7 ms·

What I'm hoping for, and this extends to all the reproduction-PCB projects (aside from the Mac ones, I've seen a replacement Atari 800XL and some Amiga designs)

by hakfoo 6y ago

What I'm hoping for, and this extends to all the reproduction-PCB projects (aside from the Mac ones, I've seen a replacement Atari 800XL and some Amiga designs) is that we can come up with a way to reproduce the ASICs rather than scavenging them from dead units. For some machines, there may be an ample supply to scavenge (old Macs being famous for barfing their batteries in a way that destroys the PCB but leaves the chips intact), but some other designs have far more limited supplies of spares. Plus there's the attrition risk-- something is gonna get damaged while parting out some percentage of the time. The introduction of replacement ASICs would allow us to grow the total population of a rare platform. I could imagine offering a turnkey kit-- a PCB plus every part necessary to populate it and have a working vintage Mac/Amiga/Atari/etc, with fresh capacitors and enthusiast quality sockets and PCB materials. They were able to offer such a kit for the IBM PC, but that was famous for not using custom components.

monoideism 6y ago
Really, the best way to solve this is likely with FPGAs, not ASICs. If you're not using an original chip, does it really matter if it's an FPGA or not? Already, most of these reproduction PCB projects, including this one, end up using PLDs to replace hard-to-find specialied ICs (and even those PLDs are at risk of EOL). So it's not like they're using 100% original chips, anyway. I understand the urge, but unless the cost of ASIC manufacturing drops dramatically, I don't see this as being practical for small runs like hobby reproduction boards. Edit: I forgot to add: this is impressive work, nonetheless!
- zeckalpha 6y ago
  Further in the article, there’s a discussion of replacing custom PAL logic with modern Atmel chips.
  - monoideism 6y ago
    I mentioned that ("including this one"). Problem is, they're Atmel chips that are still in production, but I wouldn't call them "modern". Most competing PLDs are now out of production, and it wouldn't surprise me to see the atmel f16v8 series follow. More and more people just use cheap CPLDs or FPGAs (which unfortunately aren't available in DIP packages and so not pin compatible with the old custom ICs used in the 80s).
    - throwaway2048 6y ago
      its pretty easy to make SMT to DIP adapter PCBs, doesn't work with large chips, but a small FPGA is all you need for such old ASICs anyways. http://www.the8bitguy.com/wp-content/uploads/2018/05/SwinSID-Ultimate02.jpg http://www.the8bitguy.com/wp-content/uploads/2018/05/SwinSID...
- hakfoo 6y ago
  I suppose it depends on how you're using FPGAs. It seems like a lot of the FPGA products-- things like MISTer-- are system-level emulators, which rarely provide full legacy I/O breakout. You can emulate a complete 68k Mac on a FPGA, but where do you attach that weird NuBus card? Now, replacing individual chips with FPGAs/PLDs is likely to retain that, although I could imagine it's a lot less forgiving of an emulation, when it has to match the analog-domain properties of the original chip somewhat.
  - monoideism 6y ago
    > You can emulate a complete 68k Mac on a FPGA, but where do you attach that weird NuBus card? I just looked, and NuBus has 96 pins with clock rate of like 20 MHz. There are FPGAs with > 500 GPIO that can easily handle that clock rate, and ones with over >200 GPIO that are fairly affordable. Now, I mess around with FPGAs, but I'm not an electrical engineer, and there may be other issues. But it seems doable. Any electrical engineer and/or hardcore FPGA hacker want to chime in?
    - NovemberWhiskey 6y ago
      It also looks like at least 30-40 of the pins on NuBus are either grounds or power supplies, so it's not even as bad it seems.
      - monocasa 6y ago
        Yeah, those wide parallel connections were pretty much universally power/ground heavy for signal integrity reasons.
tyingq 6y ago
If you read the comments on the story, the author did appear to have reverse engineered several of the Apple branded PAL chips on the board. So he was able to program new ATMEL PAL16R8Bs as replacements.
grishka 6y ago
> They were able to offer such a kit for the IBM PC, but that was famous for not using custom components. Are 8086 CPUs still manufactured?
- alrs 6y ago
  8088
mrweasel 6y ago
Some people prefer to tinker with hardware, and find designing a new PCB easier than attempting to implement an entire machine in an FPGA or ASIC. Some of these older systems, like the Mac SE/30, the Amiga, Ataris and C64 are simple enough that you can successfully debug and fix many issues with little to no background in electronics, and that provides a great way for people to learn. Most enthusiasts will most likely use both emulator and real hardware. There is some sense of enjoying and sense of nostalgia coming from using real hardware. What I would like to see is more work trying to recreate the chips required to repair or even recreate these old machines. If we just want to be able to access software and data from old systems, FPGA or emulators are fine, but if we also want to retain the experience, I think real hardware provides an added layer.
stuaxo 6y ago
Its the logical next step. We may have to wait a while. Maybe there will be some standard small FPGA that can sit in various packages.
- Narishma 6y ago
  At that point, wouldn't it make more sense to just emulate the whole thing in the FPGA, instead of individual chips?

