Minimal 64x4 Home Computer
Comments
ge96
cdumler
> It is like witchcraft seeing someone produce VGA out with some raw/more tangible chips.
I think what you're missing is: VGA was designed in the era when this was A Thing™. Monochrome/NTSC/CGA/EGA/VGA displays are all about "bit banging," sending signals at the right time. If you can send 1's and 0's faster than the analog reception can update, you can "fake" voltage potentials. I say "fake" because that was actually a way to do it before digital-to-analog converters were easy to implement. Today, we can easily produce chips custom for the purpose; however, "in the beginning" it was really just all about timing.
The witchcraft for me was the fact that while older cards used bit-banging to get signals out the door, it was generally designed with a specific purpose (thus specific timings). If you can get access to the underlying timing control, it [opens a whole new world that will surprise people today](https://www.youtube.com/watch?v=-xJZ9I4iqg8).
PaulHoule
Display controllers from the 8-bit era were simple conceptually but had a huge parts count, particularly it needs to have memory access logic very similar to what is in the microprocessor. The earliest home computers (TRS-80 Model I, Apple II) had a large parts count which was reduced in the next generation (TRS-80 Color Computer, VIC-20) because the glue logic and display controllers got the same LSI [1] treatment as the CPU.
People who build modern real-hardware fantasy computers [2] struggle with the cost of the display controller if it is done in an authentic style so they wind up using an FPGA or microcontroller (amazingly easy to do with ESP32 [3])
This thing addresses the problem by reusing many of the parts between the CPU and display controller, plus the contrast is not so stark since the CPU part count is greater than 1, unlike the typical retrocomputer.
It's fascinating! It's a minicomputer in the sense that it is built out of low-integration parts, but it is like a microcomputer in important ways, particularly having the closely integrated display controller.
[1] https://vaibhav-pawale19.medium.com/integrated-circuits-ssi-...
Lerc
Historically, the best example of going it alone without display hardware, was the ZX80.
https://www.8bity.cz/files/zx80_schema.pdf
VIDEO comes out of the block of gates roughly center of the schematic.
SYNC comes out of the gates on the center right.
and of course the ZX81 merged the logic chips into one ULA.
There exists a kit to go the other way for a ZX Spectrum to split the ULA logic out into discrete chips again.
ge96
That demo looked cool, would probably be more impressive if I grew up with that kind of tech.
I wonder wrt an HDMI to VGA connector how simple it is, direct connections or some chip in the middle.
Yeah for end of days being able to go back to VGA/those analog plugs would be good eg. Red/White/Yellow
Lerc
You need a chip for VGA->HDMI but they exist, and you can buy simple adapters. I think HDMI->VGA adapters might be cheaper (I have one in a draw somewhere) , One of the more tricky points with HDMI is that they are stricter on what they call a valid image and make weird assumptions like All your pixels are the same width.
A CRT can make do with signals to say "go to the next line now", "go back to the top now". and then just output whatever is coming in on the colour signal. It really means there is no concept of a display mode. It's all just in the timing of the signals on the wires. Plenty of modern hardware with digital internals look at a lot of that and just say "That's not normal so I quit".
Analog devices may make a high pitched whine and then explode, but at least they'll attempt the task they have been given.
Koshkin
> about "civilization ending" events this kind of compute
Realistically though, we'd be extremely lucky to need the abacus, even.
bloopernova
I've sometimes wondered what it would take to bootstrap up to the 555 chip or similar. Is it within the reach of most countries with universities?
cancerhacker
Not sure if it’s still being sold, but “evil mad scientist” had a kit [1] to build a giant 555 from discrete components.
askvictor
If you haven't read "A Canticle for Leibowitz" I highly recommend it for a perspective of technology after a nuclear holocaust.
0xEF
I found Canticle dull at first, then the ideas started to gel and I was salivating for more post-apocalyptic speculation like it. Do you have any other recommendations for people who enjoyed it?
ge96
Saved, I just started the 5th Bobiverse book since that "recently" came out
anonzzzies
> I am in no way associated with any activies selling the Minimal 64x4 as a product. Any such activity represents a license violation. Individual licenses apply for hardware and software parts.
Maybe I should read everything, but what does this mean?
komadori
At least part of it is licensed under a Creative Commons Non-Commercial licence.
snvzz
It means it is not Open Source Hardware.
It also means it is likely too tainted to touch even with a 10m pole.
