Hi all, So here's the state of play in designing new QL replacements...
Peter Graf is bogged down with some issues on his board. It has a decent spec but it doesn't currently "work" in a meaningful sense. He has various obstacles (working alone) that mean his design will be delayed or not released. So, for that reason, I'm looking at the resources we have available, the skills people on the list have, and seeing if we can have a serious, focused, yet public discussion about an alternate project. I have no desire to "compete" with Peter Graf. However, it is always prudent to have a back-up plan that doesn't rely on someone who has been working on his project for several years with no new product to show... I shall outline basic choices which will go far to define the machine's spec, capabilities and expandability: CPU: The choice here is between three things: A genuine 680X0: + compatibility, already designed - availability and price An emulated 680X0 in FPGA form: + highly configurable, offers single chip solution, 68k VHDL sources in public domain, very low cost - heavier design load, harder to debug. Software emulation: + Already here/done, highly reconfigurable, no hardware skills needed - Intel hardware too big for just a QL - needs to do other things too - expensive. Specification: What clock speed is needed to be useful? 25 MHz? 40 MHz? What screen resolution? It's generally agreed that basic mode 4/8 is insufficient so GD2 support seems to be a basic requirement. VGA seems to be a minimum and DVI seems desirable. What storage? We only need a gig or two - it seems SDHC is a future-proof, low cost, low power, low design-cost standard many could get behind. I had a fork in the road moment when discussing emulation on ARM embedded boards vs. emulating the 68k in an FPGA. Right now, honestly, the FPGA wins and will continue to do so for a few more years (unless the perfect ARM board can be bought off the shelf)... I'd like to support Peter in his efforts, but make sure that there's an alternative in place should he trip... Besides people I am already aware of, this seems to be something constrained by our skills... There's a complete functioning 68000 FPGA core open-licensed here: http://opencores.org/project,ao68000 which is used for Amigas and Ataris so seems eminently suitable... Knowing someone who can do further FPGA design to add the video/etc we need is a crunch point. (This is basically what Peter's doing, except I understand he designed his 68000 from scratch and that's where he's stuck) There seem to be a lot of people willing to work to get SMSQ/E running on anything if it's compatible enough - this is great and it would be smart to include SMSQ/E on any platform because it's the de facto standard. Minerva is in second place, and original QDOS isn't really in the running... If someone can make an FPGA happen, I can make PCBs happen, then someone can make the OS happen, then we have a computer. Of course, you could always decide that QPCII or Q-emuLator would replace that - they will always be faster :) However, this small project may have wider applications too, and it would be nice to see something happen. Dave _______________________________________________ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
