Nevertheless, somebody talked me into writing sort of a resume.
But now back to my hobby projects.
---------------------------------- | "Send me a message in the past | | to prevent this from happening." | ----------------------------------
It all started with the Elektor Junior Computer.
1 kB RAM/EPROM, 7 segment display.
Did meet the CBM 3016 at school.
Had much fun playing with 6502 machine code.
Did build my own keyboard.
Keyboard layout was similar to the CBM 3016,
added a few keys when connecting it to a C64 some years later.
Stopped any attempts of etching PCBs at home after this project...
Bought a C64...
Did solder/wire a 6502 computer, but I wasn't happy with the results.
While improving skills and knowledge, I decided to scrap this ugly
Despite the ugly appearance it nearly worked reliable at 1 MHz.
Got hands on a pile of PCBs, which were part of a 6800 computer.
Plenty of Eurocard sized I/O boards with one 6821 per PCB.
So I did modify the PCBs to work with the 6522 (or 6526).
Then I soldered/wired a 6502 CPU PCB, with 32 kB (GoldCap buffered) RAM,
32 kB EPROM, and a 6845 CRT controller.
Memory was shared between CPU and CRT.
Did decide to modify the C64 Kernal + Basic to run with this machine.
The end result was a computer inside a stable metal box,
with plenty of I/O ports, and a small monochrome monitor.
Problem was, it wasn't very service friendly (built from scrap, that is).
A few years later it died, and I was unable to bring it back to life again.
So it went scrapped.
Did build a tiny 6502 computer inside a metal box,
which once was a tuner from an old TV.
6532, 8 kB RAM, 2* 8 kB EPROM, 7 segment display.
(No, it wasn't compatible to the Junior Computer. In 2008, it still worked.)
Did build a big 6502 computer with 32 kB RAM and 32 kB EPROM.
It did run with a heavily modified C64 Kernal + Basic.
It also did have 512*512 vector graphics (8 colors / 192 kB DRAM),
featuring the EF9367 CRT controller.
It was a single sided PCB layout, done with knive and tape on transparent
paper and such... I wouldn't recommend this approach.
Unfortunately, wiring the outputs of the graphics card directly into
the video amplifiers of an old TV (without separated mains) turned out
to be a bad idea, and at some point the computer died.
CRT controller PCB isn't shown in this picture, it's behind this CPU
PCB and has nearly the same size.
If you wonder about all those modifications on the PCB, it's because I
did upgrade from a 1 MHz 6502 to a 4 MHz 65816, and tried to run
the CPU at 4 MHz.
Did want more computing power (and address range) as with the 6502.
From the addressing modes and the architecture, the 68k family looked interesting.
So I did buid a small 68000 based computer with a 7 segment display,
based on an article from the "Elektronik" magazine.
(In 2008, it still worked.)
Later, I did solder/wire an adapter, so I was able to use that computer
for playing with the 68020.
And the 68020 did offer a (by that time) incredible computing power...
compared with PCs at that time frame.
Unfortunately, the adapter soon started falling apart, due to cheap PCB material
and a homebrew PGA144 socket which was improvised from precision DIP sockets.
Another problem was, that the cheap DIP64 socket inside the computer
didn't make reliable contact to a 68000 IC anymore after a few adapter
Replacing it by a precision socket on a soldered/wired PCB wasn't fun.
That was the moment, when I did decide never to use cheap IC sockets again.
Did start to solder/wire my own 68020 computer.
Nine Eurocard sized PCBs, soldered together to form one big PCB.
At some point it worked, but then there was a short circuit inside
the power supply... which smoked up while sending +30V into my CPU board.
And I wasn't able to bring the computer back to life again,
so it went scrapped.
Note, that the 68020 was a bit expensive in that time frame.
Did some modifications to my C64, and increased the memory size from
64 kB to 5* 64 kB. It was fun, having five programs loaded into memory,
and then being able to switch between them by pressing a key which triggered
At some point the C64 had twice as many ICs inside as intended by the manufacturer.
But then it died, and I wasn't able to bring it back to life again.
So it went scrapped, don't ask for pictures or schematics.
Also I did buy my first PC, an 80286 with 5 MB RAM and a 50 MB harddisk...
After the 68020 disaster, I did decide to play a bit with microcontrollers.
The 8031 family looked interesting for building a homebrew PLC, an idea
which first appeared when building that 6502 computer in a metal box with
plenty of I/O pins...
