Heathkit’s New RF Meter: Who is it for?

Electronic hackers and ham radio operators of a certain age have a soft spot for the Heathkit brand. Maybe that’s why we had a rush of nostalgia when we saw the Heathkit site had a new product. You may recall that Heathkit had gone the way of the dodo until a few years ago when the brand started to resurface. Their latest kit is a precision RF meter which is available on preorder.

Before there were websites and hacker spaces and all the modern push to “do it yourself,” Heathkit was teaching people electronics through kit building. Sure, they were known for ham radio and test equipment, but many people built stereos (hi-fi), TVs, radio control gear, computers, and even robots. All with manuals that are hard to imagine if you haven’t seen one. They were world-class.

Continue reading “Heathkit’s New RF Meter: Who is it for?”

EV3DEV Lego Linux Updated

The ev3dev Linux distribution got an update this month. The distribution targets the Lego EV3 which is a CPU Lego provides to drive their Mindstorm robots. The new release includes the most recent kernel and updates from Debian 8.8. It also contains tools needed for some Wi-Fi dongles and other updates.

If you haven’t seen ev3dev before, it is quite simply Linux that boots on the EV3 hardware using an SD card. You don’t have to reflash the computer and if you want to return to stock, just take out the SD card. You can also use ev3dev on a Raspberry Pi or BeagleBone, if you like. There’s a driver framework included for handling sensors, motors, and other items using the file system.

Continue reading “EV3DEV Lego Linux Updated”

Nominal Lumber Sizes Land Home Depot And Menards In Hot Water

Hard times indeed must have fallen upon the lawyers of the American mid-west, for news reaches us of a possible class-action lawsuit filed in Chicago that stretches the bounds of what people in more gainful employment might consider actionable. It seems our legal eagles have a concern over the insufficient dimensions of their wood, and this in turn has caused them to apply for a class action against Home Depot and Menards with respect to their use of so-called nominal sizing in the sale of lumber.

If you have ever bought commercial lumber you will no doubt understand where this is going. The sawmill takes a piece of green wood straight from the forest, and cuts it to a particular size. It is then seasoned, either left to dry out and mature in the open air or placed in a kiln to achieve the same effect at a more rapid pace. This renders it into the workable lumber you expect to use, but causes a shrinkage of the wood that since it depends on variables such as moisture can not be accurately quantified. Thus a piece of wood cut by the sawmill at 4 inches square could produce a piece of seasoned lumber somewhere near 3.5 inches square. It would thus be sold as having only a nominal size of 4 inches This has been the case as long as commercial lumber has been produced, we’d guess for something in the region of a couple of centuries, and is thus unlikely to be a surprise to anyone in the market for lumber.

So, back to the prospective lawsuit. Once the hoots of laughter from the entire lumber, building, and woodworking industries have died down, is their contention that a customer being sold a material of dimension 3.5 inches as 4 inches is being defrauded a valid one? We are not lawyers here at Hackaday, but we’d expect the long-established nature of nominal lumber sizing to present a tough obstacle to their claim, as well as the existence of other nominally sized products in the building industry such as rolled steel joists. Is it uncharitable of us to characterise the whole escapade as a frivolous fishing exercise with the sole purpose of securing cash payouts? Probably not, and we hope the judges in front of whom this is likely to land agree with us.

If you have any thoughts on this case, especially if you have a legal background, we’d love to hear from you in the comments.

Sawn lumber image: By Bureau of Land Management (Oregon_BLM_Forestry_10) [CC BY 2.0].

Getting Data Off Proprietary Glucometers Gets a Little Easier

Glucometers (which measure glucose levels in blood) are medical devices familiar to diabetics, and notorious for being proprietary. Gentoo Linux developer [Flameeyes] has some good news about his open source tool to read and export data from a growing variety of glucometers. For [Flameeyes], the process started four years ago when he needed to send his glucometer readings to his doctor and ended up writing his own tool. Previously it was for Linux only, but now has Windows support.

Glucometers use a variety of different data interfaces, and even similar glucometers from the same manufacturer can use different protocols. Getting the data is one thing, but more is needed. [Flameeyes] admits that the tool is still crude in many ways, lacking useful features such as HTML output. Visualization and analysis are missing as well. If you’re interested in seeing if you can help, head over to the GitHub repository for glucomerutils. Also needed are details on protocols used by different devices; [Flameeyes] has only been able to reverse-engineer the protocols of meters he owns.

