Jamma FPGA devkit

[lix]

Blimey, you've had a busy week. Looks very slick!

Is it just a £50 USB Blaster to program this?

Thanks! All hand routed, and I like to squash things together as compact as I can.

I got a USB blaster when I purchased one of these, which has been quite good to play with:

www.ebay.co.uk/itm/292432559605

Then I thought I'de buy a cheaper blaster by itself so I could dev at work as well as home, £7 from some chinese reseller, and it doesn't work. It doesn't appear on my home PC but actually blue screen crashes my work PC as soon as I plug it in, so I haven't figured out what is wrong with it yet. The circuit inside the cheaper blaster looks identical to the one I got with the kit, however, the case has a badly printed sticker on it (well, worse looking than the dev kit one), which I guess gives itself away as being a clone.

I've just looked and HobbyComponents have recently listed some fpga stuff, they're pretty reliable as I've had stuff from them before. They've just listed a dev kit and blaster available separately for £7.49 so I'm tempted to try a couple of those.

www.ebay.co.uk/itm/141315924702

I had some quiet time at my day job recently so managed to squeeze this project in which I labelled as 'practise' since I know I might have some FPGA work incoming soon and I've not done any till now. The next couple of months are going to be very busy for me at work so I don't think I'll get a chance like this again, hence further circuit design might likely stall for a bit.

Next task would probably be board production after any revisions to the prototype. I have access to two SMT production lines at work but unfortunately they're running at full capacity for the foreseeable future. My boss has actioned getting another pick & place robot so we can do more work for others, that's a good few months away though so I'll have to spec the project up for someone else to make if I go ahead with manufacture. Allpcb are making the boards I've designed and they do assembly, so I might check them out.

 

[Macro]

Personally I'd go for the 25k gates option, as it has enough internal memory to allow recreations of the older games (if porting from mist) which means all code graphics rom etc. can be in the RBF file.

Could be useful for a project I'm currently playing with as well ... what price you expecting them to be, I may be interested in one (for recreations, not homebrew)

 

[lix]

Personally I'd go for the 25k gates option, as it has enough internal memory to allow recreations of the older games (if porting from mist) which means all code graphics rom etc. can be in the RBF file.

Could be useful for a project I'm currently playing with as well ... what price you expecting them to be, I may be interested in one (for recreations, not homebrew)

Yes, I'm swaying towards using the 25k as the default option now as I spotted that is the gate capacity of the FPGA on the mist project, and should give a fair amount of headroom for quite a few game recreations. The MAX 10 inbuilt flash on the 10M25 has 312kbyte in total, and 172kbyte available for use outside of the primary configuration block, so should get a few roms in there.

I'm still unsure what the end price will be, but it certainly won't be as much as a fully loaded mister board is currently (although clearly I have a lot less hardware and power). As soon as I have the prototype running and any tweaks done (which I certainly need to do, I found I've put the microSD card holder on the wrong way round, duh), then I'll scale up to a potential run of 25 boards and do the costing when I get them assembled. From that I should be able to start gauging interest, lock down a price and maybe do a pre-order sale. So I'll definitely keep you informed!

The components arrived earlier this week for the prototypes, just waiting on the boards now.

 

[Macro]

I have a Mister, which I am playing with at the moment, but also have a Replay and a Papillio from earlier FPGA related projects. but this looks ideal for what I am doing at the moment and probably won't need the microSD for that anyway

 

[John Bennett]

It’s logic elements, not gates.
A 68000 uses about 60,000 gates, I’ve read, so you’d be super stuffed if this thing was just 10,000 gates

It’s not something you can just work out, but I heard a factor of 10 for LE ->gates

 

[lix]

It’s logic elements, not gates.
A 68000 uses about 60,000 gates, I’ve read, so you’d be super stuffed if this thing was just 10,000 gates

It’s not something you can just work out, but I heard a factor of 10 for LE ->gates

Ah yeah, I keep forgetting Been having trouble with my bits and bytes today also!

