Key Components Part 2: Graphics and Sound Cores.

A project log for BoxLambda

A retro-style FPGA-based microcomputer. The microcomputer serves as a platform for software and RTL experimentation.

EpsilonEpsilon 05/01/2022 at 11:550 Comments

I spent some time researching graphics and sound options for BoxLambda. Here’s what I came up with.


If you’re reading this, you must be into the build-your-own-computer thing, which probably means you’re aware of the super cool Commander X16 project. Frank van de Hoef created the very elegant VERA (Video Embedded Retro Adapter) module for the X16. Here’s a high-level specification, taken from the Commander X16 website:

VERA module specifications:

Other features, not mentioned in the blurb, include:

Lucky for us, Frank recently released the VERA verilog code under the generous MIT license. You can find the code here:

I’m not particularly interested in VERA’s PSG (Programmable Sound Generator), or the non-VGA output formats, so I might remove those from the build.

The 128KB of video RAM will take a big chunk out of our available Block RAM resources, but it’ll be worth it. We’re getting a lot of bang for our buck.

Note that the VERA is designed as a separate FPGA with a SPI slave interface. Some modifications will be required to integrate it into our SoC.


I also considered, but eventually dismissed, Xosera:

Xosera is a VERA-inspired video controller, but it is being developed independently by Xarc. I like the Amiga-style Copper processor that they added. Unfortunately, Xosera doesn’t have hardware sprites. That’s a showstopper for me. I’ll keep my eye on this project though. It’s an active project and features are still being added.


A sound core is a perfect candidate for Partial FPGA Reconfiguration. There are a lot of options (Wave-Table synthesis, FM synthesis, PSG…) and a lot of open-source cores available. It would be pretty cool if the software application can just download its synthesizer of choice as part of the program.

Pretty much any core developed by Jotego sounds like a great idea.

Technically, I don’t have to select a sound core. We already have sound through VERA’s PCM audio playback. I’m going to select a sound core anyway because I like retro sounds and I’d like to mess around a bit with one of the old-school PSG chips.

I think I’ll go for a dual YM2149, one for music, one for sound FX, in a game context. The YM2149 was the Atari ST’s sound chip, so we’ll have a large music and sound FX archive at our disposal. Jotego developed an FPGA clone of the YM2149, the JT49:

Why not VERA PSG?

The only reason I’m not going for VERA PSG is that, as of yet, very little music has been written for it. I’m sure it is a perfectly adequate PSG implementation.

Why not SID?

The SID chip is partially analog, making it much harder to emulate correctly on an FPGA. Also, while I like SID, I’ve probably heard enough SID music to last me a lifetime. I’m currently more interested in finding out what other retro sound chips have to offer.

Interesting Links