Also played with the NE5050 and built my own "power line modem",
but using simple amplitude modulation turned out to be a bad idea.
Did solder/wire a 8031 based PCB, which nicely plugged into a PC ISA slot.
There was a monitor program running on the PC (upgraded to 80386),
comunnicating with the 8031 trough four 74LS670 ICs.
Also, I did attach a C64 keyboard and an EF9345 CRT controller to the 8031.
The EF9345 was an interesting IC, it was able to display 40 or 80 characters
per line in color, the character set was a bit like in those "videotext" ICs,
but it had a parallel bus interface (no slow I2C bus or such...)
Writing the software wasn't easy: I didn't have an assembler or compiler
for PC or 8031 at that time... so it was paper, pencil, and debug.exe.
At some point, I wanted to buy some more EF9345, but it turned out
that this IC already was out of production.
My first confrontation with a thing called "product life cycle"...
Did some experiments with 80535 and 80537, like attaching a DRAM
memory module from an 80386 motherboard with only a few 74xx TTL ICs.
Or controlling 24 RC servos with one 80535...
But because of my 6502/68k heritage, programming those microcontrollers
in raw machine code wasn't fun.
At some point, I decided to throw away my 8031 based stuff,
and to make another try with the 68000.
Anyhow, it looks like Infineon later axed the 80535 and 80537.
(Did pull out a few ICs, which went into other experiments...)
Back to the 68000 ! Woot !
Did try to solder/wire board a 68000 computer, which was supposed to be
(more or less) compatible to the NKC... but it refused to work.
After a few months of useless attempts of bringing this contraption alive,
I decided to scrap it.
A few weeks later, it turned out that the backplane was defective.
Avoid buying/using 'cheap' backplanes. If in doubt, build your own...
Anyhow, the CRT controller and the floppy controller used in the NKC
were already out of production.
And now, I wanted to know !
Did solder/wire a 68020/68881 based processor card, (double Eurocard size),
with 32 Bit DataBus and fast SRAM.
It really worked (at 29 MHz it stopped to work reliable), but then I felt a
strange need to buy the full version of the Eagle PCB layout software after
soldering all those wires.
Did solder/wire board some smaller PCBs (including a 16 Bit 68020 processor card).
Also some double Euro sized peripheral stuff... like a PCB with eight 6522,
able to trigger a vectored interrupt.
Also did start to play with CPLDs: the ispLSI1016.
The design software was nice: each logic block was visible on the screen,
you could click one with the mouse and then 'haxxor' your equations into it.
Felt pretty much like having a few GALs tied together inside of one IC
package, and for a year, I was happy...
...and did spend some more time playing with 68k related hardware/software.
Like interfacing to stepper motors...
...and still did play with the 68020.
Did build a tiny monochrome CRT controller with two ispLSI1016 (PLCC44),
and started first experiments with PIC microcontrollers: the 16C84.
Also did try the 68332, but while the old 68020 (manufactured in Scotland)
was a very nice/reliable IC, them 68332s died like flies in my workshop.
It was a very new IC back then, maybe the manufacturer improved quality
since then... but I don't want to know.
Since that time I'm fed up with soldering TQFP to expensive adapter PCBs
and throwing the results into the trashcan every now and then.
Did decide to build an Eurocard sized 68HC000 / 16 MHz single board computer,
with all ICs plugged into in precision DIP sockets.
1 MB SRAM, 1 MB Flash, a 68901, a PIC16C84 to connect a PC keyboard...
and so on.
It was a cute multilayer PCB, and the prototype worked with only a few
Incident ! My supply chain of 68k ICs went disrupted.
Instead of running around, holding up a sign "The end is nigh",
I did decide to go back to the 6502 again, with a vague idea of
"somehow building my own from standard parts" if my supply line
should get disrupted again.
Also had some trouble when the ispLSI1016 was replaced by the ispLSI1016E.
The old software wasn't able to program the new hardware, and the new software
refused to open my old project files.
Can't tell the full story without instantly getting sued into a smoking crater,
but it sure had some effect on enthusiasm for using CPLDs/FPGAs.
After throwing away all my ispLSI1016 related stuff,
I started to wonder what would keep IC manufacturers from playing me such
neat jokes in the future...
Did build a 6502 based single board computer.
