The "Gryphon Oh-Thirty" MC68030 board is currently in the prototype stage. Additional production from the files provided here is not recommended at this time. This page will be updated as significant events unfold.
Last update: 28 MAY 2015
ABOUT
The goal of the Gryphon project is to create a 32 bit single board computer that would have been state of the art - in the late '80's or early '90's.
Design goals were to use through hole components and to use "period correct" function specific chips as much as possible, and programmable logic limited to 16V8GALs.
Deviations from this were primarily on memory density, ethernet and video to keep the board size down.
The 68030 was chosen as the CPU for various reasons:
- The 68K/CPU32 instruction set is generally acknowledged as one of the better ISAs
- A lot of folks have a fondness for the 68K - the heart of Amigas, Ataris, Macs and others, and a time when we cut our teeth learning computers
- The '030 is the last of the Motorola line that is relatively easy to work with
- Clock speed of 25MHz is probably the limit of 2 sided PCB designs before entering RF voodoo land
Other design decisions:
- Why is the video on a daughterboard? The S1D13505 chip has 0.5mm spacing. Cost for the whole board goes up to accomodate the one chip. A decision was made to use an adapter board to help control costs.
- Why the S1D13505? The Epson chip has the most thorough documentation of anything I could find and it met the design goal of at least 800x600 color video
- Why a floppy disk interface? A true anachronism. I just thought it would be neat to understand how FDDs work. No other reason.
- Why are you doing this when you could (emulate, buy eval board, hack other 68k systems, etc)? You already know the answer or you wouldn't be on this page. It's a hobby and a learning experience.
- Why is there an image of you on the back of the board? Are you an egomaniac? This started out as a personal project. It was only after I had the gerbers ready to send to production that I decided to share the project. This is my contribution to the free hardware/software community for so much that I have benefited from over the years. Rev1.1 will not have my image.
The Gryphon68030 draws the name from the legendary creature comprised of the parts of many beings, much as the project does from other concepts.
SPECIFICATIONS
- MC68030 @ 25MHz
- MC68882 @ 25MHz
- 32M DRAM via 8422V Controller
- 512K SRAM (to help get it initially running)
- 512K FLASH
- Parallel Port
- 2 Serial Ports
- 8242 Keyboard/Mouse Controller
- AY-3-8910 Sound
- 37C65 FDC
- IDE interface
- RTC 72421 Clock
- RTL8019AS Ethernet
- S1D13505 Video (800x600) - (In a TQFP-128 to PGA-128 SMT Adapter)
Bold items have been verified operational or basic testing performed as of latest update
Datasheets for all of the components are found here.
PRIMARY DESIGNERS
Image: /f/1427063554/nm156.png Paul D. Fincato - NM156
Image: /f/1427063371/yoda.png David Mehaffy - Yoda
Please use the N8VEM Forum to contact us.
DESIGN INFORMATION
The Gryphon68030 was designed using Proteus 8.0 by Labcenter Electronics . The native design files are here.
If someone wants to take on the task of converting this project to KiCAD that would be great! I have no experience with KiCAD.
The PCB dimensions are 8.5x11 inches and is 2 layer. xuriPCB manufactured the prototype boards
The first boards were delivered to the initial builders in late October 2014. This wiki documents the progress and changes as the project develops.
REVISION 1.0 (Prototype boards - 5 produced in first run)
DESIGN FILES AS MANUFACTURED REV 1.0 Updated schematics in errata section below
SCHEMATICS - Gryphon68030-Schematic.pdf (Rev 1.0 per first run) See updated Rev 1.1 schematics below
PCB - Gryphon68030-PCB.pdf (Rev 1.0 per first run)
BOM - Gryphon 68030-BOM.pdf (Rev 1.0 per first run)
UPDATES and ERRATAS to REV 1.0 BOARDS
(Mandatory ECOs)
Please look here for the current list of erratas, documentation and updated GAL equations.
