Steel Tubes And Ground Plates Form The Skeleton Of This DIY Vertical CNC Mill

November 27, 2019 by 29 Comments

If you’re going to do it yourself, you might as well outdo yourself. That seems to be the thinking behind this scratch-built CNC mill, and it’s only just getting started.

According to [Kris Temmerman], the build will cost about $10,000 by the time he’s done. So it’s not cheap, and a personal CNC from Tormach can be had for less, but that’s missing the point entirely. [Kris] built most of the structural elements for the vertical mill from cheap, readily available steel tubing, of the kind used for support columns in commercial buildings. Mounted to those are thick, precision-ground steel plates, which eat up a fair fraction of the budget. Those in turn hold 35 mm linear bearings and ball screws for the three axes, each powered by a beefy servo. The spindle is a BT30 with a power drawbar, belt-driven by an external motor that [Kris] doesn’t share the specs on, but judging from the way it flings chips during the test cut in the video below, we’d say it’s pretty powerful.

There’s still plenty to do, not least of which is stiffening the column; perhaps filling it with epoxy granite would do the trick? But it sure looks like [Kris] is building a winner here, and if he keeps the level of craftsmanship up going forward, he’ll have a top-quality machine on his hands.

Continue reading “Steel Tubes And Ground Plates Form The Skeleton Of This DIY Vertical CNC Mill” →

The Golden Age Of Ever-Changing Computer Architecture

November 27, 2019 by 32 Comments

Given the accuracy of Moore’s Law to the development of integrated circuits over the years, one would think that our present day period is no different from the past decades in terms of computer architecture design. However, during the 2017 ACM Turing Award acceptance speech, John L. Hennessy and David A. Patterson described the present as the “golden age of computer architecture”.

Compared to the early days of MS-DOS, when designing user- and kernel-space interactions was still an experiment in the works, it certainly feels like we’re no longer in the infancy of the field. Yet, as the pressure mounts for companies to acquire more computational resources for running expensive machine learning algorithms on massive swaths of data, smart computer architecture design may be just what the industry needs.

Moore’s law predicts the doubling of transistors in an IC, it doesn’t predict the path that IC design will take. When that observation was made in 1965 it was difficult or even impossible to envision where we are today, with tools and processes so closely linked and widely available that the way we conceive processor design is itself multiplying.

Continue reading “The Golden Age Of Ever-Changing Computer Architecture” →

Your Raspberry Pi Cluster Is Not Like This One

November 27, 2019 by 47 Comments

Many readers will have had their first taste of experimentation with cluster computing through the medium of the Raspberry Pi. The diminutive Linux capable boards can easily be hooked up as a group via a network hub, and given the right software become a whole that is greater than the sum of its parts. None of us will however have reached the heights of the Raspberry Pi cluster shown by Oracle at their Oracle OpenWorld conference, a mighty rack packing a cluster of no less than 1060 Pi 3 B+ boards. This machine is touted as a supercomputer and it’s worthy of the name, though perhaps it’s not in the same class as the elite in that field.

Getting that number of individual 3Bs into a human-sized rack is no easy feat, and they have gone for custom 3D-printed racks to hold the boards. PoE would have resulted in too much heat dissipation, so instead they use USB power from an array of large multi-way USB power supplies. A set of switches provide the networking, and a conventional server sits in the middle to provide storage and network booting.

It’s certainly a cool way to wow the crowds at a conference, but we’re unsure whether it delivers the best bang for your supercomputing buck or whether it’s more useful as a large room heater. Meanwhile you can take a look at a few more modest Pi clusters, with unusual operating systems, or slightly more adherence to convention.

Thanks [Frisco] for the tip.

Just In Time For Christmas: Apple Macintosh Prototype For Sale

November 27, 2019 by 21 Comments

We do love a bit of retrotechnology around our workspace. But we have to admit, we really want to find this prototype Apple Mac under the tree this year. There’s only one problem. There’s only one for sale and only two like it known to exist, for that matter. The auction house thinks it will fetch up to $180,000. We will guess that number is low, but we will find out on December 4th.

The 1983 computer has a pre-production plastic housing and a 5.25 inch “twiggy” drive. Apple provided this machine, apparently, to Encore Systems so they could develop MacWrite ahead of the machine’s release date.

Continue reading “Just In Time For Christmas: Apple Macintosh Prototype For Sale” →

Turning OLEDs And Acrylic Into Faux Nixie Tubes

November 26, 2019 by 8 Comments

Love ’em or hate ’em, Nixies and the retro clocks they adorn are here to stay. At least until the world’s stock of surplus Soviet tubes is finally depleted, that is. The glow discharge tubes were last mass manufactured in the 1980s, and while they’re not too hard to get a hold of yet, they will be eventually. And what better way to get ready for that dreaded day than by rolling your own OLED faux Nixie tubes?

