News

3619 Articles

The Greatest Computer Ever Now Gets A New, Injection Molded Clear Case

January 21, 2019 by 55 Comments

The Macintosh SE/30 is the greatest computer ever made. It was a powerhouse when it was launched almost exactly thirty years ago today. You could stuff 128 Megabytes of RAM into it, an absolutely ludicrous amount of RAM for 1989. You could put Ethernet in it. You could turn the 1-bit black or white internal display into an 8-bit grayscale display. I think there was a Lisp card for it. These were just the contemporaneous hacks for the SE/30. Now, people are actively developing for this machine and putting Spotify on it. There’s a toolbar extension for Macs of this era that will let you connect to a WiFi network. You’ll be hard pressed to find a computer that still has a fanbase this big thirty years after release.

Now, there’s a project to create new injection molded cases for the Mac SE/30 (and the plain ‘ol SE). These cases will be clear, just like Apple prototypes of the era. It’s also one of the most difficult injection molding projects retrocomputer enthusiasts have ever taken up.

Over the years, we’ve seen some interesting projects in the way of creating new plastic cases for old computers. The most famous is perhaps the remanufacturing of Commodore 64C cases. Instead of a purely community-driven project, this was an accident of history. The story goes that one guy, [Dallas Moore], went to an auction at an injection molding factory. The owner mentioned something about an old computer, and wheels started turning in someone’s head. A Kickstarter later, and everyone who wanted a new C64 case got one. You could get one in translucent plastic to go with the retro aesthetic.

New cases for the Amiga A1200 have also been made thanks to one fan’s Solidworks skills and a Kickstarter campaign. There is, apparently, a market for remanufactured cases for retrocomputers, and it’s just barely large enough to support making new injection molding tooling.

So, about that SE/30. The folks on the 68k Macintosh Liberation Army forums are discussing the possibility of making a new case for the greatest computer Apple will ever make. The hero of this story is [maceffects] who has already modeled the back ‘bucket’ of the SE/30 and printed one out on a filament printer (check out the videos below). This was then printed in clear SLA, and the next step is crowdfunding.

While this isn’t a complete case — a front bezel would be needed to complete the case — it is an amazing example of what the retrocomputing community can do. The total cost to bring this project to fruition would be about $15,000 USD, which is well within what a crowdfunding campaign could take in. Secondary runs could include a translucent Bondi Blue polycarbonate enclosure, but that’s pure speculation from someone who knows what would be the coolest project ever.

Continue reading “The Greatest Computer Ever Now Gets A New, Injection Molded Clear Case” →

Sly Guy Nabs Pi Spy

January 20, 2019 by 38 Comments

When one of [Christian Haschek’s] co-workers found this Raspberry Pi tucked into their network closet, he figured it was another employee’s experiment – you know how that goes. But, of course, they did the safe thing and unplugged it from the network right away. The ensuing investigation into what it was doing there is a tour de force in digital forensics and a profile of a bungling adversary.

A quick check of everyone with access to that area turned up nothing, so [Christian] shifted focus to the device itself. There were three components: a Raspberry Pi model B, a 16GB SD card, and an odd USB dongle that turned out to be an nRF52832-MDK. The powerful SoC on-board combines a Cortex M4 processor with the RF hardware for BLE, ANT, and other 2.4 GHz communications. In this case, it may have been used for sniffing WiFi or bluetooth packets.

The next step was investigating an image of the SD card, which turned out to be a resin install (now called balena). This is an IoT web service that allows you to collect data from your devices remotely via a secure VPN. Digging deeper, [Christian] found a JSON config file containing a resin username. A little googling provided the address of a nearby person with the same name – but this could just be coincidence. More investigation revealed a copyright notice on some mysterious proprietary software installed on the Pi. The copyright holder? A company part-owned by the same person. Finally, [Christian] looked into a file called resin-wifi-01 and found the SSID that was used to set up the device. Searching this SSID on wigle.net turned up – you guessed it – the same home address found from the username.

But, how did this device get there in the first place? Checking DNS and Radius logs, [Christian] found evidence that an ex-employee with a key may have been in the building when the Pi was first seen on the network. With this evidence in hand, [Christian] turned the issue over to legal, who will now have plenty of ammunition to pursue the case.

If you find the opportunity to do some Linux forensics yourself, or are simply interested in learning more about it, this intro by [Bryan Cockfield] will get you started.

Building And Controlling 19 LEDs & Five Buttons From Five Outputs

January 20, 2019 by 3 Comments

Numbers are hard enough in English, but [Sadale] decided to take things a step further by building a calculator that works in Toki Pona. The result is Ilo Nanpa, an awesome hardware calculator that works in this synthetic minimal language. This is a bit harder than you might think, because Toki Pona doesn’t have digits in the same way that Neo-Latin languages like English do. Instead, you combine smaller numbers to make bigger ones. One is Wan, Two is Tu, but three is Wan Tu (1+2). As you might expect, this makes dealing and representing larger numbers somewhat complicated.

Ilo Nanpa gets around this in a wonderfully elegant way, and with some impressive behind the scenes work. The calculator has 16 LEDs, nine buttons and a slider switch, but they are all controlled and read through just five IO pins on the STM8S001J3 controller that runs the device.