Speaking of glucometers, there is a project for a Universal Glucometer which aims to be able to use test strips from any manufacturer without needing to purchase a different meter.

Thanks for the tip, [Stuart]!

Autonomous Transatlantic Seafaring

[Andy Osusky]’s project submission for the Hackaday Prize is to build an autonomous sailboat to cross the Atlantic Ocean. [Andy]’s boat will conform to the Microtransat Challenge – a transatlantic race for autonomous boats. In order to stick to the rules of the challenge, [Andy]’s boat can only have a maximum length of 2.5 meters, and it has to hit the target point across the ocean within 25 kilometers.

The main framework of the boat is built from aluminum on top of a surfboard, with a heavy keel to keep it balanced. Because of the lightweight construction, the boat can’t sink and the heavy keel will return it upright if it flips over. The sail is made from ripstop nylon reinforced by nylon webbing and thick carbon fiber tubes, in order to resist the high ocean winds.

The electronics are separated into three parts. A securely sealed Pelican case contains the LiFePo4 batteries, the solar charge controller, and the Arduino-based navigation controller. The communications hardware is kept in polycarbonate cases for better reception. One case contains an Iridium satellite tracker, compass, and GPS, the other contains two Globalstar trackers. The Iridium module allows the boat to transmit data via the Iridium Short Burst Data service. This way, data such as GPS position, wind speed, and compass direction can be transmitted.

[Andy]’s boat was launched in September from Newfoundland headed towards Ireland. However, things quickly seemed to go awry. Storms and crashes caused errors and the solar chargers seemed not to be charging the batteries. The test ended up lasting about 24 days, during which the boat went almost 1000km.

[Andy] is redesigning the boat, changing to a rigid sail and enclosing the hardware inside the boat. In the meantime, the project is open source, so the hardware is described and software is available on GitHub. Be sure to check out the OpenTransat website, where you can see the data from the first sailing. Also, check out this article on autonomous kayaks, and this one about a swarm of autonomous boats.

Continue reading “Autonomous Transatlantic Seafaring”

Catastrophic Forgetting: Learning’s Effect on Machine Minds

What if every time you learned something new, you forgot a little of what you knew before? That sort of overwriting doesn’t happen in the human brain, but it does in artificial neural networks. It’s appropriately called catastrophic forgetting. So why are neural networks so successful despite this? How does this affect the future of things like self-driving cars? Just what limit does this put on what neural networks will be able to do, and what’s being done about it?

The way a neural network stores knowledge is by setting the values of weights (the lines in between the neurons in the diagram). That’s what those lines literally are, just numbers assigned to pairs of neurons. They’re analogous to the axons in our brain, the long tendrils that reach out from one neuron to the dendrites of another neuron, where they meet at microscopic gaps called synapses. The value of the weight between two artificial neurons is roughly like the number of axons between biological neurons in the brain.

To understand the problem, and the solutions below, you need to know a little more detail.

Continue reading “Catastrophic Forgetting: Learning’s Effect on Machine Minds”

Simple Electric Bike Conversion from 3D-Printed Parts

Challenge: Perform an electric conversion on a bicycle. Problem: No significant metal working skills or equipment. Solution: 3D print everything needed to electrify the bike.

At least that’s the approach that [Tom Stanton] took to his electric bike build. Having caught the electric locomotion bug on a recent longboard build, [Tom] undertook the upgrade of a cheap “fixie,” or fixed-gear bike. His delta printer was big enough for the motor mount and weather-resistant ESC enclosure, but he needed to print the drive pulley in four quadrants that were later glued together. We can’t say we hold much faith in the zip ties that transmit all the torque through the rear wheel’s spokes, but as a proof of concept it seems sturdy enough. With a throttle from an electric scooter and a battery in a saddle bag, the bike turns in pretty decent performance — at least after a minor gearing change. And everything blends in or accents the black frame of the bike, so it’s a good-looking build to boot.

Want to catch the cheap electric personal transportation bug too? Check out this electric longboard, or this all-terrain hoverboard.

Continue reading “Simple Electric Bike Conversion from 3D-Printed Parts”