Vince Briel's Newest Replica of the Apple 1
David Betz
Posts: 14,516
Vince has retired his "Replica 1 TE" and has come out with a tenth anniversary version of his Apple 1 replica that uses a Propeller for many more functions than the TE did. If you want to get in on the numbered boards you can preorder starting on Thanksgiving Day.
http://a2central.com/5300/briel-replica-1-tenth-anniversary-limited-edition-announced/
I'm not associated with Briel Computers, just a satisfied customer. I figured people here might be interested because of it's use of the Propeller chip.
http://a2central.com/5300/briel-replica-1-tenth-anniversary-limited-edition-announced/
I'm not associated with Briel Computers, just a satisfied customer. I figured people here might be interested because of it's use of the Propeller chip.
Comments
It took a lot to not buy a nice board featuring a 6502 connected to a Propeller.
No kidding, the 6502 was the processor I learned assembly language on, and the first where I was paid for my efforts. It had a packed decimal mode which was kinda handy (but slow) for arithmetic. A few years ago I downloaded a simulator to write some code. It was both a trip down memory lane and a reminder what a Turing tarpit its instruction set was.
Honestly, I want to go the other way. Put a P2 into a card that goes into my Apple //e. I started on this with P1, then decided to wait as the P2 can really do more and offer some basic things, like VGA / RGB video output of the apple graphics, which a lot of users would enjoy, and it can overlay / augment, which would be a fun lark. Once those tasks are running, there is a ton of chip power left, who knows?
Maybe have it emulate a 65C02 / 6502 and run the Apple much faster...
In any case, that's a fun board. I just got done watching the Jobs movie and it captured that early time perfectly. Recommended, BTW. Great movie.
As much as the 6502 is used all over the place, until these guys came along everyone had completely forgotten how it actually works. It turns out that the 6502 uses a "might makes right" strategy for resolving bus conflicts. Instead of installing extra transistors to gate access to some internal buses, in several situations the designers gave the data source which is supposed to "prevail" BIGGER TRANSISTORS. That's right, two registers can fight over who puts what data on the bus and BY DESIGN one of them is given stronger drive transistors to crush the other.
I still haven't figured out whether that is brilliant or WTF.
I think that design crushed the others on a cost basis, and it worked well. Cost was huge right at that time, and the "center of gravity" was going to end up important. Better devices, such as the 6809, cost more and didn't have the mindshare, nor the simple flexibility a 6502 did, and so they just didn't get adopted.
6502's don't die easy and they are still running nicely. Woz exploited the loose design in a great way too. He noticed that the actual data transfers happened at some fairly specific point in the cycle, so he basically did a RAM read on one phase, allowing the 6502 to do it's thing on the other phase. Free video display with no cycle loss, important at 1Mhz, because that means about 1.2 - 1.4 Mhz performance if not more on a by the book system with refresh, video DMA, etc... And the refresh was free as well, given the video generator part of the design scanned the rows often enough, which it did.
The unused opcodes were an interesting artifact too. Some ended up very useful, others bizarre. Sometimes people used 'em where the same variant could be counted upon. Rockwell nixed all of that with the 65C02, then added some useful instructions.