Hi All,

I know there are a great deal of you that regularly visit this forum but don't say much so I'd like to get your comments, along with the regular contributors, on a couple of important issues.

www.6502.org is off to a good start, but still has quite a way to go before becoming the resource it one day promises to be. Right now I think that it's biggest weakness is a lack of programming information, useful subroutines, etc. But there are other areas that need improvement as well.

Please give me your comments and suggestions for the site and what can be done to expand and make it better for you. I know from the site logs that a great number of people visit, but I don't get as much feedback as I'd like.

Thanks for your help.

_________________
- Mike Naberezny (mike@naberezny.com) http://6502.org

Mike,

I tried to go to www.6502.org to check what you've got on the PCB Design, and I ended up at buzz.chisp.net. I tried also from the link in your message and ended up at the same place.

What's up ?

Ted Melton

Hi Ted,

We're having some problems with the NT server, I am looking into it. Sorry for the inconvenience, it will return shortly.

_________________
- Mike Naberezny (mike@naberezny.com) http://6502.org

This site has a good link to user's homebuilt projects.

We can also have short writeups on interesting,easy to interface ICs that can be integrated in our 6502 projects.This is specially easy if we use one of the 6522s to provide a SPI and IIC interface.Of course - bus interface ICs are always welcome.

As I suggested earlier, make the computer compatible to KIM/SYM-1/AIM-65 and make available their software such as Microsoft BASIC for our project.This will make the site much sought after among 6502 users.

We should also finalise a set of standard I/O for the board(say IBMAT keyboard and LCD attached through 6522VIA) so that users can write/adapt practical software for the 6502 project PCB.I derived a lot of satisfaction in converting and improvising games described in "Advanced 6502 Programming by Rodney Zaks" to run on my system.

In fact one of the 6522s on the project board should be standardised for SPI, IIC , IBMAT keyboard, LCD and few opto isolated I/O lines. The IIC will ofcourse be a multidrop bus. By using one extra line and a SPI MUX chip such as MAX 350, 8 chip selects for SPI bus can be generated(CS generated by I/O line is MUXed to any of the 8 selected devices).Projects on the SPI/IIC bus are truly universal in that they are CPU independent - only SPI/IIC timing requirements need to be met.

Just a thought, but the KIM/SYM/AIM memory and I/O map were devised as a demo of their various LSI's. They were not intended as a basis for a general purpose computing system.

I tried this sort of thing, but the not-so good "clever" application of memory address lines to the chip selects, while cheap was wasteful of memory map. While these systems work for a while, it's really inconvenient jumping around a 4KB block occupied by a VIA that's not of much use, and it really screws up memory in an edit buffer.

Some of us in the Denver Area 6502-Users' Group, back in the late '70's used the memory map of the Digital Group 6502 CPU, which was somewhat more tolerable. That CPU was designed with a 1702 EPROM at 0xFF, I/O in page 0xFE, and a block of 2KB of RAM either at 0000 or at 0xE000 if a memory board was present. That seemed to work fairly well. I/O didn't get in the way. OF all the 6502 boards I saw out there, I never saw one done much better from the standpoint of expandability.

I still have the text listings of the Assembler, Editor, Debugger, and a couple of high level languages, including the XPL0 that was written by a couple of the guys. There was also a hacked version of the APPLE-II basic and a floppy based OS called APEX. I've considered making a ROM of most of that stuff, but I have reservations about the editor. I think it needs a different one, and the debugger needs work, too.

There are always changes that would improve things. It might work better if I/O were mapped in to Page 0xFF along with hardware (switch/jumper or small-BPROM) programmable reset/interrupt vectors. A small single-chipper, e.g. PIC or 68HC705 would serve as an interface to an IBM PC/AT keyboard interface, while a modified ISA video board would allow the use of a really cheap monochrome video board from the local thrift store, where keyboards and mono monitors typically cost $5 or so. In fact, for $10 you can probably get a video board IN a PC/AT and get the FDD and HDD for nothing.

Anyway, there's another $.02's worth

Uli

Richard,

Please feel free to send me any or all of the software and information you mentioned for posting on the www.6502.org website. If you only have hardcopies I can reimburse you photocopying and shipping costs, drop me an e-mail.

You also mentioned in the other thread circuits for the single-step and single-instruction circuits, I'd love to post them on the site also if you have them around.

Thanks.

_________________
- Mike Naberezny (mike@naberezny.com) http://6502.org

I think a standardised I/O subsystem to include keyboard and display is a terrific idea -- most importantly, if we can develop a very basic operating system around it which everything else builds on, people will find it a lot easier to share the software they develop. The alternative is essentially chaos -- everybody developing software to run with their own custom-built hardware, so anybody else who wants to use or adapt it has to build the same expansion hardware and modify heck out of the software to patch it in around whatever I/O system they are using.

I think the most important I/O device to have, along with the supporting software (hard-coded in ROM? I think so...) would be Serial I/O. A keyboard could be serial-linked, and this would limit the hardware on the main PCB. It's a bit nebulous as I'm just shooting from the hip here, but I could see a system where we have a PIC-based translator to convert a standard $5-$10 IBM AT keyboard to serial input, and then use serial output to drive an LCD display. This way we can use our IBM's/Mac's as terminals to download software etc., or unplug the host computer and plug in the keyboard and LCD for stand-alone operation. I'm assuming most people will be writing and compiling their 6502 software on a cross-compiler running on their IBM/Mac??? I think this would be the case at first, anyway.

