On 11 March, 2011, a massive magnitude 9.1 earthquake shook the west coast of Japan, with the epicenter located at a shallow depth of 32 km, a mere 72 km off the coast of Oshika Peninsula, of the Touhoku region. Following this earthquake, an equally massive tsunami made its way towards Japan’s eastern shores, flooding many kilometers inland. Over 20,000 people were killed by the tsunami and earthquake, thousands of whom were dragged into the ocean when the tsunami retreated. This Touhoku earthquake was the most devastating in Japan’s history, both in human and economic cost, but also in the effect it had on one of Japan’s nuclear power plants: the six-unit Fukushima Daiichi plant.
In the subsequent Investigation Commission report by the Japanese Diet, a lack of safety culture at the plant’s owner (TEPCO) was noted, along with significant corruption and poor emergency preparation, all of which resulted in the preventable meltdown of three of the plant’s reactors and a botched evacuation. Although afterwards TEPCO was nationalized, and a new nuclear regulatory body established, this still left Japan with the daunting task of cleaning up the damaged Fukushima Daiichi nuclear plant.
Removal of the damaged fuel rods is the biggest priority, as this will take care of the main radiation hazard. This year TEPCO has begun work on removing the damaged fuel inside the cores, the outcome of which will set the pace for the rest of the clean-up.
Although less popular these days, wire-wrap is still a very relevant, easily reversible solder-free way to assemble (prototype) systems using wire-wrap wire and a wire-wrap tool. This latter tool can be either a hand or powered tool, but all it has to do is retain the stripped wire, fit around the wire-wrapping post and create a snug, oxidation-proof metal-metal contact fit. For the very common 30 AWG (0.25 mm) wire-wrap wire, the Jonard Tools (OK Industries) WSU-30M wire-strip-unwrap tool is pretty much the popular standard. It allows you to strip off insulation, wrap and unwrap connections all with one tool, but the question is whether you can just 3D print a wrap-unwrap tool that’s about as good?
First a note about cost, as although the genuine WSU-30M has risen in cost over the years, it can still be obtained for around $50 from retails like Mouser, while clones of varying quality can be obtained for around $15 from your favorite e-tailer website. From experience, these clones have quite sloppy tolerance, and provide a baseline of where a wrapping tool becomes unusable, as they require some modding to be reliable.
Wire-wrap tool model by [KidSwidden] on Thingiverse.Taking a quick look at the wire-wrap tools available on Thingiverrse, we can see basically two categories: one which goes for minimally viable, with just a cylinder that has a hole poked on the side for the stripped wire to fit through, as these versions by [JLSA_Portfolio], [paulgeneres], [orionids] and [cmellano]. The WSU-30M and similar tools have a channel on the side that the stripped wire is fed into, to prevent it from getting tangled up and snagging. On the clone units this channel often has to be taped off to prevent the wire from escaping and demonstrating why retaining the wire prior to wrapping is a good idea.
This leads us to three examples of a 3D printed wire-wrap tool with such a wire channel: by [KidSwidden] (based on a Radio Shack unit, apparently), another by [DieKatzchen] and an interesting variation by [4sStylZ]. Naturally, the problem with such fine features is that tolerance matter a lot, with an 0.2 mm nozzle (for FDM printers) recommended, and the use of an SLA printer probably a good idea. It’s also hard to say what kind of wire-wrap connection you are going to get, as there are actually two variants: regular and modified.
The starting guide to wire-wrapping by Sparkfun uses the WSU-30M, which as the name suggests uses modified wire-wrap, which means that part of the wire insulation is wrapped around the bottom of the post, for extra mechanical stability, effectively like strain-relief. A lot of such essential details are covered in this [Nuts and Volts] article which provides an invaluable starting guide to wire-wrapping, including detecting bad wraps.
Naturally, the 3D printed tools will not include a stripper for the wire insulation, so you will have to provide this yourself (PSA: using your teeth is not recommended), and none of these 3D models include an unwrap tool, which may or may not be an issue for you, as careful unwrapping allows you to reuse the wire, which can be useful while debugging or reworking a board.
Top image: completed wire-wrap on a post. (Credit: Sparkfun)
Raw network sockets are a curious beasts, as unless you have a strong urge to implement your own low-level network protocol, it’s a topic that is probably best left to the (well-paid) experts. That said, you can totally use raw sockets in virtually every operating system, but one should be aware of a few things, the lack of portability being one of them. This is what tripped [Swagnik] up while trying to write a low-level network ping (ICMP) utility, by reading the Linux socket documentation while testing on MacOS. It’s all BSD-style sockets, after all, right?
