All of these posts are real gems! I just had to comment on one of them:
The article "Replacing obsolete video game circuits with Xilinx CPLDs" is quite strange. The author uses a CPLD to replace a 74LS32, which contains four 2-input OR gates. This is an incredible amount of overkill given the capabilities of a CPLD, and is essentially pointless as the 74LS32 and other variants are still in production.
You'd think at the very least he'd find a truly custom or obsolete part with some moderate degree of complexity that would be better suited to a CPLD, or use a smaller scale device like a PLD. Since his focus is videogames, why not the NES PPU or a RIOT chip from a personal computer? Or a more obscure, out-of-production TTL chip?
But even that is not new or novel; arcade game manufacturers regularly replaced custom parts with daughterboards that implemented the same logic when they ran out of stock. And custom chips have been replaced with programmable logic (mainly EPLDs and FPGAs) in bootleg boards for decades.
So what I find really astonishing is his webpage:
He got a patent for this idea. A patent
. I don't even know where to begin -- I would assume there is a monumental amount of prior art for using any kind or programmable logic such as PROMs and EPROMs in this manner, and for this exact purpose. Most programmable logic vendors point out the fact that their product can be used to replace existing TTL parts, though perhaps with a focus on consolidating multiple parts into one chip.
Then there is the issue of practicality. How much more power is used for an entire CPLD and 3.3V LDO regulator compared to a single 74LS32? What is the cost of the parts and assembly of each 74LS32 replacement board? This is a very, very expensive solution to a problem.
I wonder if he deliberately picked something this trivial knowing it could get patented as existing patents would cover more complex (and actually useful) circuits.
Somebody should make a single non-inverting buffer replacement using an FPGA just to take this to even sillier extremes. The patent itself has a needlessly complex schematic repeated multiple times and a lot of filler material to take up more than one page. This is both hilarious and sad in terms of what ends up being accepted by the USPTO.
To be fair his article does cover more complex uses for this idea, but considering what he picked for a test case and that he sells
these parts to the public is just insane.