Clarification about ATF750

[IvanDSM]

Hi there!

I apologize in advance as I'm quite the newbie in the PLD/CPDL arts. I saw in the README that the reason for not considering the ATF750 in this project was the lack of JTAG programming support, which could make it difficult to program this device.

However, while reading the ATF750 datasheet I noticed it claimed to have compatibility with standard JEDEC files for the V750, which I thought might facilitate the programming process. I also found a post on 6502.org that claimed that it was programmable with afterburner and a Mega 2560. Does this warrant reconsidering the coverage of the ATF750 in this project?

Thank you for this great repo!

[peterzieba]

Hi there!

I apologize in advance as I'm quite the newbie in the PLD/CPDL arts.

Not at all. The intent was to make the topic as accessible as possible and if there's anything that could be done to make it more so it is good to hear it. I plan on updating things a little now that WinCUPL II has been released.

The TL;DR is that you can absolutely use the ATF750 and it will work just fine using WinCUPL and you can ignore any instructions related to the fitters as they are not used for the ATF750 (though the simple shell script/batch file here will refuse to work without the newer fitters, however, the new release of WinCUPL II has the new fitters built in). And yes, the Afterburner project should work. While I haven't used it personally, I engaged the author on the possibility that extending support for the ATF150x series might be possible with the remaining codespace on the Arduino UNO, and sure enough he managed to implement all of this extra functionality in the tiny space of the UNO. It's an awesome project.

I saw in the README that the reason for not considering the ATF750 in this project was the lack of JTAG programming support, which could make it difficult to program this device.

However, while reading the ATF750 datasheet I noticed it claimed to have compatibility with standard JEDEC files for the V750, which I thought might facilitate the programming process.

So, the ATF750 sits sort of in between the capability of an ATF22V10 and an ATF1502, but is more expensive than either. If you need just a tad bit more capability than the ATF22V10, have no intention of scaling something, still want to be able to place it on a breadboard without a custom PCB and don't mind the extra programming fussiness, then the ATF750 is the way to go, but you'll still be ordering a custom PCB for the afterburner itself -- to my knowledge no one sells one assembled, ready to go.

If concern for solderability of fine pitched parts is the driving factor, you can get an ATF1502 in PLCC, and then there are PLCC through-hole sockets with the usual 0.1" pitch, and you'd end up with a significantly more powerful device (Multiple clock lines, tons of IO, the list goes on) and it can be reprogrammed in a multitude of ways without the use of power rails other than 5V (unless you disable JTAG).

The use of standard JEDEC files for programmable logic is generally almost always the norm, however, all this means is that the file format a device programmer accepts is standard across these devices. In and of itself, it does not say anything about the programming algorithm (which is almost always proprietary, and even with JTAG devices, this still is not enough to program a device with something like OpenOCD. This is where conversion from a .JED to .SVF comes into play for the JTAG capable devices).

I also found a post on 6502.org that claimed that it was programmable with afterburner and a Mega 2560. Does this warrant reconsidering the coverage of the ATF750 in this project?

I believe the afterburner project uses an Arduino UNO and not a Mega 2560. You can use the ATF750 with the WinCUPL workflow and it should work fine out of the box, so I should clarify this a bit more, but still generally feel it is not a good part choice given the cost, programming requirements and alternatives that are available. Limitations are also that you won't be able to use higher level languages using Quartus or Yosys as you could with the ATF150x. The experience would be basically the same as programming an ATF16V8 or ATF22V10 in CUPL if you have a programmer that will write to it, which is not necessarily a bad thing for devices that are this straightforward.

What I want to do is figure out how the ATF2500 is programmed. While this is again a rather expensive part compared to say an ATF150x for the functionality it provides, it does have the advantage of being able to be plopped directly onto a breadboard and is a giant 5V capable DIP40. I have one in a box along with a classic programmer and if I ever have the time, I might send what I learn to the author of afterburner and see if he wants to add support for it, but it's a good bit of work just for the sake of breadboards and nostalgia, and my intent here was more about pragmatism.

The reality is that with the right device libraries in WinCUPL, or a non-Atmel/Microchip version of WinCUPL, you can quite easily create .JED files for quite a few parts beyond those by Atmel, but again, reasons to do this are not very pragmatic as all of these by other manufacturers beyond the ones mentioned here are out of production to my knowledge.

Thank you for this great repo!

Thank you! I hope it is helpful and good luck with your projects!