Mechanical keyboards use switches of a few different types. But even those types include myriad variations. How’s a hacker to know just exactly what equipment is out there?
For example, if you grab a fellow cube-farmer’s mechanical keyboard (possibly because they clacked on their Cherry Blue’s just one too many times) and angrily rip off a few keycaps to show you’re serious, what do you see? In most cases you expect to see the familiar color and stem shape of a Cherry MX switch or one of its various clones. But you may find a square box around it like a Kailh Box switch. Or the entire stem is a box (with no +) like a Matias switch. Or sometimes it looks like a little pig snout, making it a Kailh Choc.
There is a fairly wide variety of companies which make key switches suitable for use in a keyboard. Many hew to the electrical and mechanical standards implicitly created by the dominant Cherry GmbH’s common switches but not all. So if you’re designing a PCB for such a keyboard and want to use odd switches, you need to check out the Keyboardio keyswitch_documentation repo!
The keyswitch_documentation repo is an absolute treasure trove of hard to find keyswitch datasheets. Finding official information on Cherry MX switches isn’t too hard (keyswitch_documentation has 22 data sheets for MX series switches, and four for ML). But those Kailh Choc’s? Good luck (here it is in keyswitch_documentation). Did you know Tai-Hao made Matias-esque switches as well as weird rubber keycaps? Well they do, and here’s the datasheet.
We’re keeping this one handy until the next time we need data sheets for weird switches. Make sure to send a note if you find something interesting in here that’s worth noting!
[Charles Ouweland] purchased some parts off Aliexpress and noticed that the Texas Instruments logo on some of his parts wasn’t the Texas Instruments logo at all, it was just some kind of abstract shape that vaguely resembled the logo. Suspicious and a little curious, he decided to take a closer look at the MCP1702 3.3v LDO regulators he ordered as well. Testing revealed that they were counterfeits with poor performance.
Looking at the packages, there were some superficial differences in the markings of the counterfeit MCP1702 versus genuine parts from Microchip, but nothing obviously out of place. To conclusively test the devices, [Charles] referred to Microchip’s datasheet. It stated that the dropout voltage of the part should be measured by having the regulator supply the maximum rated 250 mA in short pulses to avoid any complications from the part heating up. After setting up an appropriate test circuit with a 555 timer to generate the pulses for low duty cycle activation, [Charles] discovered that the counterfeit parts did not meet Microchip specifications. While the suspect unit did output 3.3 V, the output oscillated badly after activation and the dropout voltage was 1.2 V, considerably higher than the typical dropout voltage of 525 mV for the part, and higher even than the maximum of 725 mV. His conclusion? The parts would be usable in the right conditions, but they were clearly fakes.
The usual recourse when one has received counterfeit parts is to dump them into the parts bin (or the trash) and perhaps strive to be less unlucky in the future, but [Charles] decided to submit a refund request and to his mild surprise, Aliexpress swiftly approved a refund for the substandard parts.
While a refund is appropriate, [Charles] seems to interpret the swift refund as a sort of admission of guilt on the part of the reseller. Is getting a refund for counterfeit parts a best-case outcome, evidence of wrongdoing, or simply an indication that low value refund requests get more easily approved? You be the judge of that, but if nothing else, [Charles] reminds us that fake parts may be useful for something perhaps unexpected: a refund.
I’m in the planning stages of a side project for Hackaday right now. It’s nothing too impressive, but this is a project that will involve a lot of electromechanical parts. This project is going to need a lot of panel mount 1/8″ jacks and sockets, vertical mount DIN 5 connectors, pots, switches, and other carefully crafted bits of metal. Mouser and Digikey are great for nearly every other type of electrical component, but when it comes to these sorts of electromechanical components, your best move is usually to look at AliExpress or DealExtreme, finding something close to what you need, and buying a few hundred. Is this the best move for a manufacturable product? No, but we’re only building a few hundred of these things.
I have been browsing my usual Internet haunts in the search for the right bits of stamped brass and injection molded plastic for this project, and have come to a remarkable conclusion. Engineers, apparently, have no idea how to dimension drawings. Drafting has been a core competency for engineers from the dawn of time until AutoCAD was invented, and now we’re finally reaping the reward: It’s now rare to find a usable dimensioned drawing on the Internet.
This post is going to be half rant, half explanation of what is wrong with a few of the dimensioned drawings I’ve found recently. Consider this an example of what not to do. There is no reason for the state of engineering drawing to be this bad.
It’s a fair assumption that the majority of Hackaday readers will be used to working with electronic components, they are the life blood of so many of the projects featured here. In a lot of cases those projects will feature very common components, those which have become commoditized through appearing across an enormous breadth of applications. We become familiar with those components through repeated use, and we build on that familiarity when we create our own circuits using them.
All manufacturers of electronic components will publish a datasheet for those components. A document containing all the pertinent information for a designer, including numerical parameters, graphs showing their characteristics, physical and thermal parameters, and some application information where needed. Back in the day they would be published as big thick books containing for example the sheets for all the components of a particular type from a manufacturer, but now they are available very conveniently online in PDF format from manufacturer or wholesaler websites.
Datasheets are a mine of information on the components they describe, but sometimes they can be rather impenetrable. There is a lot of information to be presented, indeed when the device in question is a highly integrated component such as a DSP or microprocessor the datasheet can resemble a medium-sized book. We’re sure that a lot of our readers will be completely at home in the pages of a datasheet, but equally it’s a concern that a section of the Hackaday audience will not be so familiar with them and will not receive their full benefit. Thus we’re going to examine and explain a datasheet in detail, and hopefully shed some light on what it contains.
The device whose datasheet we’ve chosen to put under the microscope is a transistor. The most basic building block of active semiconductor circuits, and the particular one we’ve chosen is a ubiquitous NPN signal transistor, the 2N3904. It’s been around for a very long time, having been introduced by Motorola in the 1960s, and has become the go-to device for a myriad circuits. You can buy 2N3904s made by a variety of manufacturers all of whom publish their own data sheets, but for the purposes of this article we’ll be using the PDF 2N3904 data sheet from ON Semiconductor, the spun-off former Motorola semiconductor division. You might find it worth your while opening this document in another window or printing it out for reference alongside the rest of this article.
Let’s take a look at all the knowledge enshrined in this datasheet, and the engineering eye you sometimes need to assign meaning to those numbers, diagrams, and formulas.
While [Drew] was in China for the Dangerous Prototypes Hacker Camp, he picked up some very bright, very shiny, and very cheap LED strips. They’re 5 meter “5050” 12V strips with 20 LEDs per meter for about $15 a spool. A good deal, you might think until you look at the datasheet for the controller. If you want an example of how not to document something, this is it.
A normal person would balk at the documentation, whereas [Drew] decided to play around with these strips. He figured out how to control them, and his efforts will surely help hundreds in search of bright, shiny, glowy things.
The datasheet for the LPD6803 controller in this strip – available from Adafruit here – is hilarious. The chip takes in clocked data in the order of Green, Red, and Blue. If anyone can explain why it’s not RGB, please do so. Choice phrasing includes, “VOUT is saturation voltage of the output polar to the grand” and “it is important to which later chip built-in PLL regernate circuit can work in gear.” Apparently the word ‘color’ means ‘gray’ in whatever dialect this datasheet was translated into.
Despite this Hackaday-quality grammar, [Drew] somehow figured out how to control this LED strip. He ended up driving it with an LPC1768 Mbed microcontroller and made a demo program with a few simple animations. You can see a video of that below.
Once you venture beyond the tame, comfortable walls of the 8-bit microcontroller world it can feel like you’re stuck in the jungle with a lot of unknown and oft scary hazards jut waiting to pounce. But the truth is that your horizons have expanded exponentially with the acceptable trade-off of increased complexity. That’s a pretty nice problem to have; the limitation becomes how much can you learn.
Here’s a great chance to expand your knowledge of the STM32 by learning more about the system clock options available. We’ve been working with STM32 chips for a few years now and still managed to find some interesting tidbits — like the fact that the High Speed External clock source accepts not just square waves but sine and triangle waves as well, and an interesting ‘gotcha’ about avoiding accidental overclocking. [Shawon M. Shahryiar] even covers one of our favorite subjects: watchdog timers (of which there are two different varieties on this chip). Even if this is not your go-to 32-bit chip family, most chips have similar clock source features so this reading will help give you a foothold when reading other datasheets.
There is a clock diagram at the top of that post which is small enough to be unreadable. You can get a better look at the diagram on page 12 of this datasheet. Oh, and just to save you the hassle of commenting on it, the chip shown above is not an f103… but it just happened to be sitting on our desk when we started writing.
What you see above is the biggest feature the service brings to the table, the ability to create “snippets” from datasheets by clicking and dragging the area you’d like to save (you can even get a public link to the snippet). Once you have selected a snippet there are a few tools that allow you to make annotations on it. We’ve used the rectangle tool to highlight the clock speed and divider settings in this snippet for an ATmega328 uC. The interface also offers the ability to draw arrows, freehand, or to add text to the snippet. At the bottom of this example we used the description area to notate the fuse settings (in hex) which we most often use with this chip. These snippets and annotations can then be shared with other users of the service, and there’s also a comments section below the snippet for your team to use. See examples of this in the video below.
This solves one of our biggest beefs with PDF datasheets — the ability to jump back and forth and to easily find commonly used sections. This datasheet is 567 pages long and not fun to paw through looking for the same info repeatedly. It also offers rudimentary “favorite” flagging to keep a list of your oft-used sheets — but we’d like to see more options for categorizing our collection. We also find it hard to get by without the Table of Contents functionality we’re used to in our normal document view (evince). We’ve already pestered the lead developer, [Ben Delarre], to add this feature. He’s the same guy who came up with the schematic sharing site CircuitBee. Now would be a great time to mention that this service is owned by Hackaday’s parent company SupplyFrame.
Datasheet.net has a mammoth source of datasheets available through the search, but the list of planned feature additions includes datasheet upload. Also on the list is a “Discussion” feature which sounds interesting to us. What if, through the discussion engine, searching for datasheets also turned up a list of open hardware projects that use this part? We are also drooling over the ability to embed these snippets directly in webpages. [Ben] tells us that’s already built but they didn’t have time to add it to the UI before launch. Gone will be the days of taking screenshots of PDFs for your blog writeup!
PDF delivery of datasheets revolutionized access to information about electronic components. We’re hoping that this marks the next evolution. In addition to better working features, wouldn’t it be nice if you could actually get notifications when new datasheet revisions or errata were published?