As it turns out, the network stacks in Linux and MacOS have some subtle differences, which become apparent when you read the friendly manuals. For Linux, the raw(7) man entry for IPv4 sockets make it clear that the IP_HDRINCL socket option is default by default for IPPROTO_RAW sockets. This is different from MacOS, which is effectively FreeBSD with glossy makeup. Like FreeBSD, the MacOS man page makes it clear that the IP_HDRINCL option is not set by default.
So that’s easy, right? Just fire off a setsockopt() call on the raw socket and that’s done. Not quite. The Linux man page notes that it cannot receive all IP protocols, while the FreeBSD/MacOS version makes no such exceptions. There is also the issue of endianness, which is where [Swagnik]’s blog post seems to err. The claim is that on MacOS the received IPv4 raw socket header is in host (little endian) order, while the documentation clearly notes that these are in network (big endian) order, which the blog post also shows.
Where things get really fun is when moving from IPv4 raw sockets to IPv6 raw sockets, as [Michael F. Schönitzer] covered for Linux back in 2018 already. IPv6 raw sockets drop IP_HDRINCL and requires a whole different approach. The endianness also changes, as IPv6 raw sockets under Linux must send and will receive data in network byte order, putting it in line with FreeBSD raw sockets.
Although initially defined as an issue with GPIO inputs when configured with the internal pull-downs enabled, erratum RP2350-E9 has recently been redefined in the datasheet (page 1341) as a case of increased leakage current. As it is now understood since we previously reported, the issue occurs when a GPIO (0 – 47) is configured as input, the input buffer is enabled, and the pad voltage is somewhere between logic LOW and HIGH. In that case leakage current can be as high as 120 µA with IOVDD = 3.3 V. This leakage current is too much for the internal pull-up to overcome, ergo the need for an external pull-down: 8.2 kΩ or less, per the erratum. Disabling the input buffer will stop the leakage current, but reading the input requires re-enabling the buffer.
GPIO Pad leakage for IOVDD=3.3 V (Credit: Raspberry Pi)
The upshot of this issue is that for input applications, the internal pull-downs are useless, and since PIO applications cannot toggle pad controls, the input buffer toggling workaround is not an option. ADC usage requires one to clear the GPIO input enable. In general any circuit that relies on floating pins or an internal pull-down resistor will be affected.
Although this should mean that the affected A2 stepping of the RP2350 MCU can still be used for applications where this is not an issue, and external pull-downs can be used as a ‘fix’ at the cost of extra power usage, it makes what should have been a drop-in replacement a troubled chip at best. At this point there have still been no definite statements from Raspberry Pi regarding a new (B0) stepping, leaving RP MCU users with the choice between the less flashy RP2040 and the buggy RP2350 for the foreseeable future.
There’s a special kind of satisfaction found in the act of repairing a previously broken device, which is why YouTube is full of repair channels and guides on how to do it yourself. Inspired by this, [Doug Brown] decided to give it a shot himself, with an Elgato HD60 S HDMI capture device as the patient. As per the eBay listing, the device did not show up as a USB device when connected to a computer — a quick probing of the innards revealed that not only were the board voltages being dragged down, but some of the components on the PCB were getting suspiciously hot.
One of the broken switching regulators on the Elgato HD60 S capture device PCB. (Credit: Doug Brown)
On a thermal camera the hot components in question turned out to part of the voltage regulator circuits, one a switching regulator (marked fiVJVE, for Fitipower FP6373A) and another a voltage inverter marked PFNI (Ti TPS60403DBV).
Since both took 5 V, the suspicion was an over-voltage event on the USB side. After replacing the FP6373A and TPS60403 with newly purchased ones, the voltage rails were indeed healthy, and the Elgato sprung to life and could be used for HDMI capture and pass-through. However, the device’s onboard LEDs stubbornly refused to follow the boot-up pattern or show any other color patterns. Was this just a busted Innotech IT1504 LED driver IC?
