A while ago, YouTube started recommending some ASMR videos of leatherwork being done. Handbags, laptop cases, wallets etc. It was fascinating to watch a leather hide being turned in to wonderful objects through the skills, craftsmanship and obscure processes. Over time I began to recognise the individual steps involved and wondered how hard can this craft be? Luckily, YouTube also has a lot of tutorial videos too, and before I knew it, I ordered a starter kit of tools & fittings and some sample leather.

I was given a bag of leather offcuts by a friend which was great to get to grips with the various tools. The starter kit included a key wallet thingy. Whilst I'm not a fan of these kind of key rings, it was a good starter project. Templates are easy to find online, and over a few evenings, I put the key wallet above together. It is far from perfect. Mistakes were made. Things could and should be done better. But a lot was learned. And I was very happy for the results from my first leather project.

Next up was creating a leather sheath for the skiving knife. It was a pretty simple design and works much better than the flimsy cardboard it came packaged with.

Recently I have had need for a small pocket notebook. The Travellers Notebooks are great and the Passport size refill is perfect. However, a notebook is no use without a pen, so some kind of leather cover with integrated pen loop was in order.

The sample leather pieces were 200 x 150mm and all different colours. At that size, they are plenty tall enough, but only just wide enough to fit around the passport size Travellers Notebook refill. Certainly there was no room to trim the edges, but the top and bottom could be a few mm shorter. The burgundy piece felt nice and looked great. There wasn't another piece to match for the inside, but going for the red made for a stylish fashion choice.

I had an idea of having two pen loops on one side of the cover that went over a pen loop on the other side, so that when a pen was inside it kept the notebook cover closed. Due to the tight sizing on the width I realised that I would need to use the middle section of the red piece for the loops and cut an alternating pattern between them. A 50mm wide strip was marked down the middle, and lines drawn. A circular die was used to cut the corners before the main lines were cut. To reduce the thickness where the pen loops joined back in to the panels the ends were skived down. Contact adhesive was allowed to set before using weights or clamps to create the loops.

The inner panels were glued around the edge to the outer cover. Because of the limited width there was only room for a thin bead of contact adhesive, but this was enough to hold it well. A line was marked on the outer cover about 4mm from the edge, and holes punched using a set of pricking irons. A cheap stitching pony made the saddle stitch so much easier than with the key wallet, but it still took almost 2 hours to sew together.

Until the sewing was finished I didn't know for sure if I had left enough space for the notebook covers. It was tight! just a millimetre or two less would have made it impossible to fit, but it did fit, and that's what counts. Also the pen loops as a closure worked well too. One of them needed a slight shave, but otherwise they are great. The biggest issue is that the pens I like are taller than the cover... but that can be fixed with some pen shopping!

Like many people, I had seen wild crazy looking mechanical keyboards, and had been both fascinated and confused by them. How could you get by with so few keys? What does each key actually do? They looked like some kind of wild alien input device, but, at the same time, must work well otherwise the fad would have died out years ago. The high cost and fear of failure, thought, was enough to put me off of trying them out.

While attending OSHCamp earlier this year, a talk by Alan Wood explained a lot of the ideas behind these keyboards, and answered a lot of my questions - except for one; Would I be able to cope adjusting to one?

There was a workshop the following day where all the parts were supplied to build a Reviung 34 keyboard, except for the keycaps, micro controller, and optional laser cut acrylic. This was the perfect low barrier to entry I needed to find out if these low key count options were right for me.

Reviung 34

The Reiung 34 is an Open Source keyboard design by gtips. As the name implies, it uses 34 keys - although it can be built with 33 keys. It is split column staggered, that is to say that the keys are lined up in columns, centrally split. Nominally there are 26 letters (A - Z) keys, plus 3 punctuation, return, space, delete and 2 modifiers.

