A pic of boards for the FAL65C816 computer I've been working on.
Plan is to run the '816 at 1MHz to begin with, 8MHz later if things work.
There's a CPU board (the smallest board as it just contains the cpu and some buffers). A ram/rom memory board, a video board, and an audio board. There's also a backplane board.

_________________
http://www.finitron.ca

Thx for the pics, Rob. It's nice to see a new '816 project in the works.

The memory board is the one on the top left, correct? I think you mentioned somewhere that it'll accept one of Garth's memory modules.

Looking forward to further news and comments, particularly regarding the audio & video...

Jeff

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

Yes I am curious about the video and audio too - are they 80's style parts/designs or using some modern solution like FPGA? I use a TMS9918a and AY-3-8910 for video and sound, which is basically what was used in the MSX-1 series of machines, although MSX main CPU was the pesky Z80 of course..

Yes, the memory board is top left, and has a spot for Garth's ram module. It also has room for an EPROM. EPROM and ram occupy the same memory space until the rom is switched to a higher address (ram outputs are disabled until after the switch takes place).

The video and audio are '80's style. The system is a real mash of vintage and modern parts. I'm using parts from a cannibalized C64, the VIC-II chip for video and a SID chip for audio. There may be trouble driving the VIC-II chip with an 8.192 MHz clock instead of an 8.182 MHz clock. It's only 0.1% different but the 8.18MHz clock had a fixed phase relationship with the color burst clock and 8.192MHz doesn't. I'm guessing the VIC chip doesn't care much about the relationship to the dot clock.

Another thing I worry about working is that the cpu is powered by a 3.4 to 3.6v supply and I have the address lines driving 5v logic directly. The data-bus goes through a level shifter though. Since the address is CMOS outputs it should be over the 2.x V needed for the TTL logic levels.

The I/O is decoded at $FDxxxx, so no compatibility with older systems.

_________________
http://www.finitron.ca

Looks good!

Following this with interest as I'm embarking on a '816 system myself. So-far I've seen 2 other "bus" based systems like yours and an SBC type system. My system is SBC with a VIA controlled bus for peripherals and an ATmega to manage booting, SD card and video initially.

I've not even thought of sound.. Hm.

-Gordon

_________________
--
Gordon Henderson.
See my Ruby 6502 and 65816 SBC projects here: https://projects.drogon.net/ruby/

Rob, you didn't mention why you didn't use 8.182 Mhz, and 5 volts. Some of us will be curious.

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

I don't think there's a such thing as a 8.18181818MHz clock oscillator. The C64 used a custom PLL (8701 chip) to generate the dot clock (8.18...MHz). I did not want to have to re-create the PLL using logic which would be bulky. I found a 32.768MHz oscillator on ebay which is divided by four by a couple of ff's on the backplane. The system may be able to run at half that frequency (16.36MHz) at some point. I should note that the long term plan is to use a 6567 equivalent circuit driving HDMI rather than a vintage part, so I figured the frequency wasn't critical.

Some of the vintage parts are 5v, there's even 12v needed for the VIC / SID. I believe the ram is also 5v. I have the address decoders and buffers supplied with 3.4v however as they are low voltage CMOS surface mount parts. The CMOS parts are 5v tolerant. The SMD parts allowed smaller board sizes, but it means another supply voltage is required.

The 3.4v supply comes from the 5v supply scaled down with two power diodes in series (should drop about 1.6v). This is kinda cheesy but I didn't want to bother with additional regulators.

_________________
http://www.finitron.ca

I'm pretty sure 0.1% difference will be tolerated by most displays, even old TVs. At worst you'll need to adjust the vertical-hold and similar controls, if you have a CRT old enough to still have them. The dot-clock is actually irrelevant to a CRT, only the vertical and horizontal sync matters.

Interesting... are you using DRAM? If not, a Commodore MAX contains the VIC-II 6566 that works with SRAM.

I didn’t think the 0.1% difference would matter much to the display. I’m not sure though how the VIC internally generates phased color output. Does it use pulse shaping, an analog multiplier and a PLL? Or some other means? Some arcade systems use a 14.3131MHz clock and “digital” phase shifting. If it’s some sort of digital shifting colors will likely be off.


The design is using Garth’s static ram. I have level shifting latches (‘573’s) on the outputs of the 6567. There was an additional latched added to supply the high-order address bits during a VIC cycle. The 6567 can be setup for access anywhere in the lowest 4MB of ram. There’s also 2kB of color memory as that chip size was readily available. One gotcha was that there is no char generator rom. Instead the fonts will have to be setup in the ram by the boot program.

_________________
http://www.finitron.ca

Perhaps a schematic would help.

I forgot there was only monochrome output due to board size.

_________________
http://www.finitron.ca