• Please review our updated Terms and Rules here

Commander X16 - The 8 Bit Guy's Dream Computer Discussion Thread

K

Kevin Williams

Experienced Member
Joined
Jan 12, 2012
Messages
73
Location
Fort Worth, TX
Hello Everyone!

For many months this year, we have been working with David Murray, The 8-Bit Guy, to design the motherboard for his "Dream Computer". This machine is now known as the Commander X16. David originally posted a video on this project back in February 2019.

Part 1 of the video is here: https://youtu.be/ayh0qebfD2g

Part 2 was released today, and it shows how much progress our team has made in the last six months: https://youtu.be/sg-6Cjzzg8s

Since that time, we worked with many different products, parts, and team members before we were able to finally settle on a basic architecture. We still have a lot of work to do, but now that the prototype of both the motherboard and video card are in-hand and working, it feels like real momentum to me! As such, I thought it would be nice to have a discussion group for this project.

I can not share schematics, or code for the video card at this point. Not only would we like to ensure that... uh hum... "parallel" projects do not emerge before we are complete, but we are also not complete yet. It would be premature to release more than the memory map, etc. as bad information may just forever float around. It is far better to wait until we have everything nailed down.

I would like to thank the team members:

Frank van den Hoef - Vera Video Module
Michael Steil - Kernel (kernal? :D) Development / Coding
Michael Allison - Monitor / Assembler Coding
Christian Simpson (Perifractic) - Case/Logo Design & Documentation
Andrew Owen - Case Design
Lorin Millsap - HW/SW Design Advisory

And of course thanks so much to David Murray for making this all possible!

This is by no means a comprehensive list of the contributors, however I will talk to Christian about adding a special thanks section in the documentation to name everyone who contributed to the project.

Here are links to current resources:
Facebook Group: https://www.facebook.com/groups/CommanderX16/
Emulator (WIP): https://github.com/commanderx16

Also, be sure to see David Murray, The 8-Bit Guy, in person at VCF Midwest in just a few days! He will have a working prototype with him that you can actually put your hands on! This will be the first time it has been seen in public, so get there if you are able.

Thanks so much for all of the support from the vintage community as well! There is no better community on the Internet in my opinion, heck, half of you guys probably built the Internet to begin with!

Kevin Williams
https://texelec.com
 
Watched the videos yesterday and I am beyond impressed by this. Congrats on this achievment so far.
 
I also watched the video, it was very interesting where things were and how they got to their current stage. I didn't see mention of it, curious if there's a place where current beta's of the emulator will be available. Would be nice to get that out to folks early so they can start learning it and coming up with some software options as that is what usually makes or kills this kind of project. It's good he has you involved since you're an expert in board design (well, except maybe that upside down parts.. hahaha) and retro projects, so we know that piece will be great.

Not excited about the case designs, being an 8-bit retro enthusiast, i'd rather see something more reflecting of the retro style.... but I can see why it's being done that way.
 
Neat. I'm not sure if it's something I would buy but I appreciated David's in-depth explanation as to how it would work.

I too was a little bit 'meh' by the case design ideas thus far and would have preferred a more retro wedge-shape. But I get it.. there's no way to have everything and have it be $50. I don't think it will be $50 anyway though.

Going back to the first video it's disappointing Cloanto didn't even respond to David's inquiries. What is the point of owning the Commodore trademarks and IP if you're not going to deploy them?
 
So... if you're looking for honest criticism I suppose I would probably have this main observation:

I'm not sure it really makes sense to dispense with making the greater direct addressing space of the 65816 available because doing so would require de-multiplexing the upper address and data lines. Certainly it does add some bother and the need to add... five or six (a guess based on the schematics of an old IBM PC-compatible motherboard) more chips to produce a de-multiplexed and buffered version of it, but that's standard fare for competing chips like the 8088. In an IBM PC the address/data bus is de-multiplexed once and from that point the peripheral chips don't need to worry about it at all. It sort seems at least like a similar ISA-like design would work fine for a 65816 machine. Another half dozen chips is certainly significant, but in the grand scheme of a Micro-ATX size board that doesn't seem like a deal-killer, and it could probably be incorporated into a CPLD easily enough for a more compact version.