That’s because {Sadale] did some remarkable work with multiplexing and charlieplexing. Multiplexing is controlling more outputs than there are control inputs by using rows and columns: it is how the LED display you are probably reading this on can be controlled by just a few wires. By switching through these rows and columns at a higher speed than the eye can see, you create the illusion of a single, continuous display.

Charlieplexing takes this a step further by using multiple voltages on a single connection to further split the signal. With the clever use of voltage dividers the directional properties of LEDs and multiple voltage levels, the Ilo Nanpa runs all of the LEDs and senses all of the buttons and the slider from just five pins. That’s a remarkably neat piece of design, and it is worth spending some time looking over the excellent explanation of the process that [Sadale] wrote to see how it is done, and poring over the code for the device to see how he programmed this all into a single low powered chip. And, while you are reading, you might pick up a few words of Toki Pona. Tawa Pona!

Continue reading “Building And Controlling 19 LEDs & Five Buttons From Five Outputs” →

UPnP, Vulnerability As A Feature That Just Won’t Die

January 14, 2019 by 25 Comments

UPnP — in a perfect world it would have been the answer to many connectivity headaches as we add more devices to our home networks. But in practice it the cause of a lot of headaches when it comes to keeping those networks secure.

It’s likely that many Hackaday readers provide some form of technical support to relatives or friends. We’ll help sort out Mom’s desktop and email gripes, and we’ll set up her new router and lock it down as best we can to minimise the chance of the bad guys causing her problems. Probably one of the first things we’ll have all done is something that’s old news in our community; to ensure that a notorious vulnerability exposed to the outside world is plugged, we disable UPnP on whatever cable modem or ADSL router her provider supplied.

Continue reading “UPnP, Vulnerability As A Feature That Just Won’t Die” →

Manhole Covers Hide Antennas

January 6, 2019 by 63 Comments

5G is gearing up to be the most extensive implementation of mesh networking ever, and that could mean antennas will not need to broadcast for miles, just far enough to reach some devices. That unsightly cell infrastructure stuck on water towers and church steeples could soon be hidden under low-profile hunks of metal we are already used to seeing; manhole covers. This makes sense because 5G’s millimeter radio waves are more or less line-of-sight, and cell users probably wouldn’t want to lose connectivity every time they walk behind a building.

At the moment, Vodafone in the UK is testing similar 4G antennas and reaching 195 megabits/sec download speeds. Each antenna covers a 200-meter radius and uses a fiber network because, courtesy of existing underground infrastructure. There is some signal loss from transmitting and receiving beneath a slab of metal, but that will be taken into account when designing the network. The inevitable shift to 5G will then be a relatively straightforward matter of lifting the old antennas out and laying the new hardware inside, requiring only a worker and a van instead of a construction crew.

We want to help you find all the hidden cell phone antennas and pick your own cell module.

Via IEEE Spectrum.

Green Hacking: Overclocking Photosynthesis

January 5, 2019 by 30 Comments

We think of hacking as bending technology to our will. But some systems are biological,  and we’re also starting to see more hacking in that area. This should excite science fiction fans used to with reading about cultures that work with biological tech, so maybe we’ll get there in the real world too.  Hacking farm crops and animals goes back centuries, although we are definitely getting better at it. A case in point: scientists have found a way to make photosynthesis better and this should lead to more productive crops.

We learned in school that plants use carbon dioxide and sunlight to create energy and produce oxygen. But no one explained to us exactly how that happened. It seems a protein called rubisco is what causes this to happen, but unfortunately it isn’t very picky. In addition to converting carbon (from carbon dioxide) into sugar, it also converts oxygen into toxic compounds called ROS (reactive oxygen species) that most plants then have to spend energy eliminating. Scientists estimate that if you could recover the calories lost in this process, you could feed an additional 200 million people worldwide at current production levels.

Continue reading “Green Hacking: Overclocking Photosynthesis” →

Cheating AI Caught Hiding Data Using Steganography

January 3, 2019 by 63 Comments

AI today is like a super fast kid going through school whose teachers need to be smarter than if not as quick. In an astonishing turn of events, a (satelite)image-to-(map)image conversion algorithm was found hiding a cheat-sheet of sorts while generating maps to appear as it if had ‘learned’ do the opposite effectively[PDF].

The CycleGAN is a network that excels at learning how to map image transformations such as converting any old photo into one that looks like a Van Gogh or Picasso. Another example would be to be able to take the image of a horse and add stripes to make it look like a zebra. The CycleGAN once trained can do the reverse as well, such as an example of taking a map and convert it into a satellite image. There are a number of ways this can be very useful but it was in this task that an experiment at Google went wrong.

A mapping system started to perform too well and it was found that the system was not only able to regenerate images from maps but also add details like exhaust vents and skylights that would be impossible to predict from just a map. Upon inspection, it was found that the algorithm had learned to satisfy its learning parameters by hiding the image data into the generated map. This was invisible to the naked eye since the data was in the form of small color changes that would only be detected by a machine. How cool is that?!

This is similar to something called an ‘Adversarial Attack‘ where tiny amounts of hidden data in an image or other data-set will cause an AI to produce erroneous output. Small numbers of pixels could cause an AI to interpret a Panda as a Gibbon or the ocean as an open highway. Fortunately there are strategies to thwart such attacks but nothing is perfect.

You can do a lot with AI, such as reliably detecting objects on a Raspberry Pi, but with Facial Recognition possibly violating privacy some techniques to fool AI might actually come in handy.