 

[lix]

Boards finally arrived today

I missed a bit of a cockup with the gerber files and the manufacturer held up production when they spotted the error. I've also got a couple of traces to cut and re-route, and the micro SD card holder is the wrong way round however I've bought a different holder which re-corrects the orientation, but doesn't 100% fit the grounding pads. It should do for now.

This weekend I'll put one together with the cheaper FPGA and toast it in my smt toaster oven and see if I can bring it up and get some tests done. If it works then I have enough components to do another, and if that works then I might order more parts and do a small run as I have 20 pcbs. I'll have to see. It's my first pcb with an FPGA, and my first with a small BGA part, so I'll probably take it into work and x-ray it to make sure it's all good before powering it up. I also don't know why I chose white soldermask for the prototype, it makes visually checking the tracks more difficult!

Despite printing out the circuit on a sheet of A4 the smallness of the pcb is a complete unknown quantity until you get one in your hands and find it's not much bigger than a jamma fingerboard.

Size comparison:

 

[AnDyC]

Looking good.

 

[Hurray Banana]

Great work

 

[Macro]

Can't wait to see it doing something!

 

[guddler]

Whats the development process likely to be like for this? Do you anticipate doing all coding directly to the SD card or is this the kind of board that would be simple enough for someone to create an emulator for?

Very interesting project!

 

[lix]

Whats the development process likely to be like for this? Do you anticipate doing all coding directly to the SD card or is this the kind of board that would be simple enough for someone to create an emulator for?

Very interesting project!

Well I'm hoping to create a base set of logic features to be able to get some kind of decent homebrew up and running, which indeed could be loadable as executable files from an SD card, or over a serial connection for debugging. The exact state and specification of that is under development, however it can be used as a blank canvas with the hardware as-is, and it should have enough grunt to have a decent range of arcade hardware to be simulated on it. I realise I've branched off on a tangent with slightly different memory and IO hardware compared to existing projects like MiST and MiSTer, so cores for that aren't immediately portable without a bit of work alas. I've certainly got a couple of arcade game platforms that I'm looking into building up myself that don't currently exist.

I took some inspiration from the FPGArduino project which packages a set of features for certain FPGA dev boards and links it up with the Arduino IDE. I just wanted something to be JAMMA connected, simple to use and (hopefully) cheap so I did my own hardware.

 

[Flinnster]

Amazing. Really interested, would love to develop for it

 

[Hurray Banana]

Same here, would be an interesting device to code on

 

[virtvic]

Even though I know nowt about programming or pcb design, I'm watching this with great interest. I love the idea of fpga recreations, it truly is the future of arcade gaming.

I pur myself forward as a beta tester when you have one far enough along to test roms on!

virtvic2019-03-24 08:57:06

 

[lix]

Populated one of the boards over the weekend, checked it over for shorts and bad connections, all okay. Powered it up and downloaded a verilog test project which I had written to generate a picture and test the video DACs, and it worked first time.

Also had a quick test of the fast SRAM and I could display the contents of the random memory as pixels, so that works.

Yet to bring up the audio section, the hyper ram chip, the analog inputs, the SD card and all the JAMMA edge inputs, but I've triple checked the schematic so I can't envisage there being a problem, it's just about coding the hardware logic now. I had mainly been worried about the FPGA not being happy, or my toaser oven skills over-cooking the parts. I did leave it in slightly too long as the audio amps legs went a bit off colour, but hopefully not cooked to death.

For anyone interested in the core parts I've used so far (pretty sure these will be the final components), here's some links to the parts I bought on Mouser:

FPGA

HyperRAM

SRAM

Audio DAC

So I'll spend the next week testing the peripherals and writing logic to verify everything works.

And first picture:

Steve.

 

[virtvic]

Could do with some more caps!

Awesome work!!!

 

[lix]

Could do with some more caps!

Awesome work!!!

Heh, yeah I felt like I'de gone a bit overboard. These are actually smaller ones than the caps I got in the main parts order, they were far too big! I'll pick a few lower voltage ones next time.

 

[Macro]

Now that is looking really useful!

 

[ColinD]

Yes, really like the look of this one too. Well done so far.