Granted, [Derek]’s faux Nixies, appropriately dubbed “Fixies,” require just a touch of willing suspension of disbelief. We’ve never see Nixies with tiny jam jars as envelopes, so that’s probably the first giveaway. But looking past that, the innards of these fake displays do a pretty convincing job of imitating the texture of the real thing.

The numbers themselves are displayed on a 128×64 white OLED display using a Nixie-like True Type font. An orange acrylic filter in front of the display gives it that warm amber Nixie glow, with laser etchings mimicking both the fine hexagonal anode grid and the ghostly cathodes of the non-illuminated numerals. The tubes looked convincing enough that a clock was in order, and after sorting through an I2C bottleneck with the help of a multiplexer, [Derek] had a pretty decent faux-Nixie clock, complete with a solenoid-actuated mechanical gong. The double-digit display for the seconds will no doubt cause some consternation among Nixie purists, but that’s probably part of the fun.

Of course, just because Nixies aren’t being mass-produced today doesn’t mean you can’t get new ones. You just have to be willing to pay for them, and [Dalibor Farný] will gladly set you up with his handmade artisanal Nixies, or even a clock kit using them.

Continue reading “Turning OLEDs And Acrylic Into Faux Nixie Tubes” →

How To Run ML Applications On Particle Hardware

November 26, 2019 by No comments

With the release of TensorFlow Lite at Google I/O 2019, the accessible machine learning library is no longer limited to applications with access to GPUs. You can now run machine learning algorithms on microcontrollers much more easily, improving on-board inference and computation.

[Brandon Satrom] published a demo on how to run TFLite on Particle devices (tested on Photon, Argon, Boron,  and Xenon) making it possible to make predictions on live data with pre-trained models. While some of the easier computation that occurs on MCUs requires manipulating data with existing equations (mapping analog inputs to a percentage range, for instance), many applications require understanding large, complex sets of sensor data gathered in real time. It’s often more difficult to get accurate results from a simple equation.

The current method is to train ML models on specialty hardware, deploy the models on cloud infrastructure, and backhaul sensor data to the cloud for inference. By running the inference and decision-making on-board, MCUs can simply take action without backhauling any data.

He starts off by constructing a simple TGLite model for MCU execution, using mean squared error for loss and stochastic gradient descent for the optimization. After training the model on sample data, you can save the model and convert it to a C array for the MCU. On the MCU, you can load the model, TFLite libraries, and operations resolver, as well as instantiate an interpreter and tensors. From there you invoke the model on the MCU and see your results!

[Thanks dcschelt for the tip!]

Fail Of The Week: Z-Tape Is No Substitute For Solder

November 26, 2019 by 36 Comments

Here at Hackaday, we see all kinds of mechanical construction methods. Some are impressively solid and permanent, while others are obviously temporary in nature. The latter group is dominated by adhesives – sticky stuff like cyanoacrylate glue, Kapton tape, and the ever-popular hot glue. They’ve all got their uses in assembling enclosures or fixing components together mechanically, but surely they have no place in making solid electrical connections, right?

Maybe, maybe not. As [Tom Verbeure] relates, so-called Z-tape just might be an adhesive that can stand in for solder under certain circumstances. Trouble is, he couldn’t find the right conditions to make the tape work. Z-tape, more properly called “Electrically Conductive Adhesive Transfer Tape 9703”,  derives its nickname from the fact that it’s electrically conductive, but only in the Z-axis. [Tom] learned about Z-tape in [Joe FitzPatrick]’s malicious hardware prototyping workshop at the 2019 Hackaday Superconference, and decided to put it to the test.

A card from a Cisco router served as a testbed thanks to an unpopulated chip footprint. The 0.5-mm pin spacing on the TSOP-48 chip was within spec for the Z-tape, but the area of each pin was 30 times smaller than the recommended minimum bonding area. While the chip was held down mechanically by the Z-tape, only five of the 48 pins were electrically connected to the pads. Emboldened by the partial success, [Tom] tried a 28-pin SOIC chip next. The larger pins and pads were still six times smaller than the minimum, and while more of the pins ended up connected by the tape, he was unable to make all 28 connections.

Reading the datasheet for the adhesive revealed that constant pressure from a clamp or clip might be necessary for reliable connections, which suggests that gluing down SMD chips is probably not the best application for the stuff. Still, we appreciate the effort, and the fine photomicrographs [Tom] made showing the particles within the Z-tape that make it work – at least in some applications.