Thank you Jeff, and Ed. I think I am seeing it more clearly now.

I asked Garth a long while ago about emulation mode on the '816, and his tl;dr response was "just go to native mode". Now I am seeing why. Yes, the '816 can act like a '02, it can run '02 code, but it is not an '02. I think that is what I need to keep in mind.

I altered the original image some, as per BDD's request to leave the E pin unconnected.

But now I have a new situation, given the topic at hand:

Say I design a 65C816 circuit for a drop-in 65C02 replacement. As in, if I put an '02 into the same slot on the board, it would run, albeit not at full potential. Back to the banking. The '816 starts in emulation mode, and in Bank 0. It tells the '573 that it is in Bank 0. But an '02 cannot tell the '573 anything. We would probably get mush, whatever happens to be on the databus at the time. Our banking situation could go wild! Unless I stop this by using glue logic with the non-existant E pin, results will vary.

That is another reason why I was proposing the '273 with the pull-up resistor on the E pin (I guess I just forgot about it until now). If you have ability to divide down a faster clock, this shouldn't be an issue. But leaving the E pin unconnected will again get varied results.

Any thoughts?

Thanks!

Chad

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

It was more of a 'feeling' than anything. You yourself might not have said it like that, but that is what I got from the conversation. "If you are using an '816 anyways, why restrict your abilities?" basically.

I know my proposal is a silly one: Why would I want to shove an '02 into a '816 slot? Maybe it's just for the fun of it.

Thanks!

Chad

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

Yup. But perhaps a different solution can be applied.

Since you're already talking about removing an '816 and plugging in an '02, perhaps you'd be willing at the same time to remove the '573 and plug in a male connecter with IC-like pins. The connector would be wired so that those pins which correspond to the 573's Q outputs --ie, the Bank Address lines -- would attach to connector pin 10; ie, Gnd.

Or, forget about the male connecter and, when running an '02, simply leave the '573 socket vacant. You'd have pulldown resistors permanently connected on the mobo, and they would pull the Bank Address lines low (but only when the '573 is absent).

-- Jeff

_________________
In 1988 my 65C02 got six new registers and 44 new full-speed instructions!
https://laughtonelectronics.com/Arcana/ ... mmary.html

When I tried a 65C02 on my 65816 breadboard build I just removed the '573 and tied any bank address lines low. Not exactly a drop-in replacement and much more of a hassle on a PCB!

It feels to me that you'd be OK using the NC pin as a E signal with a pullup.

If the datasheet tells you not to connect it, that could either be because it's used internally (for testing, or an option that's not advertised) or because it might in future be used internally for something. If the datasheet also tells you it's not used internally, I think you're OK. WDC are not still developing new and exciting pinouts for their 'C02 offering.

The 65xx bus trace decoder does have full 65816 support, see the 65816 Bus Decoder thread on Stardot.

Oh goodness, big oops - thanks for the link.

I don't know what a 'mobo' is, but I get the idea. Pull-down resistors on the bank lines.



Yes I am perfectly ok with using a pull-up on NC/E! It ties nicely with the '273 (again, as long as you can divide the clock down to latch the banking data properly). Thank you for this better understanding of the NC pin.



Jumpers are nice, but I personally do not like using them. But that is literally just me. They sure would make a lot of things easier though.

Thank you all!

Chad

My approach is changing or reprograming the CPLD when switching processors.
Bill

Same here—I actually have a (very untested) version of my SBC on the drawing board that could take both with a reprogram of the CPLD.

I have also been looking into a way to program the CPLD in-circuit without having to use any external hardware… my SBC includes an independent co-processor that can speak to the outside world over serial and/or Wi-Fi, and so I've been toying with the idea of using it to send SVF commands to the CPLD over JTAG, negating the need for a specialized programmer like the USB Blaster (or, as Intel now imaginatively calls it, the “FPGA Download Cable”).

A SVF player to reprogram the CPLD is an interesting idea. If you open up an USB blaster clone…uh, FPGA download cable, you’ll see an USB-to-parallel-serial-JTAG chip. So perhaps it’s possible to design a dual-use hardware that serves both as a serial port to USB adapter as well as SVF player?
Bill
Edit, uh sorry drifting off topic…

Just popping in real quick to say that "Mobo" is a somewhat common shorthand way of saying "Motherboard".

I knew I would have to at least reprogram the ROM, obviously different code for different processors. But in this way, you get to reprogram the board itself even!



Thank you. I'm still the new guy here, and will be that way for, oh, years I'll pick up on the 'slang' eventually.