ScreamingAtTheRadio

Behold! I am VERY pleased.

Fantastic to see progress... In the meantime, I imported @foft's design into KiCad and put it in a git repo: https://github.com/bleroy/pokeymax. (discussion in the other thread) I'll try to keep the repo in sync with recent progress. My own personal goal in this is to reach the point where, component availability permitting, it is possible to order assembled PokeyMaxes from the likes of JLCPCB (I added a BOM with JLC part numbers in the production folder already). Because I want Pokeys and I'm afraid of BGA Thanks again @foft for making this possible.

Hello everyone. As promised, here's a git repo with @foft's PokeyMax 4 imported in KiCad: https://github.com/bleroy/pokeymax I'll also post this on the other thread that I just learned about: I also pushed the gerbers and bom to JLCPCB to get an idea, and quite a few parts are still missing, including the fpga. I'll make an attempt with PCBWay to see if their inventory does better...

Exactly right. So the challenge here is that the space is so small, that there is no such thing as "where it was originally". The "normal" place for break, right of delete, is partially out of the allowed space, so we can't give that option. As a consequence if there's a choice it will be between relatively arbitrary locations. What do you think of putting it immediately right of space? That looks like the better option to me.

I never made the break key configurable, really (well, both places were wired), it was caps and control, and even for those, the vintage had them all at the regular place, and the modern had break right of space. 0 complaints. You will never hit it by accident in that position. Maybe then right next to space like on other keyboards? Yeah, I guess exchanging the Atari key and break makes sense. Where would you see it?

So about the proper 400 replacement keyboard (not the 400 Mini / USB keyboard)... I've been thinking about it. I had convinced myself before that there was no way to make a good replacement 400 keyboard that would both fit *inside* an unmodified 400 case and have a layout that is not completely alien. Y'all's constructive feedback with the jaw picture and all got me thinking and I was wrong. While I 100% hate the weird layouts that have been produced before, there's one weird trick that we've used with the XL modern keyboards to deal with the wider footprint... On the XL, we added an imperceptible additional gap between caps to fill the whole space and avoid a big gap on either side of the keyboard. For the 400, it's the reverse situation: we don't have enough width to fit a proper layout, but we have *some* leeway to reduce the gap between caps. Not much, but enough. I've experimented with printed test rigs and it turns out that reducing the horizontal distance between keys by 1/32U (that's 3/128" or 0.59mm), with a proper plate that holds the switches steadily in place, is no problem, and the keys still move perfectly fine without a jam. 1/16U is way too much (I also tried that just to have an upper bound) but 1/32U is fine, and that gets us 12/32U=3/8U=0.375U more. We have 253mm of width (which would normally be 12.25U with standard 3/4" key distance), with some wiggle room since there's a chamfer around the keyboard window that allows the caps to go a bit out if we don't push it too much. With keys that are 31/32U spaced, we can fit a normally 13.75U layout. So @XL Freak tried a few things, and we narrowed it down to this: I don't know about you, but I find this pretty elegant: - Wide del/backs - The missing ctrl/tab/esc left column gets rotated and falls left of space - We have two control keys at the standard PC location - Return and caps are at a familiar place for Atari users - Two comfortable shift keys - Most keys are where you'd expect them One thing that's not ideal is that the delete and return keys at this size don't exist in that specific row, which means we have to move them from other rows and they will stick out a little because they'll have a different profile. Since they're important keys, it's not absurd in my opinion. We're looking for feedback on this layout before I route and order prototype PCBs and caps.

Yeah, nevermind. MaxV should in principle be cheaper but turns out 10M02SCM153C8G is way cheaper, at least at Digikey. Weird. JLC stocks them at >$43 which is insane. JLC supports via in pad.

@Panther where did you get those prices? I just checked the prices on Digikey and the JLCPCB parts library and they are not good. Might have to send them the parts before assembly, or assemble by hand... A stereo Pokey might be achievable on a 5M2210: 2210 LEs, 271 pins, 1kB flash, 1 osc and 4 global clocks?

Yes, exactly, I was wondering how low-spec we could go, including much cheaper CPLDs (MAXV and the such).

I can prepare quotes for JLC and PCBWay sometime today and let folks know here unless someone gets to it first. PCBWay tends to be MUCH more expensive than JLC though, so I suspect even 2-layers at PCBWay is likely to be more expensive than 4 at JLC, but we'll see. I had another question: what are the exact requirements for the FPGA?

You typically need a zip with the gerbers, a position file and a bom that follow a format that can vary between fab houses. I’ll build those tomorrow for a JLCPCB order, and I’ll share a link to the files here.

So the 50MHz oscillator is giving me trouble. I can find parts that are pretty close on Digikey, but they don't seem to even be on JLC's catalog. LM339 is easy to find, but not in this package. Those two might benefit a footprint swap... Not sure what to look for for FB1. Looks like some sort of regulator but I didn't find anything with a 0402 footprint that looked like that. Apart from those, some components might need a reorder delay but seems findable at JLC.

No joke, especially on such as small footprint. Seems like I wasn't looking at the footprint right. I did find them under 0603x4: Stocks are pretty low I suppose and if by any chance it helps with routing... Probably a little more annoying if assembling by hand, but I have no intention to I'm still going through the BOM, I'll yell if I have trouble finding something else.

That's amazing! If it's not too late to make adjustments, I have a couple of suggestions. I started to send fab files to JLCPCB to get a ballpark of the price of assembly, and components that were consistently hard to find or out of stock were all those resistor arrays. Discrete resistors would probably not take that much more space on the board and would be much easier to source. On my side, I've imported the existing boards into KiCad without a hitch. I'll put those on a private GitHub at first and will give you access so you can validate you're comfortable with what I'm pushing there and once you give me the OK I'll make it public.

If it's going to be a USB keyboard, the lower-cost options would be a custom key cap set over a stock keyboard. There's no real need to make a bespoke PCB unless we want a non-standard-PC layout or a custom Atari-themed enclosure for it. I like that it gives people options.