SBC-85 8085 Single Board Computer

The SBC-85 is an open source 8085 based, expandable system with single board processor, backplane, and a growing number of support cards.

Similar projects worth following
Starting from
ccandrews has 85 orders / 1reviews
Ships from United States of America
Decades ago I wirewrapped my first 8085 3-chip system with an 8085, 8155, 8755 and I was hooked forever. Sure there are better processors, but arguably none held onto their (mostly industrial) market share for so long. I decided to rekindle my fondness for the 8085 by updating and creating an 8085 single board computer with an eye towards expansion capabilities. The end result is the SBC-85 System which includes a card edge connector and backplane to create a 70s-80s style system akin to the MIL-MOD8 or STD bus.
The 8085 SBC has an RS232 serial port (hanging on the SID - SOD), 22 pins of I/O, timer (both courtesy of the Intel 8155), 8kB RAM, and two 2732 EPROMs one on the board and one in a ZIF.

The SBC-85 CPU board is a 100mm x 100mm 2-layer with a 120 pin edge connector (mechanically PCI).   The serial SID and SOD lines, via a max232, offer a RS232 serial port for communication with a resident monitor or whatever you may decide to put in firmware.  The 120 expansion bus includes power, all of the 8085 status and system signals,  latched address bus, and many more unused terminals for future or other board-to-board signals.

RAM is provided by a 64-kBit 6264 and an additional 2kbit in the 8155.  In addition to the additional RAM, the 8155 provides an onboard timer and 22 pins of I/O.  Finally, two 2732s provide onboard EPROM, one socketed on the board and the second in a ZIF for easy swap.

Power is provided by an external regulated 5V power supply through either the barrel connector or the backplane.

Adobe Portable Document Format - 6.36 MB - 03/09/2023 at 00:32


  • SBC-85 CPU v2.0 with Universal EPROM sockets

    Craig06/03/2020 at 00:07 0 comments

    Those that do not have 21V programmers for the 2732A, or those running CP/M and other high memory requirements can rejoice. Version 2.0 of the SBC-85 CPU board will soon be completed and tested.

    When I first did the layout on v1.0, I did not think there would be room for any EPROM larger than the 24 pin 2732 and certainly not room to put the expansion EPROM into a ZIF socket. However, I optimistically created the schematic with 28-pin universal EPROM sockets, added the requisite jumpers, and began the long process of getting it to fit in the same 100mm x 100mm footprint. Quite to my surprise, I was able to get dual 28 pin EPROM sockets and still have room for the ZIF on the expansion EPROM. The routing had to be restarted several times and in some places the trace spacing had to decrease but the minimum trace width is still 10mil.  With the narrower gap, however, this means that the board will be somewhat more expensive to manufacture at any of the houses that charge extra for 5 mil spacing.

    As can be seen from the image, the general placement of all components has remained the same but the 24-pin EPROM sockets are now 28-pin universal EPROM sockets and two sets of jumpers for device selection have been added in the top left portion of the board. While both EPROMs have to be the same size, this board accepts a pair of 2732s, 2764s, or 27128s or the board can accept a single 27256 in the ZIF socket. All other features of the board remain pretty much unchanged, with the 8155 and its 22-pins of I/O, the MAX232 serial port, and the full backplane expansion. Version 2.0 also brings RESET and RESET* to the backplane which was an oversight on the first versions.

    In addition to the difficulty of simply getting this all to fit on a 100mm x 100mm 2-layer board, the new memory flexibility also created address decoding headaches. It was important to keep the address decoding straightforward yet avoid holes in the memory space and to avoid stepping up to programmable logic device. In the end, I was able to make the EPROM address space contiguous for all possible configurations, move the RAM up to a higher address, and create the hole in between the EPROMs and RAM.  In version 2.0 the EPROM memory space still begins at 0x0000 and extends up to their maximum combined address, but the RAM has been moved from 0x2000 up to 0x8000. If using a total of 256KBit of EPROM the top of ROM still bumps up against the bottom of the RAM. If using less than the maximum EPROM on this board, then there is a memory hole but rather than scattered memory it is a single hole that can easily be filled by the memory expansion board as RAM or ROM. To be more specific, in version 2.0 the 6264 RAM is now from 0x8000 to 0x9FFF, the 8155 is enabled from 0xA000 to 0xA7FF, and the 8155 ports begin at I/O 0xA0.

    With the 32KB of EPROM available on the CPU board plus [any ROM / RAM combination up to 128KB ROM or 32KB RAM] available on a memory expansion board, the SBC-85 should satisfy even the most hard-core memory needs. Add to that a 1Mbit bubble board and you have a pretty nice system on your hands.

    I am well into creating the documentation and will upload as it becomes available in draft form.  With any luck, v2.0 will be tested and a few prototypes available in June with the final version and the build files released in July.

  • 2764 EPROM Version Anyone?

    Craig03/28/2020 at 03:10 2 comments

    I have half a dozen things already in the queue, but in my next PCB run I am thinking of adding a single 2764 EPROM version of the SBC-85 SBC.   This would avoid the 21V 2732A EPROMs while not changing the overall memory space of the SBC-85.  

    Any thoughts or suggestions?

View all 2 project logs

Enjoy this project?



Craig wrote 03/18/2020 at 05:58 point

Rather than think of it as a confusing time, think of the tremendous progress made with each series between 1972 and 1977.   For a point of reference, look at a 1702 programmer like on the MIL MOD8 or the intel MP7 for the SIM8, now those were some serious programmers. The 1702 and 1702A both required programming with data inversion and multiple bipolar programming voltages totaling over 65 volts but the 1702A already programmed several times faster than the 1702.  You would be amazed at how hot a 1702 gets when being programmed!  By the time the 2708, 2716 and 2732 came around the programming voltage was already down to 25V and unipolar. Just from the 2732 to the 2732A there was a further dropped down to 21v for the 2732A and 2764.  This all happened over a span of 5 years or so and while the device capacity was also going from 2kbits to 64kbits.    When I pick up a 1702, 2708, 2716 or 2732A, it is hard for me not to think of the amazing history behind that chip.

  Are you sure? yes | no

Ken Yap wrote 03/18/2020 at 07:13 point

Yeah I've dealt with an old ROMs:

However I'm not a masochist as I have loads of (relatively) newer EPROMs and see no need to pay more money and get another burner just to use the handful of obsolete chips in my junk^wspare box. Pity the 2732s are not worth more or I could sell mine. 😉

  Are you sure? yes | no

Ken Yap wrote 03/18/2020 at 04:37 point

Using 64s and above also avoids this confusing period in EPROM history:

  Are you sure? yes | no

Craig wrote 03/18/2020 at 01:08 point

Thanks for the feedback. 

2732s run from 50 cents to $1 each, but true that most are 21V program.

For 21V the $40 Willem PCB6.0E programmer is hard to beat.  I thought about putting on a multi-purpose socket, maybe the next version.  Certainly no room for a larger ZIF but possible for the base EPROM.   The intent was to keep it to 100mm x 100mm since those are the 5/$2 from the PCB house. On the next version I may make the base socket universal, we will see.

I also have an add-on memory board in the works that has multi-size EPROM sockets and more RAM for those that feel the need to run CP/M.  Any suggestions for that board would be appreciated.

  Are you sure? yes | no

Ken Yap wrote 03/18/2020 at 00:28 point

This looks great as I have quite a few chips in this MCU family. But the 2732 is a bit unfortunate as it's harder to get and often requires higher programming voltages. Is there no way to accommodate a 28 pin socket for a 64/128/256/512 EPROM?

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates