This Library Needs To Be At Least… Three Times Bigger

Many of us have noted a tremendous price increase in many computer components for some mysterious reason. Whatever this cause is will be debated among the various modern philosophers and Diogeneses, but regardless of cause we all still have to live in this world and make do. That turns us towards getting maximum value from the things we already have rather than trying to go out and buy more computer components right now, like [svofski] using his vast swath of existing microSD cards to build an SD card library.

The library is based around a tiny robotic arm that can physically grip the cards and move them in and out of a reader. The first iteration of the arm involved rotating the two pincers, but this turned out to be overly complicated and [svofski] eventually settled on a design resembling a rack and pinion that slides the two pincers together instead. With the gripper sorted out, it’s placed in system called T-bot arrangement, similar to coreXY kinematics, that lets it pick and place among 12 microSD card slots.

Many of the parts in this build were directly from or inspired by 3D printers, making it relatively simple with so many parts available. [svofski] didn’t build it for a specific use case, though; mostly it was constructed out of fascination for robotic tape changers which perform a similar function. But for anyone who actually needs to duplicate a large number of SD cards, or other types of removable media, this could prove to be a fairly handy robot.

Training A Transformer With 1970s-era Technology

Although generative language models have found little widespread, profitable adoption outside of putting artists out of work and giving tech companies an easy scapegoat for cutting staff, their their underlying technology remains a fascinating area of study. Stepping back to the more innocent time of the late 2010s, before the cultural backlash, we could examine these models in their early stages. Or, we could see how even older technology processes these types of machine learning algorithms in order to understand more about their fundamentals. [Damien Boureille] has put a 60s-era IBM as well as a PDP-11 to work training a transformer algorithm in order to take a closer look at it.

For such old hardware, the task [Damien Boureille] is training his transformer to do is to reverse a list of digits. This is a trivial problem for something like a Python program but much more difficult for a transformer. The model relies solely on self-attention and a residual connection. To fit within the 32KB memory limit of the PDP-11, it employs fixed-point arithmetic and lookup tables to replace computationally expensive functions. Training is optimized with hand-tuned learning rates and stochastic gradient descent, achieving 100% accuracy in 350 steps. In the real world, this means that he was able to get the training time down from hours or days to around five minutes.

Not only does a project like this help understand these tools, but it also goes a long way towards demonstrating that not every task needs a gigawatt datacenter to be useful. In fact, we’ve seen plenty of large language models and other generative AI running on computers no more powerful than an ESP32 or, if you need slightly more computing power, on consumer-grade PCs with or without GPUs.

Fixing Sony And Philips’ Doomed CD-i Console

Although not intended to be a game console, the CD-i would see a a couple of games released for it that would cement its position in gaming history as the butt of countless jokes, some of which still make Nintendo upset to this day. That aside, it’s still a fascinating glimpse at the CD-based multimedia future envisioned in the early 90s, starting with its release in 1990. Recently [MattKC] decided to purchase another CD-i in a fit of nostalgic rage, and repair it to show the world what the future could have been like.

Although Sony and Philips co-developed the device, Sony would go on to release the PlayStation a few years later, which made the CD-i’s life and expectations for it that much harder, leading to it slowly fading into history. The Magnavox one that [MattKC] got is one of the later models, based on the CD-i 450 that was introduced in 1994 as one of the more gaming-oriented models.

As is typical with older devices that use optical media, it would not read discs. It also would sometimes boot up with a ‘Memory Full’ error. This is a common fault due to the built-in battery having run out, erasing RAM-stored values and causing random glitches like this when garbage values were read in on boot.

Continue reading “Fixing Sony And Philips’ Doomed CD-i Console”

The Many-Sprites Interpretation Of Amiga Mechanics

The invention of sprites triggered a major shift in video game design, enabling games with independent moving objects and richer graphics despite the limitations of early video gaming hardware. As a result, hardware design was specifically built to manipulate sprites, and generally as new generations of hardware were produced the number of sprites a system could produce went up. But [Coding Secrets], who published games for the Commodore Amiga, used an interesting method to get this system to produce far more sprites at a single time than the hardware claimed to support.

This hack is demonstrated with [Coding Secrets]’s first published game on the Amiga, Leander. Normally the Amiga can only display up to eight sprites at once, but there is a coprocessor in the computer that allows for re-drawing sprites in different areas of the screen. It can wait for certain vertical and horizontal line positions and then execute certain instructions. This doesn’t allow unlimited sprites to be displayed, but as long as only eight are displayed on any given line the effect is similar. [Coding Secrets] used this trick to display the information bar with sprites, as well as many backgrounds, all simultaneously with the characters and enemies we’d normally recognize as sprites.