Here is the Rev 1.1 schematic with all erratas incorporated to date. This will become the baseline for future board production.
Once this design has been thoroughly checked out there will be a new run of boards incorporating all of the changes. There is no current ETA for this, we will notify
the N8VEM forum when the design has been validated.
If you are interested in the next spin of the Gryphon boards, please take some time to download/review the schematic and provide us your thoughts and opinions.
Use the N8VEM Forum topic "An MC68030 SBC Design" to report any issues that you may find.
CURRENT PROGRESS and UPDATES
28 MAY 2015
- The DP8422 DRAM controller is now functional. Preliminary byte, word and long read/writes are successful. Minor changes to byte decode, DRAM glue GAL equations and a trace swap required..
- Exhaustive DRAM testing will begin soon for entire 32M address space
- There is still an issue with 8422 CS programming that fails to initialize properly after reset on occasion. Investigation into this is ongoing.
28 APRIL 2015
- With a few minor erratas applied the AY-3-8910 and VT82C42 check out
- Minimal code was written it ID the coprocessor installed. It identified the 68882 when installed and crashed F-line (as expected) when removed
- Schematics have been updated with latest errata changes and will be posted soon.
15 APRIL 2015
- A basic monitor was written to view memory, disassemble and download SRECs. Has been burned to FLASH and is now running on hardware.
- DRAM issue has not yet been resolved, but having a basic monitor will help debugging by eliminating the burn/replace FLASH cycle
- Yoda is currently working on unifying the code and working toward a standardized 68k monitor for N8VEM platforms
Image: /f/1429143220/GRYMON3.jpg
27 March 2015
- NM156 has successfully tested the serial port. Yoda reminded me that CTS must be asserted for the DUART to spit characters.
- Work is continuing on getting DRAM functional
22 March 2015
- Yoda has serial and SRAM working on his prototype. DRAM is being worked.
- NM156 has a populated and patched board with no successful testing so far
- The first set of erratas were posted (see section above)
DEVELOPMENT TOOLS
There is a binary GCC68k cross compiler distribution MinGW (Windows) on sourceforge
EASy68K is a very nice free (Windows only) simulator for the 68k. (Has been reported to run on Linux under Wine)
A Windows executable for GALASM is included in the GALS directory. This was compiled from the source on GitHub if you are developing on a platform other than Windows
Terminal program EXTRAPutty
Diagnostic terminal program BrayTerm
COM0COM (Signed 64 bit Windows driver) - Virtual COM ports - great for running a real serial terminal with EASy68k
SOFTWARE RESOURCES
An open source TCPIP stack with RTL8019 support
An open source Compact Flash FAT16/32 library
OS-9/68K
One Man Unix
CP/M-68 Simulator
REFERENCE MATERIALS
MC68030 Users Manual
DTACK Grounded Archive
MC68000 Quick Reference Guide
MC68000 Beginner's ASM guide
Javascript Atari ST Emulator
PART SUPPLIERS
I would suggest that you avoid Futurlec Electronics. I was a previously satisfied customer with many orders over the years, but it has gone downhill fast. On my initial build order for the Gryphon I was shorted components that were about 25% of the total cost of the order. It took several weeks for them to respond to my email, and they promised to make it right. They never did, all I received was lip service. I still don't have the parts and now they do not respond to my emails. I'm giving up and chalking this up to experience. AVOID THIS COMPANY.
Image: /f/1414442834/Gryphon68030Zofz.jpg
Rendering of gerber files using ZofzPCB
Image: /f/1427062577/Proteus-render.png
Rendering of Rev 1.0 using Proteus
NM156 - Checking component placement on Rev 1.0 board (October 2014)
NM156 - TOP mostly populated board (20 MAR 2015)
NM156 - BOTTOM - wire patches per erratas (20 MAR 2015)
(#30 kynar wire wrap with clear fingernail polish as adhesive)
Created from past life to perform
Illicit function, I fail this conscious
Madness I man/machine imperfection
NM156
Comments (0)