What do I really want?

[NickMilner]

My /2c FWIW.

FPGA's are a game changer for so many applications (the Oxford Oscillators in my Novation Peak synth are powered by an FPGA, for example) and if FPGA tech was available back in the day it would have been used everywhere. Personally, I don't see this as a problem.

Also, I can't say I'm a big fan of "I'm out" threads in forums. People lose interest in things all the time and don't feel the need to tell everyone. That said, I'm just a lurker on these forums so take these words with a pinch of salt. I get the feeling you're struggling a bit and I wish you all the best.

[kelli217]

if FPGA tech was available back in the day it would have been used everywhere.

It was, and it was. The PLA in the C64 was an example of an early predecessor to FPGA technology. PLA, after all, stands for Programmable Logic Array. The Sinclair ZX81 and ZX Spectrum used the Uncommitted Logic Array, or ULA, fabbed by Ferranti.

Almost all the home and personal computer manufacturers back in the day, when they developed their various 'cost-reduced' or 'reduced chip count' machines were using some flavor of this kind of technology, where they could buy massive quantities of off-the-shelf chips and just program them to function as a replacement for a dozen or more TTL chips.

[Daedalus]

I'm pretty sure those devices weren't field programmable, though. You had to mask in your end design during the fab process. So you had to pay tooling charges and order a bunch of them.

You could also get your microcontrollers programmed like that too, like the ones used in keyboards and the like.

edited to add: I remember when Motorola came out with the 68hc11 in 1984 or 5... it was a microcontroller with a UV erasable ROM built right in. It was awesome. You could reprogram it over and over and it was still a one chip microcontroller, all it needed was an oscillator and almost all of it's pins were I/O lines. I used those things for all kinds of stuff.

edited to edit the edits:
Wait... That wasn't the part number, and the year is a few off. But it was a 28 pin DIP with an EPROM window, and cost close to 50 bucks. Motorola made a little "dev board" that you could build and it allowed you to program it with a standard EPROM. So you had to use their assembler to write your program, write it to an EPROM, then use their little dev board to program the EPROM in the microcontroller.

But even that is absolutely destroyed by FPGAs now.

[DragWx]

My hot take is, most complaints about FPGAs come from not understanding what they are and why we might use one, especially when the source code for the FPGA's configuration is open and available. It's just static circuit design, and there's limits to how much stuff can be crammed in, just like how regular ICs also have limits to how they can be designed.

They're the closest a hobby community can get to making custom chips for their projects, not unlike Commodore/MOS Technologies with the VIC, VIC-II, SID, TED, and CIA chips, or Atari with the TIA, POKEY, ANTIC, Greg, John, Stephanie, Bertha, etc, of which nobody seems to have complaints about, other than "these sure are hard to get, huh?".

For the X16, listening to just criticism and nothing else, you'd get an impression as though the designers were sticking some kind of fully-fledged minicomputer (like a Raspberry Pi) onto the board and calling it "the video chip", which is a fundamental misunderstanding of what FPGAs are, and that criticism isn't unique to the X16.



Note: There's nothing wrong with programming a Raspberry Pi to work like a video chip and dropping it into a project, but it's not equivalent to what FPGAs are.

[Ed Minchau]

Back in the day I used EEPROMs to function like an FPGA. Even a 256 byte EEPROM could replace a lot of TTL circuitry. You just needed to wire it up differently than other memory, with control lines replacing address lines, chip enable, and memory read.

[DragWx]

Back in the day I used EEPROMs to function like an FPGA. Even a 256 byte EEPROM could replace a lot of TTL circuitry. You just needed to wire it up differently than other memory, with control lines replacing address lines, chip enable, and memory read.

The wisdom I keep hearing about that is, ROMs are often too slow to act as chip select decoders and the like, which is why PLAs and ULAs were used instead. I'm surprised to hear it actually worked for somebody.

[BruceRMcF]

Yes,they are fast enough if they are fast enough ... one reason PROMs were sometimes used rather than EPROMs for this, even though they couldn't be reprogrammed, was that PROMs were available that were faster.

If you have an EPROM that is just fast enough as memory for 8MHz, it won't be a good chip select for an 8MHz clock system, but it might be just fine for a 2MHz system ... especially since it can model arbitrary logic, so it will normally be only the single level of delay.

[GTR3QQ]

why not try out the emulator? see if this is for you.
From what I gather, I'd think is maybe not the hardware that interests you.
But the urge to create something, a space perhaps just for you.
Isn't that you really wanted?
I can not tell you what you should want, none of us can, even if you've been "hinted" this is for you by all the medias, deep down, I can see you still very clearly label this idea as not what I really want.

To that I'd, the only way to find out, is by trying it out yourself.
If you think emulation isn't a good enough, once the thousand is sold, I believe some will offer lending services or you can return for refund.

I hope you this useful.