Mysterious Files PH

The 3DFX Voodoo was not the first dedicated 3D graphics chipset by any means, but it became the favourite for gamers among the early mass-market GPUs. It would be found on a 3D-processing-only PCI card that sat on the feature connector of your SVGA card. The Voodoo took any game that supported its Glide API into the world of (for the time) smooth and beautiful 3D. They’re worth a bit now, but if you don’t fancy forking out for mid-’90s silicon in 2026, there’s another option. [Francisco Ayala Le Brun] has implemented the 3DFX Voodoo 1 in SpinalHDL for FPGAs.

The write-up goes into the Voodoo’s architecture. Where the parts of a modern GPU are programmable for the various functions it can do, in this part they are dedicated hardware functions for the various graphics tricks the chip can perform. Implementing such an architecture on an FPGA led to bugs and timing problems, and the write-up deals with that in detail.

The whole thing can be found in a GitHub repository if you’re curious, and is definitely worth a read for anyone interested in 1990s retrocomputing. 3DFX themselves would eventually be swallowed by Nvidia, a rival whose offerings would overtake them at the end of the ’90s, but they still represent a somewhat special moment. Don’t forget, if you have the real thing, you can probably upgrade its memory.

Header image: Konstantin Lanzet, GFDL.

On Friday, Reuters reported that Amazon is going to try to get into the smartphone game…again. The Fire Phone was perhaps Amazon’s biggest commercial misstep, and was only on the market for about a year before it was discontinued in the summer of 2015. But now industry sources are saying that a new phone code-named “Transformer” is in the works from the e-commerce giant.

At this point, there’s no word on how much the phone would cost or when it would hit the market. The only information Reuters was able to squeeze out of their contacts was that the device would feature AI heavily. Real shocker there — anyone with an Echo device in their kitchen could tell you that Amazon is desperate to get you talking to their gadgets, presumably so they can convince you to buy something. While a smartphone with even more AI features we didn’t ask for certainly won’t be on our Wish List, if history is any indicator, we might be able to pick these things up cheap on the second-hand market.

On the subject of AI screwing everything up, earlier this week, the Electronic Frontier Foundation reported that The New York Times had started blocking the Internet Archive’s crawlers, citing concerns over their content being scraped up by bots for training data. The EFF likens this to a newspaper asking libraries to stop storing copies of their old editions, and warns that in an era where most people get their news via the Internet, not having an archived copy of sites like The Times will put holes in the digital record. They also point out that mirroring web pages for the purposes of making them more easily searchable is a widely accepted practice (ask Google) and has been legally recognized as fair use in court.

Assuming we take the NYT’s side of the story at face value, there’s a tiny part of our cold robotic heart that feels some sympathy for them. Over the last year or so, we’ve noticed some suspicious activity that we believe to be bots siphoning up content from the blog and Hackaday.io, and it’s resulted in a few technical headaches for us. On the other hand, what’s Hackaday here for if not to share information? Surely the same could be said for any newspaper, be it the local rag or The New York Times. If a chatbot learning some new phrases from us is the cost of doing business in 2026, so be it. Can’t stop the signal.

Switching gears to the world of aerospace, NASA’s X-59 supersonic research aircraft had to abort a test flight on Friday after just nine minutes in the air. The plane is designed to demonstrate techniques which promise to reduce or eliminate the sonic booms heard on the ground during supersonic flight, and is currently being put through its paces at Armstrong Flight Research Center in Edwards, California.

The space agency hasn’t clarified exactly what the issue was, but after the pilot saw a warning indicator in the cockpit, the decision was made to end the flight early so engineers could take a look at the problem. Given that the X-59 went on to make an uneventful landing, it sounds like things weren’t too dire. Hopefully, that means it won’t be long before the sleek experimental aircraft is back in the air.

Friday also saw the towering Space Launch System rocket return to the launch pad ahead of a potential April 1st (no, really) liftoff for Artemis II. There are about a million things that could further delay the mission, from technical issues to suspicious looking cloud formations over Cape Canaveral, but we’re certainly in the final stretch now. The 10-day mission will see four astronauts run through a packed schedule of experiments and demonstrations as they become the first humans to swing by the Moon since the Apollo program ended in 1972.

