Like a splitter that makes two PDS slots from one? If the ethernet card uses DMA (I think many of the SE/30 ones do), I think there would need to be a bit of logic on the board to arbitrate between the two cards for memory accesses. If both cards assert /BR (bus request) at the same time, some logic on the PDS splitter must then ask the 68030 on the SE/30 to use the bus, and then grant the bus request to only one of the cards. A round-robin or more complex priority scheme can ensure that both cards have the opportunity to access main memory. Heh, a splitter could have some SRAM cache memory on it, too, so it would act like a cache card.A cheap SE/30 adapter for a Daystar card + connector for Ethernet would make my day.
What machine do you have? And what peripherals are you using?I need my NuBUS slots. I'm already short 1 slot. My vote is for PDS.
Not that I know of offhand, SE/30 cards have passthru connectors and stack vertically for the most part. PowerCache adapters for the SE/30 appear to me to place the accelerator in a horizontal position with the PDS passthru set up for vertical stacking as normal. The ARTMIX adapter places the accelerator connector vertically and has horizontal mounting provision for a PDS card as far as I can tell from pics online.(PDS splitters for SE/30 exist, right?)
The Rocket accelerates my video card. Also, it's just damn cool. Mine has 128MBs of RAM and I do launch RocketShare on occasion for giggles.If you have the proposed IIci version of this accelerator, you'll not necessarily require the Rocket. If the proposed IIci accelerator can be accessed under RocketShare, that would be a different kettle of fish and very cool indeed!
I've just redone the clock system, and fixed a problem with some of the level shifters. The way I was doing the /BG and /BGACK signals would have been way too slow for operation at 8 MHz. Maybe 2 or 5 MHz max. So I fixed it. While doing that, I managed to add the ability for the card to function as either a PDS master (accelerator) or slave (expansion card). And better, it can do both at once, in the sense that it can run the emulated 68000 while also letting the main Macintosh run too, making it similar to the Rocket in the sense that you can have a slow Mac and a really fast Mac going at once. Again, not something I want to implement myself, but a cool area to be developed in the future.If you have the proposed IIci version of this accelerator, you'll not necessarily require the Rocket. If the proposed IIci accelerator can be accessed under RocketShare, that would be a different kettle of fish and very cool indeed!
Ah, yes. Hmm. I'd like to add a passthrough connector for my eventual SE/30 card, but it may make the unit too costly. Paralleling the PDS slots is really frustrating. And anyway, I'd wanna add another set of 3 $2.50 FPGAs on the other side to talk to the card you attach. That would help with throughput. Adds a ton to the cost though. Just fantasies at this point. Let me focus on SE and Plus. I know exactly how I'm gonna do the 68000 emulation. 68030 is not as clear, and, especially when MMU emulation is enabled, will be slower than 68000 emulation.Not that I know of offhand, SE/30 cards have passthru connectors and stack vertically for the most part. PowerCache adapters for the SE/30 appear to me to place the accelerator in a horizontal position with the PDS passthru set up for vertical stacking as normal. The ARTMIX adapter places the accelerator connector vertically and has horizontal mounting provision for a PDS card as far as I can tell from pics online.
What you're thinking of would likely be the PDS splitters for the Macintosh IIsi.
I love the early Radius stuff. One question, though. What is the separate ROM board for?edit: dunno if it might be helpful, but you can see traces on the Radius 16 Accelerator for the SE in the HiRes pics posted in applefritter's NuBus Mafia gallery. Info on slot decoding/interfacing for your card might be apparent.
Amazing. I have been wanting one since I was 12 hahah. Do you have any technical description of the Rocket? I would be dying to read it and make sure Rocket acceleration is possible to implement with my design.The Rocket accelerates my video card. Also, it's just damn cool. Mine has 128MBs of RAM and I do launch RocketShare on occasion for giggles.
That the pic was of the original Mac Plus version of the card was my first guess. Mine was the Radius 16 for the SE as in the NuBus Mafia's HiRes pics, no additional ROM necessary. WAG would be that the ROM board bypasses (disables with an INIT?) the 68000 on the Mobo, its 68000 booting from the expanded (?) ROM on board and then the Accelerator's 68020 bypasses that.So there are two 68000s in this system and one 68020? Trash, do you know anything about this?
Edit: that's different from the board you linked though. That one seems to be installed in a Plus.
Not so much, I'm a visual thinker and the only time I did a layout for over 50 components was the first and they were all DIP CMOS at about the time Don Lancaster wrote that designing a board without a MicroProcessor on board was becoming counterproductive. My brain doesn't do schematics or much electronics for that matter. So board design to me is an exercise in electron plumbing and a visual schematic development process.That's crazy! I've never heard of doing them in Illustrator..
About the power budget for the Portable's PDS: ISTR reading a spec that an external power source was required. Just checked DCaDftMF2e and there's no mention of it there. Hrmmm? Have you considered decoupling the Power pins and jumpering from 5V on one of the peripheral interconnect cables. A M-F adapter for the cable ought to be easy enough to knock out.The portable accelerator is hard in terms of getting it within the power budget. Maybe an ARM Cortex-M7 running an interpreter, with a bit of SDRAM and an FPGA accessible through the MCU's external bus interface would provide a small boost to a Portable. That would probably still exceed the power budget for the Portable's PDS, but only barely