The keyboard is designed to run a standard Arduino Pro Micro that runs the QMK firmware, although there are lots of compatible micro controllers and firmware that can be used instead. The QMK firmware runs on hundreds of different keyboards and can be easily customised to for any layout you want. It handles modifiers through the use of layers, and can control things like RGB LED lighting and other controls too.

All of the keys (except space) have multiple uses, and some of the regular keys double up as modifiers too. For example, the "Z" key, bottom left corner, which is near to where shift would be on a normal keyboard, will give a z if tapped, but if held down and another key is pressed, will give the shifted output of that other key. So that means you can get any of your upper case keys - except for Z. Luckily, the "/" key on the opposite side also doubles up as a shift when held, so this can also get your upper case keys, including Z.

Other keys work in a similar way for control, alt and the OS (Windows / Mac) key.

The two modifiers either side of the space allow access to other keys such as the number row by holding down the right one, or the shifted number row by holding down the left one, as well as other combinations.

Why would you want to use a keyboard like this?

The biggest advantage of a reduced key keyboard is that your fingers never need to move far from the home keys (ASDF JKLreturn) to type what you want. This reduces fatigue due to less stretching. It also allows for much faster typing too.

A smaller keyboard also takes up much less desk space than a full size keyboard. And, if you travel around, carrying your favourite 102 key mechanical keyboard adds a lot of bulk and weight to your kit.

Of course, it isn't the perfect solution for everybody. There is quite a steep learning curve, although if you're already a reasonable touch typist, it isn't as big as you'd expect. Some key strokes feel like a Dance Dance Revolution challenge for your fingers. Luckily, most of the characters you're likely to type most often are quite intuitive. There are times too when that particular symbol seems like it keeps moving around the keyboard to avoid every being typed!

Although I could already type pretty well, I recommend doing at least the intermediate typing course on typing.com. Whilst typing.com is not aimed at custom keyboards like this, the lessons get you quickly acquainted with your keyboard with regular typing, and familiarity makes things much easier.

How did it work out for me?

Mistakes were made!

I was very excited to build my Reviung 34 at Alans workshop. However, it quickly became apparent that without my own soldering station that I'm used to, and the extra magnification that I am reluctantly realising that I need, that things weren't going well with the assembly. There is a small surface mount diode for each key and 9 surface mount LED packages. Frankly, I disappointed myself with how well this came out on the day. On top of that, the top plate which holds the key switches may look like it is mirrored - but it is not. fitting this on upside-down means that the micro controller cannot be fitted.

The only solution was to unsolder every single key switch to remove the top plate. This is an epic job in itself, but after getting them all out, I could flip the top plate over, re-fit the switches and connect the micro controller. This was where I discovered that there is a good way and a bad way to remove the switches - and I had used the bad way (gripping them vertically with pliers on the top), which damaged most of the switches! Luckily, I had some spare switches, but it still meant removing most switches again. And the replacement switches I had were MX Blue (clicky) vs Yellow (silent). Eventually, though, after fixing many errors and broken pads, I had a working keyboard.

I laser cut the acrylic base plate, and chose to add a vinyl layer to hide the mess hidden below. Time to learn how to type!

Typing on it felt surprisingly good. Much more natural than I expected. Regular typing that doesn't involve too much punctuation seemed to flow really easily. As mentioned earlier, the lessons on typing.com really helped. I initially had issues with mixing up the B and V keys, often trying to type B with the wrong hand. But, like to many things, practice really helps.

One of the first decisions that had to be made was to make this as a 33 or 34 key keyboard. The difference is the space bar can either be a central 2U wide key, or 2 individual keys where the left one is delete and the right is space. Having more keys sounds like a good idea, and a dedicated delete key must be good, right? Well, for me, I found the smaller space key annoying and I kept hitting the delete key with my left thumb when I wanted space. A minor issue, really, but if I had the chance again, I would choose a 33 key layout.