Finally, the National Museum of the U.S. Air Force has released a video taken by a drone flying around their collection of Cold War era aircraft. Seasoned FPV pilots will probably notice it’s not the most technically impressive flight out there, but it does provide some viewpoints that simply wouldn’t be possible otherwise. It’s also a bit surreal to see these aircraft, once the absolute state-of-the-art and developed at an unimaginable cost, collecting dust while a $300 drone that packs in higher resolution optics and far more processing power literally flies circles around them.

See something interesting that you think would be a good fit for our weekly Links column? Drop us a line, we’d love to hear about it.

Odds are, you’ve taken pills before; it’s a statistical certainty that some of you reading this took several this morning. Whenever you do, you’re at the mercy of the manufacturer: you’re trusting that they’ve put in the specific active ingredients in the dosage listed on the package. Alas, given the world we live in, that doesn’t always happen. Double-checking actual concentrations requires expensive lab equipment like gas chromatography. It turns out checking for counterfeit pills is easier than you’d think, thanks to a technique called Disintegration Fingerprinting.

It’s delightfully simple: all you need is a clear plastic cup, a stir plate, and a handful of electronic components — namely, a microcontroller, a servo, and an IR line-following sensor. You’ve probably played with just such a sensor: the cheap ones that are a matched pair of LED and photodetector. It works like this: the plastic cup, filled with water, sits upon the stir plate. To start the device, you turn on the stir plate and actuate the servo to drop the pill in the water. The microcontroller then begins recording the signal from the photo-diode. As the pill breaks up and/or dissolves in the water, the swirling bits are going to reflect light from the IR LED. That reflectance signal over time is the Disintegration Fingerprint (DF), and it’s surprisingly effective at catching fakes according to the authors of the paper linked above. Out of 32 different drug products, the technique worked on 90% of them, and was even able to distinguish between generic and brand-name versions of the same drug.

Of course, you do need a known-good sample to generate a trustworthy fingerprint, and there’s that pesky 10% of products the technique doesn’t work on, but this seems like a great way to add some last-mile QA/QC to the drug distribution chain, particularly in low and middle-income countries where counterfeit drugs are a big problem.

We’ve featured pill-identifiers before, but machine vision is going to be much more easily fooled by counterfeits than this method. If your problem isn’t worrying that your pills are fake, but forgetting to take them, we’ve had projects to help with that, too.

Thanks to [Zorch] for the tip!

Thermal energy storage is pretty great, as phase-change energy storage is very consistent with its energy output over time, unlike chemical batteries. You also get your pick from a wide range of materials that you can either heat up or cool down to store energy. Here, the selection is mostly dependent on how you wish to use that energy at a later date. [Hyperspace Pirate] is mostly interested in cooling down a house, on account of living in Florida.

As can be seen in the top image, the basic setup is pretty straightforward. PV solar power charges a battery until it’s fully charged. Then an MCU triggers a relay on the AC inverter, which then starts the cooling compressor on the water reservoir. This proceeds to phase change the water from a liquid into ice. The process can later be reversed, which will draw thermal energy out of the surrounding air and thus provide cooling.

Although water is not the most interesting substance to pick for the

thermal energy storage, it can provide 1 kWh of cooling power in 10.8 kg, or 92.8 kWh in a mere m³. This makes it much more compact as well as cheaper than chemical storage using batteries.

After charging the main compressor loop with R600 (N-butane), the system is trialed with a small PV solar array that manages to freeze the entire bucket of water. Courtesy of insulation, it’s kept that way for a few days, giving plenty of time for the separate glycol-filled loop to dump thermal energy into it and push cold air into the surrounding environment. This prototype managed to cool down [Hyperspace Pirate]’s car in just two hours, which is good enough for a proof-of-concept.

Although the CRT has largely disappeared from our everyday lives, there was a decades-long timeframe when this was effectively the only display available. It’s an analog display for an analog world, and now that almost everything electronic is digital, these amazing pieces of technology are largely relegated to retro gaming and a few other niche uses. [Maurycy] has a unique CRT that’s small enough to fit in a handheld television, but since there aren’t analog TV stations anymore, he decided to build his own with nothing but an 8-bit microcontroller and a few other small parts.