Buying UARTs turned out to be difficult, so I did decide to integrate two
PIC16C65 into the design... to work as UART, keyboard interface and such.
Also modified the C64 Kernal + Basic, to make it work with the 6522
instead of the 6526 (which seemed to be out of production).
(Sorry for the bad quality of the picture.)
The 6845 also appeared to be out of production, so I did decide to
build my own CRT controller with 74xx TTL and an EPROM.
Also connected to the internet, which turned out the be a strange new world.
Until that time, I was all on my own for building my hardware/software,
equipped with inapprobiate tools, and with ancient datasheets/articles
printed on paper... the 'lone wolf' approach.
Wouldn't have started a 6502 related project without the existence of 6502.org.
Incident ! My supply chain of 6502 ICs went disrupted.
They disappeared out of all the cataloges here in Germany, and it looked
like there was no distributor left selling 6502 related parts.
Ordering things from "abroad" wasn't an option, and WDC had no minimum
order policy back then.
[Edit: The 6502 showed up at Mouser/Germany in 2012 again.]
So I stumbled into the world of hobby TTL CPU design, what resulted in the
M01 prototype, which still had quite a few serious bugs. It was intended to be
a 180*100mm sized single board CPU. A high density layout...
(Some ICs were salvaged out of the corpse and recycled for building M02.)
While fixing the bugs, I realised that I would like to have a 24 Bit
address bus, and it turned out that it wouldn't have fit into one PCB anymore.
So I did a complete redesign, resulting in the M02 CPU.
Finding most of the bugs inside the microcode took me six months,
and there really were quite a few strange effects.
But in the end, I was able to run 6502 machine code with my first TTL CPU.
Also did build some peripheral PCBs... like a CRT controller,
able to display text and graphics at the same time.
Did write a M02 emulator for PIC16C65, then ported it to PIC18F452.
Unfortunately, it did turn out to be ca. 3 times slower than M02.
M02 had only about the speed of an 1 MHz 6502, but the timing (especially
the bus timing) wasn't compatible.
Spend some thoughts on this, but it became obvious that the M02 CPU core
would have to run at 4 MHz to allow a 6502_like bus timing.
But with an ALU built from EPROMs, The M02 "core" wasn't able to run
much faster than 3 MHz.
So I would have needed a faster ALU, but getting hands on the 74181
turned out to be very difficult.
Did decide to invest some time into research, trying to find out how to
build the best/fastest ALU design possible with standart parts...
Also, I did feel a need to write a few articles about the info I would have
needed for starting a TTL based hobby CPU project.
Stuff, that isn't mentioned/explained in the internet or by school books.
Also, I did stumble over a MC14500 datasheet, and noticed that this
unusual architecture would be a nice playground for testing some concepts
related to ALU design (and to RISC architecture).
Point is, that it has function blocks like a 'real' CPU, but they are
Did notice, that some of the multiplexers I would need to build a fast
74xx ALU (without resorting to the 74181) disappeared out of some of the
So I had some thoughts about building sort of a Bitslice CPU with NAND
gates... one PCB per Bit. But I didn't want to start such a crazy project.
Did some more ALU research, and did build a (more or less) MC14500 compatible
CPU from NAND gates.
Since I figured out in theory, how to do the building blocks for a CPU
with NAND gates, I did feel a need to try building logic gates with transistors.
Due to RoHS, distributors did decide to "phase out" some parts that would
have been useful to TTL CPU hobbyists, but that did not make much money.
Now I wanted to know.
Just imagine: a future proof CPU design, no matter how many accountants
and bueraucrats out there may catch the mad cow disease...
Had some thoughts about building sort of a transistorised MC14500,
an architecture which turned out to be only useful/efficient for
But then I thought that a 16 Bit CPU would be more useful.
Soldering thousands of SMD components wasn't funny.
And until it was completed, I had some doubts that a design with so many
transistors would work at all... the most complicated thing I did build
with transistors before starting the "MT15 project" was a one_Bit binary
counter. Go figure...
Transistorised flipflops built from NAND gates take plenty of PCB space,
so I did try to investigate some alternatives.
Building dynamic (capacitive coupled) flipflops, running fast and reliable
at a 5V supply turned out to be tough, and I did not have much success on that.
Did try building logic gates from small signal FETs like BS170 or BSS138,
unfortunately the capacitance of such transistors still was too high, what
means you would have to burn more energy than with ECL for a design
running at some MHz.