Swapping it with a pin-compatible Macroblock MB15040 didn’t improve the situation, and the LEDs worked fine when externally controlling the MB15040 on its SPI bus, as well as with the original IT1504. Asking Elgato whether there was a known issue with these status LEDs didn’t lead to anything, so [Doug] decided to tackle the presumed source of the problem: the Nuvoton M031LD2AE MCU that’s supposed to drive the LED driver IC. The PCB helpfully provides a 4-pin JST connector on the board for the MCU’s SWD interface, but Elgato did protect the flash contents, so a straight dump wasn’t going to work.
The binary firmware is however helpfully available from Elgato, with the MCU already running the latest version. This is the point where [Doug] loaded the firmware into Ghidra to begin to understand what exactly this firmware was supposed to be doing. Putting on a fresh MCU with the correct firmware definitely worked, but debugging was hard as the new MCU also enabled protections, so in Ghidra the offending code for this was identified and neutralized, finally allowing for on-chip debugging and leading down another rabbit hole only to find an SPI flash chip at the end.
Ultimately it turned out that all the hardware was working fine. The problem ended up being that the LED patterns stored on the SPI EEPROM had become corrupted, which caused the Nuvoton MCU to skip over them. The same issue was confirmed on a second HD60 S, which makes it seem that this is a common issue with these Elgato capture devices. As a next step [Doug] hopes to figure out a way to more easily fix this issue, as even the streamlined process is still quite convoluted. Whether it is an issue with Elgato’s software or firmware (updater) is unknown at this point, but at least there seems to be a fix for now, even if Elgato support seems to just tell owners to ‘ignore it if capturing works’.
Recently, the EPA and COBB Tuning have settled after the latter was sued for providing emissions control defeating equipment. As per the EPA’s settlement details document, COBB Tuning have since 2015 provided customers with the means to disable certain emission controls in cars, in addition to selling aftermarket exhaust pipes with insufficient catalytic systems. As part of the settlement, COBB Tuning will have to destroy any remaining device, delete any such features from its custom tuning software and otherwise take measures to fully comply with the Clean Air Act, in addition to paying a $2,914,000 civil fine.
The tuning of cars has come a long way from the 1960s when tweaking the carburetor air-fuel ratios was the way to get more power. These days cars not only have multiple layers of computers and sensor systems that constantly monitor and tweak the car’s systems, they also have a myriad of emission controls, ranging from permissible air-fuel ratios to catalytic converters. It’s little surprise that these systems can significantly impact the raw performance one might extract from a car’s engine, but if the exhaust of nitrogen-oxides and other pollutants is to be kept within legal limits, simply deleting these limits is not a permissible option.
COBB Tuning proclaimed that they weren’t aware of these issues, and that they never marketed these features as ’emission controls defeating’. They were however aware of issues regarding their products, which is why they announced ‘Project Green Speed’ in 2022, which supposedly would have brought COBB into compliance. Now it would seem that the EPA did find fault despite this, and COBB was forced to making adjustments.
Although perhaps not as egregious as modifying diesel trucks to ‘roll coal’, federal law has made it abundantly clear that if you really want to have fun tweaking and tuning your car without pesky environmental laws getting in the way, you could consider switching to electric drivetrains, even if they’re mind-numbingly easy to make performant compared to internal combustion engines.
Images that can be interpreted in a variety of ways have existed for many decades, with the classical example being Rubin’s vase — which some viewers see as a vase, and others a pair of human faces.
When the duck becomes a bunny, if you ignore the graphical glitches that used to be part of the duck. (Credit: Steve Mould, YouTube)
Where things get trickier is if you want to create an image that changes into something else that looks realistic when you rotate each section of it within a 3×3 grid. In a video by [Steve Mould], he explains how this can be accomplished, by using a diffusion model to identify similar characteristics of two images and to create an output image that effectively contains essential features of both images.
Naturally, this process can be done by hand too, with the goal always being to create a plausible image in either orientation that has enough detail to trick the brain into filling in the details. To head down the path of interpreting what the eye sees as a duck, a bunny, a vase or the outline of faces.
Using a diffusion model to create such illusions is quite a natural fit, as it works with filling in noise until a plausible enough image begins to appear. Of course, whether it is a viable image is ultimately not determined by the model, but by the viewer, as humans are susceptible to such illusions while machine vision still struggles to distinguish a cat from a loaf and a raisin bun from a spotted dog. The imperfections of diffusion models would seem to be a benefit here, as it will happily churn through abstractions and iterations with no understanding or interpretive bias, while the human can steer it towards a viable interpretation.
