Hello All!

I'm re-designing my SBC to include more I/O and would like to invite those of you who are interested to look over my design.

I'm about 90% complete on the circuit board layout.

So, take a look at the "SBC-2 Prototype" tab and let me know what you think!

Thanks!

Daryl

_________________
Please visit my website -> https://sbc.rictor.org/

It's nice to see your project coming along. I can't claim any recent progress on my own next workbench computer for the last couple of months.

At this late stage of the game, it's understandable that you wouldn't want to make any major changes, but I might make some suggestions.

On the layout, you might want to make pins 1 and 14 of the VIA headers to be ground, and 2 and 13 to be Vcc so that if you plug in a ribbon cable backwards, at least power and ground orientation are still correct and nothing gets damaged from power being applied backwards.

For all ACIAs, and all VIAs except WDC's, you need the IRQ and NMI to have a pull-up resistor of around 3.3K. The line may float up without it, but it will take so long that the line will still be in a logic-low state when you come out of an ISR, and the processor will think it already has yet another interrupt to process. I learned that one the hard way (like so many things). In my case, I even had the pull-up, but needed to clear the VIA's interrupt sooner in the ISR so the line would have time to come up to a solid logic-high level before the RTI instruction.

A lot of RS-232 products do use the hardware handshaking; so since you have another line driver and receiver in the MAX232, it would be good to connect them to the ACIA's RTS and CTS. A plug-in jumper on a pin header could hold CTS true if it doesn't get connected through the interface cable. (The ACIA transmitter is disabled anytime CTS goes false.)

The ACIA can be run with a crystal and couple of 22pF capacitors on it instead of using the 1.8432MHz can oscillator. Then you could use the freed-up room for a schmitt trigger gate for the reset input. The RC works on the uP itself because it has a schmitt-tigger RST input; but the peripheral ICs don't, and you can get some strange effects as the voltage across the capacitor slowly wanders through no-man's land between logic states. If you leave it the way it is, make sure the reset routine thoroughly sets up the VIAs and ACIA, after the capacitor voltage has had plenty of time to come up, but before interrupts are enabled. Actually I did get bit by this on the uP itself once too, because there was enough electrical noise picked up on the line between the capacitor and the uP's RST input pin to break out of the tiny hysteresis area of the RST input. Without re-doing the board, our solution was to scab a .1uF capacitor on the back of the board from pin 1 to pin 40 of the uP, with the leads as short as possible.

Since you're getting boards made by the board house, will you be offering them to sell to others here on the forum too? I'm sure some of us would be interested.

Garth

_________________
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?

Garth,

Thanks for the feedback. As this is just a hobby of mine, I am ALWAYS willing to listen to the advise of others.

I know I took several shortcuts from "text-book" design, both in the no-frills reset circuit and the RS-232 connection. I had included a pull-up resistor (SIP Package) for the RDY, RESET, IRQ, & NMI pins. Would it be safe to use the 2.2k resistors accross these for pull-ups?

I'll take a look at your other suggestions and try to incorporate what I can. I am already thinking of a daughter board for the expansion port that would include a 2x16 LCD, 16 button keypad, and perhaps a real-time clock chip of some sorts. I've been looking over Chris Ward's DIY design and think these might be desireable as well.

And yes, having ExpressPCB manufacture several boards at once make the cost cheaper so I'd be glad to invite others to join me in a group order, when the time comes.

Thanks again for the advice! Anyone else want to add anything???

Daryl

_________________
Please visit my website -> https://sbc.rictor.org/

> I know I took several shortcuts from "text-book" design, both in
> the no-frills reset circuit and the RS-232 connection. I had
> included a pull-up resistor (SIP Package) for the RDY, RESET, IRQ,
> & NMI pins.

I guess I looked too fast. I missed that. My apologies.

> Would it be safe to use the 2.2k resistors accross these for
> pull-ups?

According to the spec.s, 2.2K is too heavy; but I recently tried the port pins on a Rockwell VIA and found, as I expected, that their true current-handling capability is waaaaay beyond what's specified. They could pull down to 0.4V with a 220-ohm resistor to +5V, meaning they could achieve a guaranteed logic-low at 21mA. That's more than ten times what's specified. A direct short to Vcc gave anywhere from 90 to 110 mA. They could also pull up to 2.4V with a minimum of 10mA (connecting to a 220-ohm resistor to ground). I didn't try the IRQ output though. I want to try WDC's VIA next.

> I am already thinking of a daughter board for the expansion port
> that would include a 2x16 LCD,

I've used these intelligent character LCD modules many times. Let me know if you have any trouble getting it going. There's kind of a trick to getting them to initialize dependably. After a lot of effort, we finally got some special info from a Hitachi applications engineer that wasn't in the LCD data books. After that we never had any trouble. If you're hard up for I/O bits, you can take care of the LCD with only 6 lines: ground the R/W\, and use 4 data lines, 1 RS, and 1 E. On the other hand, since you're only running at 1MHz, you might get away with putting the LCD directly on the data bus without going through a VIA.

> 16 button keypad,

If you don't need to press more than one button at a time, you can use some of the same lines you used to interface the LCD. The uP can feed the LCD and watch the keypad at the same time if, for example, you output 4 bits to the keypad rows connected to the LCD data lines, and have the keypad columns be inputs to 4 other VIA bits.