MarkusWandel
I don't quite get the complexity/performance graph's X axis.
Around the time of the early microprocessors in the 1970s, vastly greater performance was obtainable via discrete parts, for the simple reason that more parallel stuff could be built than the LSI of the day could accommodate (e.g. 32-bit datapath). He even has a specific X axis point with the MSI 74181 bit slice ALU chip - which was in all the high-performance discrete part CPUs of the day, either that one or the AM2901 which also contained registers.
So X axis should be in net complexity in transistors or gates, regardless of implementation. That being said, the device presented still gets an enormous amount of compute per TTL chip, especially considering no dedicated ALU is involved.
nottorp
> Monochrome VGA output 400 x 240 pixels
Seen the VGA output on a bunch of retro projects.
I've also seen that there are VGA to HDMI converters available.
But can those do any random non standard res like this one, or are most (or all) of them limited to standard historical resolutions?
jdboyd
There are two aspects to consider. They are can a VGA->HDMI adapter support 400 pixels wide, and can it support 240 lines.
VGA has no pixel clock. A line is a set of continuous waveforms for R, G, and B. This means that there is no problem with a mis-matched horizontal resolution, except that the edges may be ugly.
If a VGA->HDMI convert supports 240 lines is a bigger question, and the answer is less certain. Some might not support it. Vintage VGA cards could be convinced to do all sort of weird line rates, much stranger than 240, so a good VGA->HDMI adapter should support it happily.
aidenn0
Adapters specifically for 8-bit and 16-bit consoles will probably do 240 lines, since 240p was what most of those used.
The OSSC can probably do it. I don't think any of the retrotink boxes support separate sync (which is what VGA provides), which is a shame because those are a great bang-for-the-buck.
indrora
The Rt5x can handle various sync pulses (Sync On Green, Sync On Luma, SCART sync) and the RT4K can handle "Just about anything"
aidenn0
Can it handle separate hsync and vsync signals though? That's what VGA outputs.
ofrzeta
> Vintage VGA cards could be convinced to do all sort of weird line rates
Yeah, with XFree86 Modelines :-)
komadori
A DVI* signal is logically very similar to a VGA one, but with a different physical layer. The analogue colour signals are replaced with TMDS encoded digital ones, but the pixel clock and sync signal work more or less the same way.
I would guess that a simple VGA to DVI converter simply syncs to the VGA pixel clock, samples the analogue colours, and outputs the digitally encoded values with the same timings.
From a quick look, the oscillator in this machine's schematic runs at 16 MHz. I assume that the pixel clock is derived from this. The DVI specification has a minimum pixel clock of 25 MHz so you couldn't produce a valid DVI stream from this without buffering the pixels and retiming the output in summer way. Well, I suppose since the pixel clock isn't explicit on the VGA cable you could have an imaginary clock which is higher by doubling pixels horizontally.
Ultimately though, success probably varies depending on the converter and the display used. There are quite a lot of standard VESA modes and you can often get away with generating something close-ish to spec.
For more exotic video signals you can use devices like the RGBtoHDMI: https://github.com/hoglet67/RGBtoHDMI
It decodes the input signal into a framebuffer and uses the Raspberry Pi's video core to output the result.
* HDMI is, broadly speaking, a proprietary extension of DVI. You can feed DVI signals though an HDMI connector and it will display anyway.
nottorp
> HDMI is, broadly speaking, a proprietary extension of DVI
It is but it's becoming hard to find monitors with dvi connectors, so i asked about hdmi which should be more common.
How you get your ancient analog output to a display with only digital in is becoming a problem. I don't know shit about how good or bad your average solution is so I ask.
[I don't have anything like the toy we're talking about in this thread, but I have a 486 with a Trident 512kb vga card and Syndicate on it in a closet]
When I think about "civilization ending" events this kind of compute comes up since it seems more simple/more robust granted it can do much less than a modern computer. I also have heard of the argument at that time it doesn't matter, you can die from sepsis or whatever vs. trying to run a computer.
It is like witchcraft seeing someone produce VGA out with some raw/more tangible chips.
I also think about how like on an Arduino you can have a filesystem, make a makeshift notepad via printing/serial in.
Idk... same with RISC V, if a computer works to me do I care what's under the hood. Anyway I did take a course where you learned about flip-flops/registeres, design your own 8-bit ALU, etc... I can appreciate it again "tangible" vs. a single IC that does everything. It's like those cheap games where they have a single chip under a black blob.