I have been looking at the RC2014 and I think that something like that would certainly be really cool looking (if a bit less cost effective than the shield / stack system).

Cbmeeks, have you put any thought into the design of such a system? If so, would you care to work on a sort of standardized connector (I am thinking some sort of card edge connector so that there is no danger of bending pins when inserting / removing cards).

_________________
- Billy

Not at the moment.

I've recently been given a $45 credit with PCB Way for a first time purchase. They're obviously trying to make a loyal customer out of me and if they keep giving me money, they just might. lol

Anyway, $45 will go a long way for the SBC that I am designing so I've been focused on that lately.

_________________
Cat; the other white meat.

I like this idea very much indeed. Mostly because I tried breadboarding a 6502 computer, and found that some of the connections were dodgy/not connecting. I then tried rebuilding it to make such things easier to debug, and suddenly found it very hard going. I do eventually want to put this on a PCB, but the design is not yet totally final.

Personally, I'd prefer that everything be socketed, and that the clock circuit be an oscillator can(again, socketed) on the CPU board. I'd also prefer to do my address decoding on a separate brick, as I am using a 16V8 GAL for address decoding

I would pay for something like this, socketed for the appropriate ICs. I really want to get something going, but it's been really painful doing that the way I've been trying to. This would make it so much easier to get the design working, and then design a PCB based on it.

There is a cheap and accessible half-way house, which is to use a protoboard and point to point wiring - the Verowire system is pretty handy. It's like a poor man's wirewrap - wirewrap is great in principle, but not cheap.

I built such a thing around the 65C134. I never put together a working machine though because I got disillusioned with the disadvantages of the 65C134 shortly after building the first board. I'll have to see if I can find it. It was basically the CPU with the exact clock circuits and layout specified in the datasheet, plus a reset circuit, and a bunch of pin headers and resistors.

There's really no reason I couldn't use it the way it is, but the whole reason I built it was to prototype an application I thought I had for the '134. I cancelled that project when I realised the '134 plus all the support and glue would not fit in the space I needed it to fit in. If I ever get time (not looking likely anytime in the next 12 years) I'd like to start over with a 65C02. I think I can make that fit.

I stumbled across a web page where a guy is tweaking the RC2014 design for 6502: http://danceswithferrets.org/geekblog/?p=582
Initially running at 1Hz, in order to be able to visibly observe everything that happens, without single-stepping.. nice one.

I have also thought of a prototype system. But rather having bricks I thought of having a base system with the CPU the RAM a fully programmable GLUE and a MCU assisted bootstrap loader and a small expansion bus. I prefer a prototype system that has as many connections already made on the PCB. My idea would be to have the data bus already be wired as well as most of the address bus (let's say A0 to A13). Also some other signals like PHI2, a read and write signal qualified with PHI2,reset, et. al. would be prewired. The rest of the PINS would be connected to some breadboard style section where you can complete the required connections. I'm actually looking for such a breadboard that can be soldered to a PCB but was not able to find one.

My L-Star project might be doing something in the direction you want to go, except there's no breadboard or glue logic.

It has a Propeller to bitbang the 6502 address bus and provide (PS/2) keyboard input, (monochrome) video output, a serial port (also used for downloading firmware). The Propeller also emulates ROM and RAM, but it's limited to 32K (including its own firmware) so it's possible to add an SRAM chip where the Propeller controls which addresses are used as ROM, which are RAM and which are neither.

I'm working on a kit (available soon on Tindie) with a 128KB SRAM chip and an expansion port that includes all the signals (data bus, address bus, R/~W, clock, ~IRQ/~NMI/RDY/~SO/BE/~RESET, RAM enable, Propeller reset). I think my first expansion board will be a KIM-1/Elektor Junior board but it should be easy to implement other boards with programmable glue logic and a breadboard.



===Jac

You guys have really inspired me to get back on the wagon and get some of my ideas finalized as real products.

I admit, I've been in a slump lately with this stuff.

I can also highly recommend the L-Star computer. It's a great product!

_________________
Cat; the other white meat.

I've gone and designed my own(that may be bad).
I have them on a sheet of paper(photos attached in zip), and I'm in the process of putting them into KiCAD. If/when I finish them all, I'll post up the KiCAD project files here in a ZIP file.

The biggest problem with the whole stackable header idea is that such headers are nearly unobtanium, especially in Australia, unless you buy a set for making Arduino shields, which are not too usable for this purpose.
Thankfully, I found an Instructable(here:Link) on making your own.

I buy mine here:

http://www.taydaelectronics.com/connectors-sockets/stackable-headers.html

Not sure if they ship to Australia but I've had lots of good luck with that company.

_________________
Cat; the other white meat.

I've actually built some crude prototypes using dual row headers, right-angle dual row, etc. But I was curious as to what people think of such headers. Seems like single row male/female headers are super cheap and easy to find. But, as you say, makes the board much longer. I would hate to build something that people couldn't find headers for. Or, have too much trouble soldering.



Would you happen to have any Digikey or Mouser part numbers? The 96-pin connector sounds interesting. And, what are the card cages? I'm afraid that's a new one to me.

Thanks.

_________________
Cat; the other white meat.

http://www.mouser.com/Connectors/Backpl ... /_/N-axj5j

Yes, stackable headers are a problem, they're difficult to find and they cover up hardware that you might want to probe with an oscilloscope or logic analyzer. I used standard .1" IDC headers with 50 pins so it's possible to use old SCSI ribbon cables to connect boards (though I realize there might be problem with reflection and delay if the ribbon cable becomes too long).

I also initially designed my board so that the connectors and jumpers could be removed (though I'm not sure if it's still possible to physically do that without the need to add wires). I really like the RC2014 idea of having small (non-stacked) expansion boards plugged in to a motherboard. I may do this in a future version of the project but I'm trying to avoid further feature creep for now.

===Jac

Hi Jac,

yes your L-Star is exactly something in the direction I was thinking of. The only difference would be a bus for the external signals. I would use the 64 pin DIN 41612 connector form C with 2 x 32 pins. You can get very cheap PCBs for the bus when you restrict yourself to 10x10cm.

Do you mean like these?

http://www.mouser.com/ProductDetail/HARTING/09032326845/?qs=sGAEpiMZZMud62t8luTOJmGADOg%2feYQXFB8KSyzv3Cs%3d

_________________
Cat; the other white meat.