So I started playing with ECL/DECL.
First with BC547 transistors, then BC847, then BFR93A.
DECL flipflops indeed need less components than TTL flipflops,
unfortunately building them turned out to be tough, for more details
read the "DECL test run" section of my homepage.
Also, I had a nice discussion with Tom about what the instruction set
for a minimalistic hobby CPU should look like, if it still should be
(more or less) efficient when running C code.
Found a hobby shop which still had a handfull 54LS181 in the drawer.
TI chips in ceramic package, with readable text on top. Woot !
Had the intention to build another transistorised CPU, but after buying
those ICs I could not resist to try a TTL implementation.
TREX, TTL RISC experiment.
A lumbering mostrosity, supposed to be extinct,
but still able to give people the creeps.
Unfortunately, fixing all the critters in the control circuitry turned
out to be more difficult as I tought, and changing the design parameters
maybe a bit too often during that process also didn't help much.
So there were a few bugs/critters left to play with in 2008.
Did build some peripheral PCBs for my TREX CPU, including a TTL based UART.
------------------------------------------------ | Due to some 'real life problems', | | the 28 years long timeline of hobby projects | | had to end here. | ------------------------------------------------
Burnout, serious health problems, and some real life trouble.
Had no choice but to withdraw from electronics,
and to go into exile, for sorting things out.
Also, I was forced to tear down my CPU workshop.
Was forced to scrap/destroy most of the things mentioned here
in this timeline section... plus a few other things.
...and so on.
I'll spare you some more pictures.
Two cubic meters of electronics went to scrap.
Didn't feel a thing while doing this.
It hit me later.
Another year in exile.
We did a similar cleanup at my workplace.
In in the end, all that stuff is only dead weight,
filtering the dust from out of the air.
On the bright side, maybe we are getting less spam
because of all this... someday.
Started to think about returning from exile.
Two voices are arguing in my head right now:
"Go back into electronics again, and show the world what you can."
"It ain't worth all the trouble, and you just want to be left alone."
The coin is still in the air.
To me, some crackers/spammers took the fun out of building hardware.
Somebody broke into my mailbox again.
This settles things.
I'm fed up with getting stalked by The Enemy.
I hereby resign.
One thing left to to before going back into exile again:
Writing/spelling my curse:
May they get happy with what they achieved so far.
Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.
Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.
Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.
Stranger in a strange land.
when you are fed up with getting hacked,
when you turn your back on technical things,
when you try to enter a non_technical forum...
and when you are instantly getting first mistaken as a bot,
then as 'the evil hacker/cracker'...
because of your "odd writing style".
Didn't want to return.
Didn't want to touch hobby electronics ever again...
despite that I missed it pretty much.
What brought me back was,
that The Enemy showed up less often.
And that the amount of spam around me
was significantly reduced...
And I'm very, very thankful for this.
Because I was out of hobby electronics for nearly 5 years,
I'm now taking things a bit slow,
and rebooting step by step.
So I decided to return to my roots,
and started to build DRC2:
a 6502 single board computer.
In other news, it looks like the PC which contained the files
for my articles was hacked some months ago, so I unpacked/packed
the ZIP archives again by removing *.EXE files and such...
what doesn't improve the chance to anybody
for rebuilding my projects.
But better safe than sorry.
And I had to run all the images for all the articles
through netpbm to make sure that they don't contain
anything harmful for the visitors of my homepage.
As a result, quality of all the images didn't improve,
and it could be that some of the images are not correctly
displayed by some of the web browsers out there.
So please don't blame me...
This also explains, why there won't be any new articles
for a long time.
Again: thanks to everybody who tried to help.
---------------------------------------------- |Maybe I'm just too old fashioned... | | | |But I would like to mention, | |that without a certain sort of hobby projects,| |computers wouldn't exist nowaday... | |or they would look pretty different. | |Messing with such projects | |only brings bad luck... | |for all of us. | ----------------------------------------------
Invader. Invader nearby.
Invader. Invader is nigh.
And again I'm getting hacked/spied,
and again somebody is breaking into my mailbox.
I would say, that the internet nowaday became
a too hostile place for sharing technical knowledge.
And I somehow doubt, that I'll be able
to keep my homepage alive in the long run
because of all this.
I'm very sorry to say this now:
But the only option I have left is deleting my homepage.
I just want to be left alone by those guys, you know...
(c) Dieter Mueller 1980..today