by hakfoo 6y ago

monoideism 6y ago

Really, the best way to solve this is likely with FPGAs, not ASICs. If you're not using an original chip, does it really matter if it's an FPGA or not? Already, most of these reproduction PCB projects, including this one, end up using PLDs to replace hard-to-find specialied ICs (and even those PLDs are at risk of EOL). So it's not like they're using 100% original chips, anyway. I understand the urge, but unless the cost of ASIC manufacturing drops dramatically, I don't see this as being practical for small runs like hobby reproduction boards. Edit: I forgot to add: this is impressive work, nonetheless!

zeckalpha 6y ago
Further in the article, there’s a discussion of replacing custom PAL logic with modern Atmel chips.
- monoideism 6y ago
  I mentioned that ("including this one"). Problem is, they're Atmel chips that are still in production, but I wouldn't call them "modern". Most competing PLDs are now out of production, and it wouldn't surprise me to see the atmel f16v8 series follow. More and more people just use cheap CPLDs or FPGAs (which unfortunately aren't available in DIP packages and so not pin compatible with the old custom ICs used in the 80s).
  - throwaway2048 6y ago
    its pretty easy to make SMT to DIP adapter PCBs, doesn't work with large chips, but a small FPGA is all you need for such old ASICs anyways. http://www.the8bitguy.com/wp-content/uploads/2018/05/SwinSID-Ultimate02.jpg http://www.the8bitguy.com/wp-content/uploads/2018/05/SwinSID...
hakfoo 6y ago
I suppose it depends on how you're using FPGAs. It seems like a lot of the FPGA products-- things like MISTer-- are system-level emulators, which rarely provide full legacy I/O breakout. You can emulate a complete 68k Mac on a FPGA, but where do you attach that weird NuBus card? Now, replacing individual chips with FPGAs/PLDs is likely to retain that, although I could imagine it's a lot less forgiving of an emulation, when it has to match the analog-domain properties of the original chip somewhat.
- monoideism 6y ago
  > You can emulate a complete 68k Mac on a FPGA, but where do you attach that weird NuBus card? I just looked, and NuBus has 96 pins with clock rate of like 20 MHz. There are FPGAs with > 500 GPIO that can easily handle that clock rate, and ones with over >200 GPIO that are fairly affordable. Now, I mess around with FPGAs, but I'm not an electrical engineer, and there may be other issues. But it seems doable. Any electrical engineer and/or hardcore FPGA hacker want to chime in?
  - NovemberWhiskey 6y ago
    It also looks like at least 30-40 of the pins on NuBus are either grounds or power supplies, so it's not even as bad it seems.
    - monocasa 6y ago
      Yeah, those wide parallel connections were pretty much universally power/ground heavy for signal integrity reasons.

tyingq 6y ago

If you read the comments on the story, the author did appear to have reverse engineered several of the Apple branded PAL chips on the board. So he was able to program new ATMEL PAL16R8Bs as replacements.

grishka 6y ago

> They were able to offer such a kit for the IBM PC, but that was famous for not using custom components. Are 8086 CPUs still manufactured?

alrs 6y ago
8088

mrweasel 6y ago

Some people prefer to tinker with hardware, and find designing a new PCB easier than attempting to implement an entire machine in an FPGA or ASIC. Some of these older systems, like the Mac SE/30, the Amiga, Ataris and C64 are simple enough that you can successfully debug and fix many issues with little to no background in electronics, and that provides a great way for people to learn. Most enthusiasts will most likely use both emulator and real hardware. There is some sense of enjoying and sense of nostalgia coming from using real hardware. What I would like to see is more work trying to recreate the chips required to repair or even recreate these old machines. If we just want to be able to access software and data from old systems, FPGA or emulators are fine, but if we also want to retain the experience, I think real hardware provides an added layer.

stuaxo 6y ago

Its the logical next step. We may have to wait a while. Maybe there will be some standard small FPGA that can sit in various packages.

Narishma 6y ago
At that point, wouldn't it make more sense to just emulate the whole thing in the FPGA, instead of individual chips?