Easy Raspberry Pi framebuffer screenshots with raspi2png

I gave a talk about retro-gaming on the Raspberry Pi yesterday. I was describing RetroPie, and I really needed lots of screenshots to illustrate games. I’m used to grabbing screens under X, but RetroPie runs without it, so all my usual tools were of no use.

I’d just found out about raspi2png, and it works really well! Usage is simple: just call it like

raspi2png -p outfile.png

and it’ll save whatever’s on the screen. It doesn’t play well with X, but there are already tools to take screenshots with that. As I was playing games, I didn’t want to have to pause the computer to take a shot, so I ran the command every five seconds for 30*5 seconds like this:

for f in {1..30}; do raspi2png -p retro_$(date -Iseconds).png; sleep 5; done

Here are some quality shots via raspi2png from Deathchase, officially the best ZX Spectrum game ever:

VM-CLAP1 👏 sensor + gpiozero on Raspberry Pi

Well, that was easy!

Since the Verbal Machines VM-CLAP1 sensor is an open collector type — that is, it sinks current when triggered — it behaves like a simple button to gpiozero, the Raspberry Pi Python GPIO library. If you attach a callback function to the sensor’s when_pressed event, your Python script will call that function every time it registers a clap.

The wiring is as simple as it could be:

 VM-CLAP1: Raspberry Pi:
 ========= =============
      GND → GND
      PWR → 3V3
      OUT → GPIO 4

This example code just prints clap! when the board picks up a 👏:

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

# Raspberry Pi gpiozero test for
# Verbal Machines VM-CLAP1 clap sensor
#   scruss - 2017-06
# Wiring:
#  VM-CLAP1:    Raspberry Pi:
#  =========    =============
#    GND     →   GND
#    PWR     →   3V3
#    OUT     →   GPIO 4

from gpiozero import Button
from signal import pause

def clapping():

clap = Button(4)
clap.when_pressed = clapping

This is a trivial example, but at least it shows that anything you can do with a button, you can also do with this hand-clap sensor.

Clap Sensor part for Fritzing

I have no idea if this wiring would work, but it’s pretty …

For work, I blogged about Verbal MachinesHand Clap Sensor VM-CLAP1 (“Clap on 👏👏, Clap off 👏👏 – Elmwood Electronics”). I’ve made a preliminary part for Fritzing for the board: VM-CLAP1 Clap Sensor for Fritzing.zip

It should work in Breadboard and Schematic mode, but absolutely doesn’t work in PCB mode. This shouldn’t be a problem, as it’s only available as a standalone board. Fritzing doesn’t have any way to create new parts from scratch any more, so I had to base it on a somewhat similar-looking board, the SparkFun Electret Microphone Breakout.

I’m looking forward to see what I can do with gpiozero and the clap sensor.

Headless Raspberry Pi + VNC: useful resolution

Breadboards of Canada …

I just set up a Raspberry Pi Zero to be a little breadboard computer. Running a headless machine only through SSH gets a bit dull at times, so the inclusion of VNC Connect in Raspbian is handy.

Only problem was that the default screen size — something like 720×480 — was too small for most dialogue windows. Here’s how to enable a more useful resolution of 1024 × 768.

All of these are enabled from the raspi-config tool, so open a terminal and start it with:

sudo raspi-config

Enable Boot to Desktop

Select 3 Boot OptionsB1 Desktop / CLIB4 Desktop Autologin:

Enable VNC on Startup

Select 5 Interfacing OptionsP3 VNC, and answer Yes to Would you like the VNC Server to be enabled?:

Set Screen Resolution

Select 7 Advanced OptionsA5 ResolutionDMT Mode 16 (1024×768) …:

Once you’ve enabled all of these, raspi-config will ask if you wish to reboot your Raspberry Pi. Once it has rebooted, you should have a usable remote desktop.

(All of these screenshots were taken from a headless Raspberry Pi Zero via VNC.)

