Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!ucsd!swrinde!zaphod.mps.ohio-state.edu!sdd.hp.com!caen!math.lsa.umich.edu!math.lsa.umich.edu!hyc From: hyc@math.lsa.umich.edu (Howard Chu) Newsgroups: comp.sys.atari.st Subject: Re: Any hardware hacks anyone ? Keywords: hardware hacks? Message-ID: <1990Oct6.062139.24891@math.lsa.umich.edu> Date: 6 Oct 90 06:21:39 GMT References: <1425@abekrd.UUCP> Sender: usenet@math.lsa.umich.edu Distribution: comp Organization: University of Michigan Math Dept., Ann Arbor Lines: 69 In article <1425@abekrd.UUCP> mike@abekrd.UUCP (Mike Quinn) writes: >I have been following the discussions on comp.sys.amiga (just out of interest) >and they're talking about doubling the processor clock speed. Now I know >about hacks to upgrade the memory (done that); add my own second internal >drive (done that); make the display bigger (thinking about it :-))... A second internal drive, huh? Kind of an interesting idea. Don't think I'd care too much one way or the other. (Especially since the Mega's power supply is already straining just to keep the fan and floppy running...) >Has anybody done any interesting (and fairly simple) hardware projects to >boost this wonderful machine in other ways? Can _we_ double the ST clock to >16 Mhz ? How about a homebrew remote keyboard ? Or details on putting all >these disks & ST motherboard in a small(ish) PC type box ? FAST Technologies' Turbo-16 has been around for a long time now. The installation process is tedious but straightforward. ICD also has a new accelerator, which is said to be even easier to install, also uses a cache, and is more compatible. (I.e., Turbo-16 doesn't let you switch back to true 8 MHz operation, the ICD accelerator is supposed to allow this. This is crucial for time-sensitive code like Spectrum 512 uses, or copy-protection code for many games.) You can double the clock going to the floppy controller, and drive high density (1.2 and 1.44 meg) floppy drives. I think Best Electronics has a kit for making the standard ST keyboard detached. There are a couple companies advertising PC-style cases for ST systems as well. Just find yourself a copy of AIM or some other good Atari magazine and look thru the ads... >Come on guys, lets make this machine _work_! Either E-mail me or post details >to comp.sys.atari.st.tech. I'll summarize any mail there too. > > Mike. > >PS. The second drive mod involved replacing the internal drive with two thin > 3.5" drives, a new ribbon cable, adding one wire to the motherboard and > a little bit of sawing on the chassis (Not recommended if your machine is > still under warranty :-)). For a *really* intense project... Reverse engineer the Atari custom chips, then rebuild for faster operation. (Of course, ya gotta figure, if Atari couldn't do it themselves, having the original chip designs onhand, it's probably not something you wanna try on your own....) What we need at this point are ST clones. That's what drove the prices down and the performance up in the PC market.... Systems with compatible hardware but offering higher throughput, clock rates, etc... Lessee... How 'bout a dual-ported RAM system, and a faster memory control unit, or heck, a PMMU. I don't think there's too much of the OS that's dependent on the way the GLUE/MCU/DMA stuff are split out. Just keep the DMA and MCU controller registers in the same place.... The blitter would probably pose some problems though. (If you wanted to duplicate it. Chances are, you won't care.) So what do we need... 68030, fast memory, better video hardware [JRI or ISAC can provide that, right?], the WD floppy controller, the Yamaha sound chip (or the improved model, if you can find it), the 68901 MFP, a pair of 6850s, hm... that about covers it. Dunno how you could build such a system for cheaper than Atari can, tho, considering that they already have their mass-production facilities for economy of scale, and based out of Taiwan, so they already have the cheap labor. Hm. Mebbe not such a good idea after all. Who knows. -- -- Howard Chu @ University of Michigan one million data bits stored on a chip, one million bits per chip if one of those data bits happens to flip, one million data bits stored on the chip...