I suppose I could could go really out on a limb here and suggest that you could make the expansion slots of such a computer partially ISA compatible; it's a simple, well-understood bus after all, and there were a few 68008-based hobby machines that had ISA slots so there's precedent. With 24MB of address space available you could set aside 1MB for memory-mapped cards, and... this may be a terrible idea, but maybe you could support Port I/O by mapping things so access to a neighboring memory space (or simply to the bottom...10 bits of the slot address space) would activate the IOR/IOW signals instead of MEMR/MEMW. DMA support might be a bit much to ask but IRQ mapping might be possible?

Of course, I'm saying this mostly out of ignorance because I don't know that much about the 65816. Would you still need alternate memory mapping hardware to access I/O placed higher than the 64k mark in the memory map if your main OS kernel is a direct 8-bit port of the Commodore 65 Kernal, or can that be fudged over inside the 65816 itself?

(Note: to be clear, this is just general spitballing on the topic of making some kind of "hackable" alternative-architecture 8/16-bit retrocomputer; I totally get that these ideas probably don't align with the driving vision behind this particular dream computer. What you have certainly looks like just the thing to scratch a Commodore-specific itch, and I totally wish you luck with carrying it to conclusion!)
 
Last edited:
Kevin Williams said:
I can not share schematics, or code for the video card at this point.
Click to expand...

Not knowing what the capabilities of the video card are made it very difficult to talk to David after VCFMW about a game programming idea he had. He was wondering out loud about clipping and I started to describe bit-level clipping until I realized I had no idea if the card even had built-in clipping, making the conversation moot. Are the official video specs online somewhere?
 
The 65c816 does work, sort of. The Kernel timing on initialization / video refresh seems to be a tad off, but code will run and it will be designed to be interchangeable on the final board. We'll likely address that before the end, so it should work. (Of course, you will have a few hurdles extending the bus based on the motherboard design.) I'm sure you guys are aware that the 02/816's are weird beasts in that the data/address lines are multiplexed on the 816 and the 02 has an unknown state on Phi2 low. When we initially began development, we did a lot of research into the inherent stability of this 816 and even spoke to a few folks at WDC and decided it may not be worth the hassle. In retrospect, I don't think it would have been as bad as we thought, but the reality is we are pretty married to the design now for a lot of reasons. I'm sure David will talk about this more, as there have been a lot of hoops, hurdles, and compromises made, just like with any system. In addition, Michael Steil, who is a WIZARD, btw :), has expressed how much of a kludge the 816 really is, and how it's not fully compatible with the 6502. Also, the 65c02 is a bit more advanced and is quite a bit different then the original NMOS 6502. While it does not have 16-bit registers, this machine seems to be plenty fast to write stuff in C. Heck, the Kernel mostly is. The Kernel makes extensive use of RAM and ROM banking already and Michael has extended basic into 2 or 3 ROM banks now and no one can even tell. :) There isn't one design that will please everyone, and I must admit, I too wanted to use the 65c816 originally. As time has gone on, I'm actually quite happy with the performance and behavior of the system. Dealing with 8-bit registers is nothing new for the seasoned 6502 crowd, and with the amount of dev that is happening already, I think we're going to be good. :)
 
Once it becomes available for sale I hope they can actually deliver something. My Spectrum Next and Mega 65 are over a year past the delivery date. :cautious:
 
I respect the Mega65, but stylistically I can't get over the faux 3.5" bay :)

To me, the X16 is looking like what the almost-was Color Computer 4 would have been (recall the CoCo3 had 512KB RAM, but with the time/budget they had it focused on cool-new-video and tragically neglected built-in audio -- the X16 fixes that).

As we all know, the Tandy TRS-80 was the first meaningful home computer :eek:😲 (haha) The first to hit 100,000 units sold (prove me wrong 🥊 -- the others barely filled half a stadium worth of sales). But yes, then in 79-80 Tandy lost that edge. The "80" meant Z80 processor, which they transitioned away from with the subsequent CoCo series. Would be interesting to resurrect CP/M, but there were some interesting OS's on the CoCo3 (FLEX, OS-9).
 
