pointer to DFFF, then RST 00H onto the bus at startup for the few games that don't initialize the SP properly and expect a BIOS (which the Genesis doesn't have for Z80) to have already done it. So perhaps this was one of the few games that needed it. If this game uses a mapper, it would actually be built into the ROM chip. So it's back to all you need is dealing with the SMS pin remapping, and grounding S_RES (aka M3).
  4. I printed out both pictures (with the bottom flipped) and started marking down various signals. I quickly realized that top chip was NOT a 74LS153. (first of all it has 20 pins rather than 16) In fact it is a 315-5342, which is the main chip from the Power Base Converter. Or maybe it's a SMS mapper chip. So in a way, it's cheating. Anyhow, here is a link showing some of the wiring for a PBC: https://pyra-handheld.com/boards/threads/making-a-power-base-converter.53277/
  5. I had both a Studio II and a 4-port 5200 and I thought that was pretty clever. Of course that was long before VCRs made composite jacks a thing. The 5200 controllers were so bad that even now I can deal with the power switch box, or even with taking an afternoon to mod it for composite and power jacks. (I should do that someday with the rare 4-port I found that works with the 2600 adapter, probably was a service upgrade motherboard.)
  6. Ugh, what a mess. But the top half of the legs are still there, so it can be fixed. The best pin donors are old socketed PAL chips, since you can't reprogram them. Desoldering a chip is not a simple skill. After at least 30 years of it, I'm still coming up with new tricks. Basically it's about melting ALL the solder to the other side, and sucking it all out at once. The most important thing is making sure the pin is free (with those red squeeze bulb types I gently wiggle the pin around until it can move) and adding fresh solder if it doesn't suck out properly. Ground pins are the worst because they absorb so much heat and tend to have more solder on the other side. The other trick that I recently learned is to put some extra flux on, to help it melt better. (that's probably why the fresh solder helped) And it'll be funny to see a half-broken Pokey chip being used just for the audio part.
  7. Main problems I had were an apparently slightly flaky power connector (it could have been the wire, sticking up vertically doesn't help, but it only has to disconnect just once!), and problems running from battery. It really doesn't work with NiCd/NiMH batteries because of the lower voltage, and they last almost no time. (At least it's not using a 7805, or it would be even hungrier!) That's probably why the Game Gear battery pack won't work. It really needs a 7-cell battery pack for NiMH. A lithium pack ought to be great. I suppose replacing the voltage regulator with something more efficient would be good too. And it's just so stupid chunky, like an Xbox Duke controller. I know it had to be that big back in the day, but it's still on the upper limit of acceptable.
  8. I catweaseled all my disks back in '08 or so, though I wrote my own code to decode Apple II and Commodore disks. My TRS-80 disks even had the same bad sectors where they were back in the day. The important thing is to know what format you plan to rip to. FM/MFM for TRS-80 was well supported. I never quite got what the preferred formats were for the other systems, but I never had those systems back in the day, most of my Commodore disks were a few hundred I got one day in a Sam's Club parking lot when some boomer had expected to meet someone there to give them away, and the guy didn't show up. Writing code to decode GCR taught me that Apple's GCR was better than Commodore, simply because the Apple GCR had unique extra codes that could be used as proper address/data marks. Commodore's was only straight nibbles, and you had to guess where sectors began. I also learned that for some reason Atari (8-bit) used the negative-logic version of its FDC, so all bytes are inverted, and it also ran the motor at a slower speed to get an extra sector. I even ripped a few old 5 1/4" CP/M disks for Osborne and Kaypro. It's still a "someday" project to rig up an 8" floppy drive to read the box of those that I have. (Some are RSTS-11 disks, it's amazing what you can find at thrift stores sometimes!)
  9. Also the back panel is going to be different with pre-TMSS units. If you don't see a modem port, don't even bother, the chance of it being pre-TMSS is probably zero. The modem port was completely removed after TMSS. I have two of them right now that I just looked at and tested, HRG lid with modem ports, both have an FCC ID of FJ8USASEGA. One is broken (looks like slot damage), and the other has TMSS. The broken one is missing its screws (my fault for sure), and there are chip dates of 9143 and 9149 inside. According to Wikipedia, the US Genesis was introduced in 1989, and TMSS happened in 1990, so they were making modem port units for at least a year after TMSS. They may be rare as hell now, but they were around long enough for EA to make half a dozen games that were broken by TMSS, while apparently Sega had already written their own game code in anticipation of it. As for "very late" units, somewhere I have one that had a modem port motherboard and shell, but no DB-9 plug. I had to add one myself (to use as a serial debug console), so I used a male plug because that's what I had handy. Gonna guess it was from early 1992, gotta take a look if I find it. I thought I had just found it, but I was wrong. Anyhow, that one had TMSS too. And there's always the chance that despite the FCC ID, the original motherboard died and was replaced (either as official repair or more recent history) with a TMSS motherboard. Still, it's good to know about the FCC ID thing.
  10. Still staying home and digging into the attic now that the Texas weather has finally stopped being hot. But I finally found my Gamecube (and the missing Zelda TP disc that I knew was in it), my fat PS2, and my "dev" Sega Genesis (with a DB-9 installed in the "modem" port for serial debugging at a blazing 4800 baud!). Also a big box full of Sega stuff including a floppy disc game copier. (it supports SNES but didn't have the slot adapter)
  11. If I recall correctly, IBM made some mistake when implementing their first Centronics interface, so there is a slight difference with the "original" spec, which became the only way it was done afterward. So watch out for that.
  12. 8VAC 1.5A, gonna guess you can use 9VDC of the same amps or higher if it has the same size plug.
  13. The MC-10 was Tandy's ZX-81. Except even Tandy couldn't bring themselves to make something crappy enough to compete with Sinclair.
  14. SSN might imply a military family. Back in the 70s a lot of my family's stuff had my dad's SSN on it.
  15. Don't forget the clock cap in the original Xbox! Most of them are PowerStor caps (black wrapper) and they will slowly leak and mess up your board. (There are also some gold ones which are Nichicon and supposed to be okay.) If you find one and don't have time to take the board out, wiggle it until its legs break and fix it later. Better to have to set the clock than to let the board get any worse. One of my two "good" units had a Nichicon and the other had a PowerStor. As I was removing the PowerStor, an SOT-23 diode fell off the board, probably related to the charge circuit, but I was able to save it, get the board pads re-tinned (not easy!), and put it back. Then I put a header in the wider pair of holes (0.2" apart) and connected it by a wire to a larger PowerStor that was scrap from a place I used to work. (We had a few of those leak too!) Then I then snugged it into a place away from the board. But yeah, the biggest problem with electrolytic capacitors is from the early '00s, when China stole a formula for a "better" electrolyte, but they didn't get it right.
  • Create New...