If you had to get really fanatic about saving I/O bits, you could have the 4 LCD data bits also go to a 74HC154, and each of its 16 outputs could go to one key and a series diode. The other side of all 16 key switches go to a single VIA input with a pullup resistor. If the input bit is low, the key being pressed is the one whose 4-bit code is being fed to the 74HC154's inputs. This takes a little more software overhead and an extra IC, but then you can add the keypad with only one extra I/O bit, and press more than one key at a time without fowling up the LCD.

Another way would be to use the VIA's synchronous serial port... Woops, I think I'm getting carried away. This wasn't supposed to become "1001 Ways to Interface Everything."

> and perhaps a real-time clock chip of some sorts.

The VIA's T1, automatically generating a continuous stream of evenly spaced interrupts, can be used for a real-time clock with no extra hardware. Of course it all goes away if you power down. You might want to have that VIA on NMI instead of IRQ. I've used the Saronix RTC58321 clock and calendar IC. It's nice, but slow. Even at 1MHz, you can't connect it directly to the uP bus. Using the VIA's T1 with interrupts will let you time things with millisecond resolution or better, unlike the RTC IC.

Garth

_________________
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?

Hi again!

Garth, I read all your suggestions and incorporated most of them. My design is now complete. I removed the 6551 TTL osc and replaced it with a crystal and cap. I added pads for a DS1813 reset device, and added pads to strap about any combination of handshaking for the serial port, including from 6551 to Max232 and also from max232 to the serial port connector. I decided to save space by replacing the 9 pin sub-d socket with a 10 pin header. Most PC parts stores sell the 9 pin and 25 pin pigtail cables that connect to the PC motherboards and one of these should work here.

I have uploaded new schematics and board pictures to my web site. I am still verifying the traces but thought more eyes would be better here. Everyone, please look them over and let me know if you see any mistakes.

This time I saved the pictures as .gif files, reducing their sizes from 200k to 25k so they should load faster.

As far as the daughter board, I'll save that for later. You are right, there are many ways to overlap IO port for multiple applications, such as LCD and keypad on one 8 bit port.

I will start collecting names of those who may be interested in making a bulk purchase of this board and if we can get enough interest, we should be able to get the cost down to about $25-$30 per board. Right now, the minimum order is 2 boards for about $115. Ten boards run about $260 ($26 each). Twenty would drop the price to $440 ($22 each). Not a bad price for a quality produced board.

I am also looking for suggestions for purchasing the 65CXX chips. Would anyone be interested in getting a bulk order together to build these boards with? Any suggestions on supply sources would be appreciated as so far I've used recycled parts for all my projects.

Take care!

Daryl 65c02@softcom.net

_________________
Please visit my website -> https://sbc.rictor.org/

If you can keep them close to $25/ea., put me down for a couple. I'll probably use at least one them to help our son get into home-made computers without having to wire-wrap another one. He's done some C64 assembly programming, but this will afford a different kind of work, with fuller control over the entire project.

I don't think I need any more parts myself, and the first quantity price break for most places that will sell one part at a time may not be worth the double shipping unless you send the parts out with the board itself.

WDC does have a $100 minimum order, but their VIAs have the faster totem-pole outputs on the IRQ line that can't be wire-ORed like older VIAs and like your board does. Since you haven't gotten the board built yet, maybe it would be good to add a series diode in the IRQ line of each VIA. (I should have brought this up earlier.) I expect the VIA could still pull the voltage low enough that the processor would recognize a valid low logic state and respond to the interrupt. If silicon diodes have too much voltage drop, there should be a small shotky diode that would do the job. Worst come to worst, the diode space on the board could be jumpered and you use non-WDC VIAs. In other words, you'll probably gain something but you won't lose anything by adding the diode spaces.

California Micro Devices' (CMD's) parts are not as fast as WDC's (only spec'ed to 6MHz). Their distributors are:

Arrow Bell Components www.arrow.com
El Segundo, CA branch: (310)563-2348
Jaco Electronics www.jacoelectronics.com
Hauppauge, NY (800)541-9371

These are industrial distributors. Such distributors have line cards, but usually no catalog. They don't usually mind selling to hobbyists, but they usually have minimum orders (like $100) and you should be prepared with exact manufacturer's numbers you want and acceptable substitutes (like a wider temperature range or faster speed) if your first choice would take too long to get. CMD still makes the 65c51, but I don't believe WDC ever did. For my own future projects, I may just use UARTs with I2C interfaces so the processor can communicate with them without slowing down to 4 or 6MHz. I have 100 1MHz 6512 processors here. I'd like to keep a few, but I might be willing to let most of them go for a reasonable price. The 6512 is like an NMOS 6502 without an on-board clock generator. Boards can be made with a couple of jumper options to accomodate the 6502 or 6512 either one.

To again mention a couple of hobbyists' sources of 65xx parts:
Debco Electronics www.debco.com/db_online/micro.htm#6000
Jameco Electronics www.jameco.com

Happy building, happy programming, and all that good stuff.

Garth

_________________
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?

I've finished verifying the traces on the PCB (found 1 mistake) and have added pads for diodes on the interrupts of all three peripheral IC's. I'll have the final board layout pictures up soon.

In order to get $25 per board, we need to order at least 10. So far, 4 have been requested.

Thanks for the sources.

This should be a useful board with lots of potential for expansion.

Daryl

_________________
Please visit my website -> https://sbc.rictor.org/