Simple — like, really simple — Arduino periodic timer with Brett’s MillisTimer library

I don’t know how many times I’ve written bad Arduino code to call a function every few milliseconds. Sometimes this bad code works well enough for my sketch to actually work. Often, it either doesn’t work at all or does something I really didn’t expect.

So on Arduino Day 2017, I’m glad I found out about bhagman/MillisTimer: A Wiring and Arduino library for working with millis(). It couldn’t be simpler to use: include the library, write the function you want to call every N milliseconds, set up the timer to run every N millis, and put timer.run() in a loop that’s called frequently. The library handles the timing and resetting all by itself.

As an example, here’s the eternal “Hello, World!” of the embedded world, Blink, rewritten to use MillisTimer:

// MillisTimerBlink - blink LED every second
//  using Brett Hagman's MillisTimer library
//  https://github.com/bhagman/MillisTimer
//  (or use Sketch → Include Library → Manage Libraries … to install)
// scruss - 2017-04-01

#include <MillisTimer.h>
MillisTimer timer1;               // new empty timer object
const int led_pin = LED_BUILTIN;  // use the built-in LED

void flash() {                    // function called by timer
  static boolean output = HIGH;
  digitalWrite(led_pin, output);  // set LED on or off
  output = !output;               // toggle variable state High/Low

void setup() {
  pinMode(led_pin, OUTPUT);       // use built-in LED for output
  timer1.setInterval(1000);       // set timer to trigger every 1000 millis
  timer1.expiredHandler(flash);   // call flash() function when timer runs out
  timer1.setRepeats(0);           // repeat forever if set to 0
  timer1.start();                 // start the timer when the sketch starts

void loop() {
  timer1.run();                   // trigger the timer only if it has run out
  // note that run() has to be called more frequently than the timer interval
  //  or timings will not be accurate

Note that MillisTimer only triggers when timer.run() is called. Sticking a delay(2000) in the main loop will cause it to fire far less frequently than the interval you set. So it’s not technically a true periodic timer, but is good enough for most of my purposes. If you want a true interrupt-driven timer, use the MsTimer2 library. It relies on the timer interrupts built into the Arduino hardware, and isn’t quite as easy to use as MillisTimer.

Keypunch029 — for all your punched card font needs …

A fairly accurate rendition of the 5×7 dot matrix font printed at the
top of punched cards by the IBM Type 29 Card Punch (1965).

Local copy: Keypunch029.zip.
Fontlibrary link: Keypunch029

The 029 (as it is sometimes known) generated a bitmap font from an engraved metal plate pressing on a matrix of pins. A picture of this plate from a field engineering manual was used to re-create the pin matrices, and thus an outline font.

029 Code Plate
029 Code Key

Historical Accuracy

The 029 could have many different code plates, but the one used here contained the characters:


The character glyphs have been sized such that if printed at 12 points, the 029’s character pitch of 0.087″ is accurately reproduced. No attempt to research the pin matrix pitch or pin diameter has been made: the spacing was eyeballed from a couple of punched cards in my collection.

The earlier IBM Type 26 Card Punch (“026”) included a glyph for a square lozenge (Unicode U+2311, ⌑). The 029 code plate did not include this character, but I added it here for completeness.

The character set was extended to include:

  • all of ASCII, with lower case characters repeating the upper case glyphs;
  • sterling currency symbol; and
  • euro currency symbol.

While there may have been official IBM renditions of some of these additional glyphs (with the exception of euro) no attempt has been made to research the original shapes. This font set is intended to help with the visually accurate reproduction of 1960s-era punched cards, mostly coinciding with my interest in the FORTRAN programming language. No attempt has been made to use historical BCD/EBCDIC encodings in these fonts. We have Unicode now.

The 029 card punch could not produce any bold or italic font variants, but FontForge can, so I did.

