65c02 computer motherboard design feedback

[kalj]

GARTHWILSON wrote:

I have ±12V (actually anything from about 9 to 13.5V, adjustable at my power supply but very non-critical in most cases) on my workbench computer for several things.

Wow, that build is something out of this world. Thanks for writing that detailed page about it!

[Proxy]

kalj wrote:

Regarding flash vs EEPROM, I guess I went with what I that suited the style / era.

if that's the goal of your build then go for it, it's your build after all!
i'm just usually trying to squeeze as much function out of as little Space, Components, and Money as possible. and i wasn't alive when all of the 8 bit stuff happend so i don't feel bad for throwing very modern components into my 8 bit computers.

kalj wrote:

But sure, there are probably endless improvements possible. Now that you mention it, are there any good flash chips that suit the bus of a computer like this?

the 65C02 bus is very simple so basically any asynchronous parallel Memory (ROM/RAM) will do.

though personally i like and use the SST39SF0x0 series (x indicates the capacity, 1 = 128kB, 2 = 256kB, 4 = 512kB)
they come in easy to use DIP packages, or PLCC if you need them a bit more compact. and they also have 5V or 3.3V variants.

and i mean, just look at these prices... ~2 USD for the 512kB/5V/DIP version!
https://www.mouser.com/ProductDetail/Mi ... zrqJ6nyQ==

[GARTHWILSON]

Proxy wrote:

and i mean, just look at these prices... ~2 USD for the 512kB/5V/DIP version!
https://www.mouser.com/ProductDetail/Mi ... zrqJ6nyQ==

I thought you were talking about something faster than EPROM. This one is not.

[BigEd]

Your link is to a 70ns chip, I think - that's more than fast enough for a 1MHz design! Have I missed something?

[Proxy]

GARTHWILSON wrote:

I thought you were talking about something faster than EPROM. This one is not.

when i looked up the 28C256 on mouser most of them showed an access time of 150ns.
there are obviously faster ones, but i don't think they'd be as cheap...

though i'd kinda love to be proven wrong, faster ROM at an equivalent price would be very neat!

BigEd wrote:

Your link is to a 70ns chip, I think - that's more than fast enough for a 1MHz design! Have I missed something?

yes it's 70ns, and i also just kinda realize how that might be a bit overkill for a 1MHz system, seeing as the ROM is fast enough for RDY-less 16MHz operation.
but then again it gives the glue logic more time to do it's thing.

[BillO]

3 or 4 years ago I got 10 SST 27SF256 EEPROMs for $15. They are 70ns units. They also came in a 55ns version if I'm not mistaken. There were probably even faster EEPROMs that I don't know about. I have EPROMS that run at 45ns and there are 35ns ones available too. Although these are 'ancient' technology, flash has not exactly blown them away in the speed department.

[Proxy]

BillO wrote:

3 or 4 years ago I got 10 SST 27SF256 EEPROMs for $15. They are 70ns units. They also came in a 55ns version if I'm not mistaken. There were probably even faster EEPROMs that I don't know about. I have EPROMS that run at 45ns and there are 35ns ones available too. Although these are 'ancient' technology, flash has not exactly blown them away in the speed department.

alright there seem to be quite a few fast EEPROMs, even directly from STMicroelectronics, that i somehow never found when browsing for components....
i'd still say Flash is cheaper in terms of kB per USD, but i should stop with claims like these before i derail things further... sorry.

to get a bit more on-topic again, one thing i would add on the motherboard is an Interrupt Button.
on your schematic the NMI input is simply pulled high so i assume you didn't intend to use it for anything anyways, so might as well use it for a Button (with a debounce circuit).

the main reason for an NMI Button is that it allows you to instantly interrupt any running program, which can be useful for debugging or just if you want to get out of an endless loop or crashed program.
others will probably have some better examples for the uses of such a button, but they are basically endless.

also i hope you plan on using a socket for the Oscillator instead of directly soldering it onto the board. that way it's less of a loss if the board doesn't work or you want to replace the Oscillator with a faster one to see if it would work.

