Project Logs

Collapse

• The kit is complete and working

  senily64dx • 09/22/2026 at 21:44 • 0 comments

  That took a little longer than it should, and it's all due to my stupidity. I managed to forget to hook up the Output Enable from the 2111 to the 6116 RAM chip and in the meantime, tried different addressing schemata which didn't work and I ended up ANDing the outputs from the RAM mux and since that's working and I'm also using all the chips from the box (except for the blank 2708) which will probably end up going in an arcade machine where a single 2708 completely died, rendering the screen black&white except for the sprites. Another part sacrificed to fix an arcade machine.
  
  So it now has 2k of RAM. Due to my addressing issues, it's in two disjoint areas, one at 1000-13FF and the other at 3000-33FF.
  
  The Siemens made components are all a bit slow. The UART is only stable to 4800 baud and can transmit but not receive at 9600. At 19200 it's completely dead.
  There's also three 74161 - the first one in the timing chain is made by AMD because the Siemens parts are just too slow. At least they run lukewarm - the AMD one gets almost too hot to touch.
  
  By the way, this kit is actually a licensed version (I hope) of the Intel MCS-80 aka SDK-80 dev kit. The Intel one was colored in Intel's corporate light blue. As you can tell, the Siemens one is orange-red. Siemens' capacitors from that era are also orange-red.
  
  The dumb terminal's screen shot shows the dump of the monitor ROM. I did dump it using a small contraption I made for reading two 2708 as one 2716. They won't program of course and if you tried, you'll probably fry the programmer. Also sorry about the dumb terminal. As the owner of a small computer museum, it's embarrassing that my only terminal is this modern (very early 90s vintage).
  
  So next I'll edit part 1 of the video (from the pre-build phase to the working machine) and publish it on Youtube.
  Part 2 will be about building and programming the EPROM programmer. It'll probably begin with some type-in example programs and testing of the 8255, which I still don't know if it's working (remember it's not dead but it may be borked)

• Assembly complete

  senily64dx • 09/08/2026 at 21:23 • 0 comments

  I've finished assembling the machine according to the build manual which I'll probably have to scan in so anyone can build a copy.

  So far I've tested the power rails. While the 7905 was a bit hot at -5,8V during out-of-circuit testing, it did regulate well enough that I didn't think it would pose any problems. Now in the circuit I get -4,97V which is bang on no worries.

  I also tested the clock generator chip (bottom left corner) and it works. Haven't checked the exact frequencies but the oscillograph shows they're in the right ballpark. The two clock outputs to the CPU have odd duty cycles and 12V levels which I bet is what the 8080 needs (I know the 8080 needs a 12V clock).

  According to the manual, the 8251 UART can only go up to 4800 baud which is kinda lame, but the internet says it's fine up to 19,2kbaud which is what I have my dumb terminal set to anyway. Also the baud rate selector has two unpopulated spots which I bet were intended for 9600 and 19200. Of course I will run it at 4800 to see if it works and then go to 9600 and if that works, try 19200.

  You might have spotted I don't have the RAM or its MUX populated. I will start with known good 2114 which require putting in some jumpers and rewiring the RAM sockets. I might later populate the machine with known bad 2111 to test them - at least one of these passed in the tester but completely wrecked the machine it was in which took me like 6 hours to find.

  You might also have spotted I populated the two optional i8212 buffers/BUS transceivers. I have a bag full of those on hand and they need to be either jumped (for a minimal functionality) or populated (for full functionality). Since I intend to use this system for future stuff requiring more memory, I'll better put them in.

  Now I have another party this weekend so there's not much time for more work... I also have to clean the basement just enough so it's usable for party mode. My parents aren't too happy I converted their party basement into an electronics repair workshop...

  Full test next week.

  I've recorded 3 hours (in SD) of build footage (which I'll fast forward) and a few dozen minutes of other stuff (in HD). There will be (at least) two videos - one with the build process which will be released when the system has achieved the status of booting to the Machine Language Monitor and has at least proven that the basic functionality works (manually blink an LED on the 8255, testing RAM access), and another video when I have at least the EPROM programming functionality working.

• Whoopsie - forgot to scan the PCB

  senily64dx • 09/02/2026 at 21:48 • 0 comments

  Oops - after around 1 hour of populating components (video upcoming, probably next week or so), my brain told me it would have been a good idea to scan the bare PCB... well OK so I added scans of both sides of the PCB with partially populated components, better than nothing.

  The footage I shot so far does have a good look under the red IC socket that's populated in the scan, definitely good enough to see where the traces are going and what the silkscreen says.

View all 3 project logs