Since most modern computers support 5-volt serial, I don't see a need to include the +/-12V converters for true RS-232 on the main PCB (keeping with the spirit of a small main PCB) -- the circuit is very easily built into a small external self-powered or 9-volt-battery powered box for those who need it. I don't think I'd even have DB-9/DB-25 connectors on the main board, just 2x5 pin headers and use commercially available adapters like the IBM PC hardware has. I do think I would like a dual-port serial interface though, one for keyboard/LCD interface and the other for external devices as desired by each user.

Any thoughts?

-- Brian Farley

Hi, Mike

Okay, I've made a couple of posts but not a lot. Still want my input? Of course you do.

Anyway, all kidding aside, the main thing I'd like to see is an ongoing schematic page for this. I haven't seen one at www.6502.org, or am I just missing it??? We've got a LOT of good ideas floating around, but getting the details down on e-paper has to happen sooner or later so we can make sure we are all on the same page. It will also allow the real go-getters to start hand building boards and see if we are being really dumb about something -- you know how things pop up in production that everybody SHOULD have seen during development, but just DIDN'T....

If somebody wants to hand-draw a shematic and scan it in for everybody to rip on, I'll be happy to convert it to OrCAD and .PDF formats so it will be nice and clean. For that matter, if somebody wants to hand-draw it but doesn't have a way to scan it in, just fax it to me at (925) 734-8618 and I'll put it into an OrCAD .SCH file, print it as a .PDF, and E-mail both files to you so you can post them where you see best. For that matter I'll be happy to do the whole thing, but that will take longer than just doing a capture-and-clean-up job, since original schematic generation takes a lot of time and cross-checking, and what free time my job doesn't suck up, my robots usually make off with...

-- Brian Farley

To share software, it must be written to use only 'logical' I/O devices rather than 'physical' I/O devices. The logical devices can be likened to the standard i/o of C using the stream concept.An 'I/O distributor' in the software then links them to physical devices such as serial i/o,terminal etc.All such redirection may be done using vectors in RAM(indirect addressing).Each physical device driver has its own encapsulated initialisation and access functions. The physical device is not visible to any other functions.

On the hardware side, we make it easy to interface and write drivers if we interface to standard buses like SPI, IIC etc.

As a start, I suggested allocation of port lines on a 6522 VIA earlier to form a eight channel SPI bus, a multidrop IIC bus, an IBM AT keyboard port(a standard in its own right) among others. Most boards will also have a UART and so a serial port already built in. The suggested allocation and its driver is by no means a final solution - you are most welcome to have your own hardware/software combo for the interfaces you require.One may even design a custom chip using a FPGA for the same.

I was therefore trying to suggest that we decide upon certain standard I/O routines.

The idea of a serial I/O port as a standard device is very good as it has become the most common standardised peripheral.It is an excellent solution for small simple boards.However, remember that you need to plug different devices to the rs232 connector in place of redirection.

I think the PC keyboard controller is a great idea- you can't beat a $10 keyboard.

The circuit for the interface is in the original PC Hardware Reference manual - I can send a scan of it to anyone interested. It used 2 LS175 gates, an S74 gate, and LS322 for the bus buffer, and 3 logic gates. I think it's overly complex, and a 6502 could easily bit-bang the interface. The actual connection is a serial link, with a bit clock. It returns 8-bit "scan codes" though, rather than ASCII, requiring a translation table.

On another note, small currents of higher voltage can be pumped from a 5V base supply, so the machine could still have only a +5V power supply and yet provide true RS-232C compatability.

WizWom - Wandering Kernel of Happiness

Wizwom was quoted as saying:

I think the PC keyboard controller is a great idea- you can't beat a $10 keyboard.

================================

Well, the AT keyboard, available everywhere (e.g. the thrift stores) for $5-10, is pretty tricky and not that similar to the PC-keyboard to which you refered. The 'LS322 is a shift-register, by the way, and the remaining hardware is used to manage the data transfer. There are several published ways to interface the AT keyboard to function as a parallel interfaced keyboard, but they all require a significant effort. That effort is provided in some cases, though, and the one I like best is that one that uses a 20-pin 68HC705J1. The reason for the J1 as a minimum is that the device has to host a lookup table for the various multi-byte scan codes the keyboard returns and convert them into single-or-double-byte strings where possible. That in itself presents a problem if your plan was to use the keyboard as a paralle port-interfaced one because those usually don't have any handshaking, which this situation absolutely demands.

It's still a good and practical idea, though!

For a display, however, you can't beat the IBM monochrome circuit or a similar interface, e.g. one of the Hercules knock-offs. If you simply build your hardware on an ISA bus, perhaps "improved" with negative rather than positive interrupts, you'll find that to be a really decent system bus. It could be simplified a mite, but with lots of backplanes available either as you would cut from a motherboard, with your bandsaw, say, or as one of those adapters allowing the bus to be oriented at right angles to the motherboard, which you'll find in many DELL, Packard Bell, or many other thrift-store specials, it's quite realistic to consider using that as a backplane. You can build your own hardware, or you can use the old 8-bit ISA stuff you find lying about. You can even make changes as the situation demands, replacing old PC stuff with your own. It would be silly to overlook this opportunity to recycle those old video cards (and the $5 monitors at the thrift stores) and the old FDC's and HDC's nobody wants any more.

As for schematic diagrams, I'd still recommend the use of the freebie EAGLE v3.55 available at www.cadsoft.de. The autorouter limits you to a board size of about one playing card, but the schematics can be as big as you like. Think about it! No scanning, no translating formats . . . just use the free software to draw and view schematics.

Uli