Yes, the subject was FPGA-based implementation of MCU cores, but I was caught up lamenting that, while programmable logic has made standard "family" logic, passe', it has not replaced it.

The MOSFET issue was relating to the means for driving the solenoids. The 3-volt logic outputs of which most programmable logic of current generation is capable, are not suited to driving low-cost MOSFETs, with which we accomplish useful work. If I had to replace the board-full of 5-volt logic that drives, say, the hammers on a line printer (yes, an old and uninteresting problem, but one which addresses my issue) I'd have to provide buffers capable of driving the MOSFETs. Those hammers, are, after all, solenoids, and require a considerable "whack." Recently, I've learned of some low-voltage compatible MOSFETs, but they're not as cheap as the old types that would run from 10-12 volts. However, there were some "logic" FET's that would serve without additional circuitry to drive substantial loads from logic levels. Those were what was used back when FPGA's weren't so readily available. Back then, you controlled the solenoids with an MCU and a PIA of one sort or another, all operating from 5 volts. Those parts are still reasonable in cost.

Yes, you can easily justify an FPGA as a forerunner to an ASIC, which, BTW, can have 5-volt outputs if they're needed, while the FPGA can't. The economy doesn't come from the FPGA, though. It comes from the eventual use of the ASIC which won't require $40 in level shifting buffers and passives together with the $30 FPGA. The ASIC costs a major fortune to develop and build, but, once in production, costs ~$1 or less. That means you can't fool around with what you require. FPGA allows you to "play" with your design. They're the weapon of choice if you want to play. If you want to do useful work, you have to think through the reuirements first and compute the costs on a real BOM, not just the word of a salesman.

RE