For over 6 weeks now, the Reviung has been my daily driver. Honest opinion is that it's great. For anything number intensive it isn't ideal, however, I do have a separate number pad that makes life much easier. For regular typing, I can easily hit over 60wpm on most typing tests. The common punctuation and symbols that I use, ' @ # ! :-) have all become second nature. A handy cheat-sheet isn't too far away though in case I need ² or ½ or ±

One of the advantages of the QMK firmware is that it is totally customisable. I decided to go with the default layout to start with before making any changes, and ultimately decided that it does everything I want. Some of the layout was a little questionable at first, but clearly it has been developed by people that know a lot more about good keyboard layout than I do. So, I haven't made any modifications, and probably won't either.

Redemption and further down the rabbit hole

Spurred on by adapting to the Reviung better than expected, ashamed by the mistakes I had made at the start, and a desire to do things properly, I decided to try again. As mentioned earlier, the Reviung keyboards are open source, so the PCB design files are available and free to use. Whilst placing an order for some RC2014 circuit boards it was easy enough to add in an order for 5 (minimum order quantity) top and base plates.

This time I did things properly. I laser cut a stencil so that I could use solder paste and hot air reflow for all the surface mount parts.

I have now built 3 of these since the original workshop one. One is my daily driver, and has clicky MX Blue switches and full 3 layers of acrylic. This is a really solid keyboard and perfect for desktop use.

Another was used to upgrade the keyboard from tactile switches on my RCade cabinet with the lower acrylic plate being incorporated in to the chassis. This will undoubtable make me quicker at arcade games like Zork, Hitchhikers Guide To The Galaxy and WordStar!

A third one has been fitted with quiet MX Red switches, and only the lower base plate. This gives a lighter keyboard to carry around, and is quiet enough to use out in coffee shops.

Downloads

I have made a cheat-sheet with all of they key combinations on them for whatever character you want to type. I think I have found them all, although it wouldn't surprise me if there are still some combinations that I haven't discovered yet. Download the pdf here

Where's That Fn Key

A couple of years ago I saw that somebody on the internet had made a Slow Movie Player, and that piqued my interest. The basic idea is that your favourite film is displayed on a screen at a really slow framerate. Eink displays are ideal for this as they use no power except for when updating the image. Their refresh rate is pretty poor compared to OLED displays, but if you're only updating once a minute, that's fine. A decent size eink display is pretty expensive though, so this idea went on the back burner.

Earlier this year, Pimoroni released the Badger. A small eink display with a built in RP2040 (Raspbrry Pi Pico) which was ideal as a programmable name badge - hence the name. This was really cool, although I didn't have a need for a name badge. This also went on the back burner.

The Screenplay

Last week at EMF Camp, I picked up a badger, and suddenly my back burners collided. What... If... A... Badger... Was... A... Slow... Movie... Payer?

The Badger runs Micro Python, and the documentation for this and the specific Badger stuff is great for a non-programmer like me. Using Pimoroni's Python image converter, I could easily get the Badger to display a single image. That was all I needed to know to go ahead with this crazy idea. All I needed was every single frame from a film and set it to display them one at a time. How hard can that be?

Enter Stumbling Block The First - Stage Left

A raw image of 296x128 pixels is pretty small (4.7k actually) but a film has a LOT of them! Typically 24 for every second of it. That soon adds up and is way too big for the Badgers memory. So some external storage is needed. The good folk at Pimoroni thought ahead and were kind enough to add i2c pins and some other signals to the back of the Badger. Sadly, an SD Card needs SPI, and although some of the SPI pins were present, SCK wasn't. I really thought this was going to be a deal breaker here, however, the SCK that serves MISO and MOSI 0 is simply unattached on the RP2040. An ideal excuse to add a Badger Bodge Wire!

There was a certain amount of trial and error getting the Pico SD Library to work here, although this was mainly my error in not understanding how it worked Once I'd got it working, displaying an image worked a treat.

Act 2: A Dress Rehearsal.