Last edited:
Fair enough, there were some photos where it looked pretty blank behind that enclosure. I'm ok with SD cards on retro systems, not messing with mechanical drives (although I suppose some folks might enjoy hearing them spin around) - but sure, that's just a casing style preference, can saw that quarter of the case and route the drive somewhere else.

My understanding on the X16 is the BASIC was updated to have commands/API for mouse support, so I wouldn't want a super wide case, rather have space the for the mouse (or trackball).

See MOUSE command here
github.com

x16-docs/X16 Reference - 03 - BASIC.md at master · commanderx16/x16-docs

Commander X16 Documentation (CC BY-SA). Contribute to commanderx16/x16-docs development by creating an account on GitHub.
github.com github.com
 
Last edited:
Well, the Mega65 uses the case design of the Commodore 65 / 64DX - that was the whole point behind its creation. If you do not care for that, the X16 is certainly the better choice.
 
Saw the latest video about the X16 - I gotta say it's kinda fascinating watching the team trying to figure out manufacturing processes and such while discovering some of the pitfalls of their plans. Plus seeing the soldering machine and pick a part machines and such. The $500 price tag for the first units is a bit steep for me but I respect the heck out of them for getting it this far. I wonder how much of their development challenges mirror those of early pioneers like SWTPC, etcetera. Obviously some things are different - like the Yamaha chip inventory disaster. But I imagine a lot of pioneers found themselves not able to find enough of certain key components from time to time, albeit different reasons. I am doubtful about being able to get the per-computer price down to $50 (I think that's what David said). I'm thinking in consideration of various costs and inflation, maybe $200-300 is more realistic.

BTW the Yamaha chip problem didn't surprise me at all - I find electronics parts vendors to be kind of oily generally. If they're not misrepresenting or not paying attention to their inventory, they're jacking up prices as soon as someone hits them with a substantial order. I thought the latter would be the problem for the X16, not someone having a 10 or 100,000 unit discrepancy in their inventory.
 
falter said:
If they're not misrepresenting or not paying attention to their inventory, they're jacking up prices as soon as someone hits them with a substantial order.
Click to expand...
The problem is that most don't even have an inventory at all. One wholesaler says he has 10.000 pieces in stock and then many resellers rely on that - thinking they can just order from the wholesaler on demand. But if ten resellers do, the total stock to begin with is still 10.000 and not 100.000.
 
Timo W. said:
The problem is that most don't even have an inventory at all. One wholesaler says he has 10.000 pieces in stock and then many resellers rely on that - thinking they can just order from the wholesaler on demand. But if ten resellers do, the total stock to begin with is still 10.000 and not 100.000.
Click to expand...
Yeah that sucks. What drives me nuts with these guys is they have these incredibly long RFQ forms, they make you spend a half hour just setting up for a request, and then it's like - actually, no we don't have that.

I guess that's a hazard of these 'retro' projects that rely on original parts.
 
falter said:
Saw the latest video about the X16 - I gotta say it's kinda fascinating watching the team trying to figure out manufacturing processes and such while discovering some of the pitfalls of their plans. Plus seeing the soldering machine and pick a part machines and such. The $500 price tag for the first units is a bit steep for me but I respect the heck out of them for getting it this far. I wonder how much of their development challenges mirror those of early pioneers like SWTPC, etcetera. Obviously some things are different - like the Yamaha chip inventory disaster. But I imagine a lot of pioneers found themselves not able to find enough of certain key components from time to time, albeit different reasons. I am doubtful about being able to get the per-computer price down to $50 (I think that's what David said). I'm thinking in consideration of various costs and inflation, maybe $200-300 is more realistic.

BTW the Yamaha chip problem didn't surprise me at all - I find electronics parts vendors to be kind of oily generally. If they're not misrepresenting or not paying attention to their inventory, they're jacking up prices as soon as someone hits them with a substantial order. I thought the latter would be the problem for the X16, not someone having a 10 or 100,000 unit discrepancy in their inventory.
Click to expand...
I don't understand why they don't outsource the assembly of the gen 1. It would be faster and cheaper.

I think the $50 target is the gen 3 which is going to be just one FPGA or SoC/emu. Mouser is now selling the Agon Light for $60 so it might be possible to get close.
 
You must log in or register to reply here.
Back
Top