[Video] Building the DFB1r5

[Badwolf]

Afternoon all,

Boards for revision 5 of the DFB1 have arrived and it's time to embark on what I (dearly!) hope will be the first public version.

A few bodge wires I don't mind but if there are any cut trace requirements, I'll be a sad panda.

Here's part one of the assembly series. Passives and the CPLD.

Cheers,

BW

[Steve]

Very cool! Good luck!

[JezC]

@Badwolf I'll be following the progress with interest - this one seems to have been turned round pretty quickly!

(Either that or I've been so preoccupied with all the other developments on the forum plus my own H4 + TF536 woes that time really has gone that quickly)

[Badwolf]

Morning all.

Part two in the series is now available.





Spoilers: it's good news.

BW

[derkom]

Spoilers: it's good news.

Looks like the spoiler is already in the YouTube title image.

[pixelpusher]

Morning all.

Part two in the series is now available.

Spoilers: it's good news.

BW

Great project and entertaining and educational videos.

Thx!

[Badwolf]

Spoilers: it's good news.

Looks like the spoiler is already in the YouTube title image.

Haha. That’s from r4 and it’s the target. Not there with r5 yet

BW

[Badwolf]

Part three is up.

If nothing else works from this point foward, I think we have a viable release.

Todays target: TT-RAM.







BW

[Badwolf]

The final part of the DFB1 build series, part four, is up.





In this episode we target the FPU and run some benchmarks. Frontbench plays us out, as ever.

There are likely to be other videos on DFB1 as I improve the firmware and investigate issues (currently I have a suspected faulty CPLD), but I now consider the DFB1 project to be suitable for full release.

I'll be making firmware available on the GitHub page later this week, with luck, and I've already started updating the wiki if anyone wishes to look into how this was made.

~~~

I started thinking about this back in 2019. I loaded the MC68030 user's manual onto my iPad in July of that year and that kept me quiet on a transatlantic flight to Toronto whist my wife was asleep. I became distracted for a few months with an Acorn A3000, but picking this back up in November of the same year I asked a question on Atari-Forum looking for some guidance (https://www.atari-forum.com/viewtopic.php?t=37897).

Within the week I'd wired up an Arduino to the expansion port to investigate bus arbitration. I succeeded in stealing cycles from the CPU whilst leaving the system stable. As one of my friends put it, I'd successfully built a decelerator!

Later that month I ordered the circuit board for DFB1r1. Getting the first revision board to work at all was the hardest part, but once I was able to boot the diagnostic cartridge from an 'external' CPU, the problems, whilst maybe taking as long, became more trivial at each step.

The next big leap was about a year into the project: the move to SDRAM to take my expansion card from 1MB TT-RAM to 128. Another year went past finessing niggling smaller issues. The DSP took the longest time to figure out. The FPU needed a number of redesigns. The amount of money spent on blown up CPLDs doesn't bear thinking about.

Of course I've not been working on this full time and I've been distracted by software projects (like FrontBench) or other hardware hacks (like the ST booster), but here we are, two and a half years later and, despite perhaps a bit of a stability issue with the CPLD on this first rev 5 board, I'm drawing a line and calling it job done for the hardware.

I'd like to go and sit in the garden with a gin.

But it's only Monday. C'est la vie.

BW

[JezC]

I think you deserve that G&T no matter what day of the week it is! Great result...I'd better start saving up the pennies to buy the bits for one of these!