The Image Converter program does a great job of converting .png files in to something the Badger likes. All I needed to do was find a way to get a .png for every frame in a movie. there are a lot of free online converters, but I tried the demo version of Total Video Audio Converter and it seemed to do exactly what I wanted, albeit with Trial Version stamped on ever frame. It was $29 for the full paid for version though, so after being happy that it did what I wanted it to do, I paid for a licence code.

The key settings for TVAC is to output .png files at a 296*128, keeping the original aspect ratio and adding padding where necessary. A full movie takes ages to process (and would take much longer to watch!) so I picked a few seconds worth and gave it a try.

Lights! Camera! Action!

With dress rehearsals going well, it was time to make the final edits to the script. Firstly the wiring between the Badger and the SD card module needed tidying up. They are held together with Velcro, and this works pretty well.

With a full version of TVAC purchased, it was time to pick something suitable for the opening night. What better to demonstrate it than the ultimate demo Bad Apple. Every single shot in this stands alone as a great image, and it is perfect for the black and white constraints of the Badger (although dithering works well)

Straight to VHS

A suitable enclosure can really make or break a hardware project. For the Slow Badger Movie Player, it seemed appropriate to reuse an old VHS cassette box. The dimensions were easily found, so I decided to use the laser cutter to make things as accurate as possible. Although technically this worked, I would not recommend laser cutting a VHS box. The amount of (probably poisonous) smoke was quite worrying.

With a suitable image printed out and put inside the case, everything was complete.

Over all, I'm really pleased with how this turned out. It works really well, and takes just over 2 days to get through Bad Apple (Vs 244 seconds original speed) at 2 frames per minute. This is also just within the Lipo battery life too, so a full length movie would definitely need a mains adapter to power it. The small size is a bit of limitation, although larger screens increase the cost and complexity a lot, so I'm happy to call this a wrap.

During the 2021 lockdown I started a laser cutting project I'd wanted to do for a while. It turned out that the research on this went much much deeper than I first expected.

You probably recognise this image, and think you know where it comes from. But what if I told you the story goes back further than you think.

CP1919 is the first radio pulsar discovered in 1967 by Jocelyn Bell Burnell. It transmits radio waves with a pulse 0.04 seconds wide every 1.3373 seconds. This is part of a plot recorded in 1970.

Not every pulse is the same shape or size, and for a while it was thought that it was an extraterrestrial beacon, so it was nicknamed LGM-1 for Little Green Men. This was subsequently disproved. Burnells original plot from Wikipedia.

Interpreting a long wiggly line is difficult, and hard to get info from it, but it can be split up in to sections of 1.3373 seconds for a clearer view of each pulse.

Lining up successive pulses vertically gives a clearer impression of what the pulses look like, although rather than a very wide page, you're going to end up with a very tall page.

In 1970 Harold Craft published a PHD thesis on "Radio observations of the pulse profiles and dispersion measures of twelve pulsars" which featured 80 plots stacked on top of each other.

This plot went on to feature on the cover of the IAU General Assembly in 1970, a 1971 Sci Am article, in 1974 in Graphics Diagrams: The Visualization of Abstract Data by Walter Herdeg and the cover of The Cambridge Encyclopaedia of Astronomy.

If the individual 1.3373 second sections are converted to an appropriate file format, they can be cut from clear acrylic, in a similar way to cutting up the paper tape.

Sequential sections of acrylic can then be stacked up, similar to what Harold Craft did in 1970 with the black and white plot in his thesis.

Putting all 80 pieces of acrylic in the right order (and right orientation!) is both satisfying and frustrating. However, I think the end results are absolutely stunning.

If you think you recognise this plot, but didn't know about CP1919, then you probably know it from the cover of Joy Divisions Unknown Pleasures album which was designed by Peter Saville, taking the image from Cambridge Encyclopaedia Astronomy and inverting it.

After a visual check of the keyboards (see this post), I had a couple of questions, which were do the keyboards work? and how do the keyboards work?

