FPGA/Programmable Device Content: Seeking your feedback on what you are interested in

There is a serious learning curve to become proficient with FPGAs so the incentive needs to be pretty strong. It often takes an application where an FPGA is really the only clean solution. I chose video because if you have multiple cameras and need video processing in real time, any normal MCU will be hopelessly under powered. Of course applications where FPGAs really shine take many months to develop, so they aren't likely to be suitable for element14 projects unless somehow most of the work is already done. Some of the FPGA modules out there may have reference designs that are suitable.

Although it is easier to get into FPGAs if you are just implementing some simple logic, FPGAs are not essential for this, so simple logic applications in themselves are not that great a motive to tackle the learning curve. Obviously the learning process would involve simpler designs, but the incentive needs to come from heavy duty applications.

Sorry Doug, I have to disagree with you about the simple logic - FPGAs are utterly wonderful in such applications.

Even fairly ancient parts like the Lattice XP2 flash based family will happily clock at 100MHz, more modern parts like Spartan7 can cope with 300MHz.. I've been working on a faster than usual interface for IO on the Raspberry Pi 4  - the logic is quite simple and I've spent less than a day on the VHDL - but it would take you maybe 20 or so 16 bit wide buffers, registers and counters to do it with logic chips. You couldn't breadboard it (100MHz clock), and every time you found a new wrinkle in the Pi you would need to re-spin the multi layer pcb.

MK

I almost agree with you, Michael. I think - just for example - that some complex logic mechanisms in robotics can be easily solved by FPGA, as well as all the security and protocol interpreting and converting features. What I have not yet had the opportunity to experiment with is how a repogrammable hardwired logic as offered by FPGA - and here it is interesting just the speed you mention - can be reliable in some parts of the robotic context like the fast calculation of motion points, optimization of multi stepper motors syncronizaion, like those needed in inverse kineamtic and more and more.

Enrico

That is a good point. FPGAs are good for lots of applications. Certainly once you are well up the learning curve, it becomes much more attractive to use them in more applications. All the FPGA specialists I know love to use them for almost every application. They also like to put them on a card just in case they are needed, or as insurance against scope creep or they will use a bigger one than needed to make the design easier. There are lots of pros and cons in each FPGA decision, but ultimately they need to satisfy some weighting of factors like NRE cost, development schedule, production cost, product sustaining costs and circuit performance.

That is a good point. FPGAs are good for lots of applications. Certainly once you are well up the learning curve, it becomes much more attractive to use them in more applications. All the FPGA specialists I know love to use them for almost every application. They also like to put them on a card just in case they are needed, or as insurance against scope creep or they will use a bigger one than needed to make the design easier. There are lots of pros and cons in each FPGA decision, but ultimately they need to satisfy some weighting of factors like NRE cost, development schedule, production cost, product sustaining costs and circuit performance.