The microcontroller in question is a fairly standard 8-bit AVR. These microcontrollers have one major limitation when generating the VHF and UHF radio signals needed for analog TV: their natural clock speed is much too low. The maximum output frequency of a pin on this microcontroller is only 6 MHz, and [Maurycy] needs something about two orders of magnitude faster. To solve this problem, [Maurycy] uses a quirk of the square wave generated by toggling a pin at its maximum frequency, which is that a wide range of harmonics will be generated, some of which will have a high enough frequency to be picked up on the handheld analog TV. The microcontroller is configured to use two pins.  Toggling the pins into various states allows the humble AVR to generate a usable TV signal.

The scan rate for CRTs is comparably low as well. At the beginning of each frame, there’s enough processing power left on the microcontroller to play Conway’s Game of Life, which is then sent out over the airwaves to the TV. [Maurycy] notes that his harmonics-based video generation method is extremely noisy and probably wouldn’t pass FCC muster.  However, the signal Power is so low that it’s unlikely to interfere with anything. If you’re curious about these unusual sideways-built CRTs, though, we recently saw someone take two apart and use them to build a CRT-based VR headset.

The best part about retro computing is the idea that you’ll save some poor system from being scrapped and revive it to a working state, at which point you can bask in the glory of a job well done. That’s when reality often strikes hard, and you find yourself troubleshooting a maddening list of issues as you question everything about your life choices. Such was the case with [This Does Not Compute] over at YouTube with a Sun SPARCstation IPX that decided to put up a big fight.

This is the second video of a series. In the first installment, the PSU was repaired, and a boot failure was diagnosed. The system’s onboard diagnostic led to the assumption that one of the 8 kB SRAM ICs was defective.  You can readily get SRAM replacements, so it seemed to be an easy fix. Unfortunately, the fun was only beginning as the system reported the exact same error after the SRAM was replaced.

After flipping a virtual table, the mainboard was swapped with a donor one from a scrapped laptop system. With this, it booted, but the video output showed only vertical lines. Obviously, the solution here was to insert a replacement video card, which not only fixed the display output but also demonstrates once again that you can fix many vintage computing issues by simply replacing hardware.

With the system now seemingly happy, a disk drive was added so that Solaris 2.3 could be installed. This turned into another confusing job. Getting into the GUI was seemingly impossible. Ergo, this virtual table got flipped, too, and Solaris 2.6 was installed. Finally, the system got to the desktop GUI.  At this point it was clear that this 40 MHz mid-range SPARCstation from 1991 is no speed monster.  Just drawing windows was slower than an X11 remote session over 9600 baud dial-up.

Installing more RAM might have helped here, but the system requires parity FP RAM, which could have been purchased, but at some point, you have to decide how much money you want to throw at an old system like this. Although these are still pretty interesting systems in their own right, it should be clear that they are not easy systems to repair or maintain.

We’re used to electronic parts of the same type staying predictably the same, sometimes over many years. An early Z80 from the mid 1970s can be exchanged with one from the end of production a few years ago, for example. This week, we’ve had DMs from several readers who’ve found that this is not always the case, and the culprit is surprising. Espressif has released a new revision of their P4 application processor, and though it’s ostensibly the same, there are a couple of changes that have been catching people out.

The changes lie in both hardware and software, in that there’s a pin that’s changed from NC to a power rail, a few extra passives are needed, and firmware must be compiled separately for either revision. The problem is that they are being sold as the same device and appear in some places under the same SKU! This is leading to uncertainty as to which P4 revision is in stock at wholesalers. We’ve been told about boards designed for the old revision being assembled with the new one, a situation difficult to rework your way out of. Designers are also left uncertain as to which firmware build is needed for boards assembled in remote factories.

The ESP32-P4 is an impressive part for its price, and we’re sure that we’ll be seeing plenty of projects using this new revision over the coming years. We’re surprised that it doesn’t have a different enough part number and that the wholesalers have seemingly been caught napping by the change. We’re told that some of the well-known Chinese assembly houses are now carrying the two chips as separate SKUs, but that’s scant consolation for a designer with a pile of boards carrying the wrong part. If you’re working with the P4, watch out, make sure your board is designed for the latest revision, and ask your supplier to check which chips you’ll get.

If the P4 is new to you, we’ve already seen a few projects using it.