On the basis that neither of the Alphameric had their EPROM windows covered, and that one of the Cherry keyboards had already been modified, I chose the other Cherry one as the most likely to work.  After opening it up (again), I checked the 20 pin header.  It was quite easy to find out what the 5v and Ground pins were (traceable to the power pins on the logic chips and corresponds to red and black wires on pin 1 & 2!), but the others were going to need something more to diagnose

Time to break out the oscilloscope!  It didn't take long to discover that the pins 3-9 changed upon pressing a key.  Curiously, they didn't didn't change when releasing the key though.  Also, depending on which key was pressed, they might not change at all.  Pin 9 was most likely to change and pin 3 was least likely to change.

So, there we have a 7 bit ASCII pattern!  The value was being latched, and pin 11 was a pulse to indicate that a valid signal was ready to be read.

For last years Retro Challenge Andy Collins had done pretty much the same but opposite thing, where he had a faulty Alphameric keyboard that he replaced with an Arduino & USB keyboard.  The timing diagram was particularly useful, so thanks Andy!

If an Arduino can be used to simulate being a keyboard, then it can also be used to read one!  So I plugged in a cheap Uno that I had lying around and wrote a quick sketch to read the data pins D0 - D6 every time the strobe pin goes high, then convert this in to an ASCII number and print the character on the serial port.  I also got it to print out the ASCII code itself, which was handy for working out what keys had non-printable characters

The regular keys A-Z were affected by the two modifier keys to give 3 different characters (a, A and ASCII 1, or b, B and ASCII 2 etc.).  The Del key was the only other one to have 2 modified outputs, but most of the others had just one output regardless of the shift or control key being pushed.

I had a spare Arduino Mini Pro clone lying around and also a 25 pin D socket, so I figured it might be handy to make up a slightly more permanent adaptor that will plug in to the keyboard.  Some of the soldering is a bit dodgy here, but it's all solid enough :-)

Also, the important thing is that it all fits inside the 25 pin D shell!

The Cherry keyboard worked exactly as expected, and allows me to use it with my RC2014!  However, I wanted to know if the other, Alphameric keyboards would work too.

They had the same connection, although I hadn't done any of the diagnostic work that I did with the Cherry one - but it was worth taking a chance.  And it turns out that yes, they work too!  Well, the regular Alphameric keyboard works exactly as the Cherry one does!

The one with the hex keyboard, though, only partially works.  The main keyboard section works as expected, however, the hex keypad sends out odd key codes.  Some of them output several characters, and can even put the RC2014 or the Arduino adaptor in to a weird mode that means the main keyboard types weird stuff until it is reset.  Not sure if this is something to do with bit corruption in the EPROM, or some other kind of error, or if it's actually by design - however, that's for diagnosis another time.  There's still lots more stuff I need to go through yet and time is ticking away...

Arduino Code

#define Strobe 9
#define DATA0 8
#define DATA1 7
#define DATA2 6
#define DATA3 5
#define DATA4 4
#define DATA5 3
#define DATA6 2

int DATABYTE = 0;

void setup() {
  pinMode(DATA0, INPUT);
  pinMode(DATA1, INPUT);
  pinMode(DATA2, INPUT);
  pinMode(DATA3, INPUT);
  pinMode(DATA4, INPUT);
  pinMode(DATA5, INPUT);
  pinMode(DATA6, INPUT);
  pinMode(Strobe, INPUT);
  digitalWrite (Strobe, HIGH);
  Serial1.begin(115200);
  //Serial.println ("Started");
}

void loop() {
int pulse = digitalRead(Strobe);
  if (pulse == 0 ) {
    printASCII();
  }
}

void printASCII() {
  delay (1);
  DATABYTE = ((digitalRead(DATA0)) + (digitalRead(DATA1) * 2) + (digitalRead(DATA2) * 4) + (digitalRead(DATA3) * 8) + (digitalRead(DATA4) * 16) + (digitalRead(DATA5) * 32) + (digitalRead(DATA6) * 64));
  //Serial.print(DATABYTE);
  //Serial.print(" ");
  Serial1.print(char(DATABYTE));
  delay(10);
}