Of course, using built-in hardware to do something the computer was designed to do isn’t necessarily a hack, but it does demonstrate how intimate knowledge of the system could result in a much more in-depth and immersive experience even on hardware that was otherwise limited. It also wasn’t free to use this coprocessor; it stole processing time away from other tasks the game might otherwise have to perform, so it did take finesse as well. We’ve seen similar programming feats in other gaming projects like this one which gets Tetris running with only 1000 lines of code.

Continue reading “The Many-Sprites Interpretation Of Amiga Mechanics”

39C3: Hardware, And The Hard Bit

The 39th annual Chaos Communication Congress (39C3) is underway, and it kicked off with a talk that will resonate deeply with folks in the Hackaday universe. [Kliment] gave an impassioned invitation for everyone to start making hardware based on his experience both in the industry and in giving an intro-to-surface-mount workshop to maybe thousands of hackers over the years.

His main points are that the old “hardware is hard” cliche is overdone. Of course, working on a complicated high-reliability medical device isn’t child’s play, but that’s not where you start off. And getting started in hardware design and hobby-scale manufacture has never been easier or cheaper, and the open-source tooling gives you a foot in the door.

He tells the story of an attendee at a workshop who said “I kept waiting for the hard part to come, but then I was finished.”  Starting off with the right small-scale projects, learning a few techniques, and ramping up skills built on skills is the way to go. ([Kliment] is a big proponent of hand-placed hot-plate reflow soldering, and we concur.)

This is the talk that you want to show to your software friends who are hardware-curious. It’s also a plea for more experimentation, more prototyping, more hacking, and simply more people in the hardware / DIY electronics scene. Here at Hackaday, it’s maybe preaching to the choir, but sometimes it’s just nice to hear saying it all out loud.

Back To The 90s On Real Hardware

As the march of time continues on, it becomes harder and harder to play older video games on hardware. Part of this is because the original hardware itself wears out, but another major factor is that modern operating systems, software, and even modern hardware don’t maintain support for older technology indefinitely. This is why emulation is so popular, but purists that need original hardware often have to go to extremes to scratch their retro gaming itch. This project from [Eivind], for example, is a completely new x86 PC designed for the DOS and early Windows 98 era.

The main problem with running older games on modern hardware is the lack of an ISA bus, which is where the sound cards on PCs from this era were placed. This build uses a Vortex86EX system-on-module, which has a processor running a 32-bit x86 instruction set. Not only does this mean that software built for DOS can run natively on this chip, but it also has this elusive ISA capability. The motherboard uses a Crystal CS4237B chip connected to this bus which perfectly replicates a SoundBlaster card from this era. There are also expansion ports to add other sound cards, including ones with Yamaha OPL chips.

Not only does this build provide a native hardware environment for DOS-era gaming, but it also adds a lot of ports missing from modern machines as well including a serial port. Not everything needs to be original hardware, though; a virtual floppy drive and microSD card reader make it easy to interface minimally with modern computers and transfer files easily. This isn’t the only way to game on new, native hardware, though. Others have done similar things with new computers built for legacy industrial applications as well.

Thanks to [Stephen] for the tip!

Continue reading “Back To The 90s On Real Hardware”

A Non-Sony Playstation Motherboard Replacement

As hardware ages, it becomes harder and harder to keep it in service. Whether that’s because of physical aging or lack of support from the company who built it in the first place, time is not generally good for electronics, especially when it comes to our beloved retro gaming systems. The first Playstation, for example, is starting to see some of the deleterious effects of having originally been built in the 90s, and [LorentioB] has a new, third-party motherboard to bring to the table to keep these systems online as well as adding some features in that Sony removed.

The motherboard is known as the nsOne, meaning Not Sony’s One since this is the first motherboard built by a single person outside of Sony. It’s not based on any FPGAs or emulators and is completely compatible with all of the original hardware, chips, and other circuitry of the original Playstation. Based on the PU-23 series, it even revives the removed parallel port, which Sony removed after the first versions of the hardware because of region locking concerns and other pro-consumer issues. Every chip footprint and connector was reverse engineered manually, using optical sanding, scanning, and net-by-net tracing.

For such a complex piece of hardware this is quite the feat, and for anyone who wants to restore old hardware or add the parallel port back on to their system this could be a game changer. [LorentioB] is not quite finished yet but hopes to have a finished version shortly. As far as fully opening up the system goes, there are some software hacks to look at that allow more games to run on the system and some hardware hacks that open the system up as well.