[kalj]

Yep, the oscillator will sit in one of these: https://www.mouser.se/ProductDetail/535-1108800

Regarding an NMI button for getting out of a crashed program - couldn't I already achieve the same thing by pressing the reset button?

[BigDumbDinosaur]

kalj wrote:

Regarding an NMI button for getting out of a crashed program - couldn't I already achieve the same thing by pressing the reset button?

No.

Reset starts the microprocessor (MPU) from scratch, whereas NMI interrupts the MPU and temporarily gives control to another program. If, for example, your program gets stuck in a loop and you can interrupt it with an NMI you can then examine registers and/or memory to determine why the program got stuck. That assumes you have a machine language monitor resident in the firmware or whatever constitutes the unit's operating system.

On the other hand, if you reset, some of the MPU registers get returned to power-up values and depending on how the unit's firmware is built, memory may be altered such that you won't be able to do a post mortem on your program when it got stuck. Also, reset may affect periperal devices and return them to their default state. So a reset can be considered "destructive," whereas an NMI is not.

[GARTHWILSON]

The NMI affords the advantages BDD cites, about being able to examine registers and where the program was stuck possibly in a loop whose exit condition is never met. I have such a button on my workbench computer, but I've never used it. I just do the reset, since usually as soon as it crashes I know what bone-headed thing I did wrong; and then I have the startup give me three options:

1. start fresh, discarding all material I had compiled and assembled in its RAM
2. keep the compiled and assembled material that's in RAM and go to the command prompt
3. keep the compiled and assembled material but then execute the routine whose starting address is in the INIT-AP variable which may set up the I/O, timers, etc. again and re-start whatever process I was working on, so from crash to being back in the ap may only take a couple of seconds

[kalj]

Okay, so sounds like something that is rarely used, but which could still come in handy.

What is the most space efficient way of doing the NMI debouncing? I'd prefer if I didn't have to fit an extra 74HC14. I guess I could try to duplicate the reset circuit with the ds1813, but maybe that gives strange effects on power on?

[kalj]

Oh, and one final question. Which would you go with for the power socket: DC barrel connector or USB B / USB mini? I can't make up my mind.

[GARTHWILSON]

I personally would definitely go with the DC10 barrel-type connector.  Common, thin USB cables and tiny connectors will have significant resistance.  I designed a thing for work a couple of years ago that they wanted to charge by USB, and I had to limit the USB current draw to 250mA, because although baseline USB power supply current capacity is supposed to be .5A, pulling that much caused so much voltage drop that the 5V microcontroller I used to control the battery charging would go into brown-out reset.  There are at least nine kinds of USB connectors, making one wonder how they think they can call it universal serial bus.

[Proxy]

i would personally go with USB, USB-B to be exact. i've been using it ever since my first SBC and it's been working very well.

like a Barrel Jack Connector the USB-B Connector is very sturdy and easy to solder, which is one reason i recommend it over something like Micro USB or USB-C, which would require SMT Soldering and could snap off the board easier.
also i have rarely seen a thin USB-B Cable, even the cheap ones are usually thick and somewhat well shielded.

and while there is just 1 USB-B Connector, there are a few very similar looking Barrel Jack sizes. but then again you can just buy one of those universal Barrel Wall Adapters where you can swap out the connector at the end and even adjust the voltage.

i doubt voltage drop will be an issue unless you try to draw more than ~500mA or use a really really long cable.
my 65C02 SBC+VGA Card connected to my PC draws around 250mA @ 4.82V. but with just the SBC it's around 60mA @ 4.95V.
with your system as it looks right now i doubt it would raw any more than around 100-150mA (with some kind of serial expansion card).

or you could just go the Arduino route and include both, so you can just experiment around to see whatever works better for your use case, and then on your next project you can just leave out the connector you didn't like.

[kalj]

I believe I found an answer to my NMI/DS1813 question:
viewtopic.php?f=2&t=2871&start=0#p31909
Nice!