Things I learned in making these fonts

  1. The 029 card punch printer could be damaged if you tried to print binary cards, as there was no way to disengage the code plate from the punch mechanism.
  2. FontForge really hates to have paths in a glyph just touching. Either keep them more than one unit apart, or overlap them and merge the overlapping paths.
  3. EBCDIC is weird.


In the unlikely event you need to represent Emoji in RTF using Perl …

Of all the niche blog entries I’ve written, this must be the nichest. I don’t even like the topic I’m writing about. But I’ve worked it out, and there seems to be a shortage of documented solutions.

For the both of you that generate Rich Text Format (RTF) documents by hand, you might be wondering how RTF converts ‘💩’ (that’s code point U+1F4A9) to the seemingly nonsensical \u-10179?\u-9047?. It seems that RTF imposes two encoding limitations on characters: firstly, everything must be in 7-bit ASCII for easy transmission, and secondly, it uses the somewhat old-fashioned UTF-16 representation for non-ASCII characters.

UTF-16 grew out of an early standard, UCS-2, that was all like “Hey, there will never be a Unicode cope point above 65536, so we can hard code the characters in two bytes … oh shiiii…”. So not merely does it have to escape emoji code points down to two bytes using a very dank scheme indeed, it then has to further escape everything to ASCII. That’s how your single enoji becomes 17 bytes in an RTF document.

So here’s a tiny subroutine to do the conversion. I wrote it in Perl, but it doesn’t do anything Perl-specific:

#!/usr/bin/env perl
# emoji2rtf - 2017 - scruss
# See UTF-16 decoder for the dank details
#  <https://en.wikipedia.org/wiki/UTF-16#U.2B10000_to_U.2B10FFFF>

use v5.20;
use strict;
use warnings;
use utf8;
sub emoji2rtf($);

my $c = substr( $ARGV[0], 0, 1 );
say join( "\t⇒ ", $c, sprintf( "U+%X", ord($c) ), emoji2rtf($c) );

sub emoji2rtf($) {
    my $n = ord( substr( shift, 0, 1 ) );
    die "emoji2rtf: code must be >= 65536\n" if ( $n < 0x10000 );
    return sprintf( "\\u%d?\\u%d?",
        0xd800 + ( ( $n - 0x10000 ) & 0xffc00 ) / 0x400 - 0x10000,
        0xdC00 + ( ( $n - 0x10000 ) & 0x3ff ) - 0x10000 );

This will take any emoji fed to it and spit out the RTF code:

📓	⇒ U+1F4D3	⇒ \u-10179?\u-9005?
💽	⇒ U+1F4BD	⇒ \u-10179?\u-9027?
🗽	⇒ U+1F5FD	⇒ \u-10179?\u-8707?
😱	⇒ U+1F631	⇒ \u-10179?\u-8655?
🙌	⇒ U+1F64C	⇒ \u-10179?\u-8628?
🙟	⇒ U+1F65F	⇒ \u-10179?\u-8609?
🙯	⇒ U+1F66F	⇒ \u-10179?\u-8593?
🚥	⇒ U+1F6A5	⇒ \u-10179?\u-8539?
🚵	⇒ U+1F6B5	⇒ \u-10179?\u-8523?
🛅	⇒ U+1F6C5	⇒ \u-10179?\u-8507?
💨	⇒ U+1F4A8	⇒ \u-10179?\u-9048?
💩	⇒ U+1F4A9	⇒ \u-10179?\u-9047?
💪	⇒ U+1F4AA	⇒ \u-10179?\u-9046?

Just to show that this encoding scheme really is correct, I made a tiny test RTF file unicode-emoji.rtf that looked like this in Google Docs on my desktop:

It looks a bit better on my phone, but there are still a couple of glyphs that won’t render:

“The Error Message as a bourgeois construct”

If you try to run the (unmodified) BASIC code for Oregon Trail (1975) on PDP-8 BASIC, you get this:

 DI 30
 XC 45
 XC 205
 IF 700
 NM 730
… (many, many more lines …)

I thought at first it was a stack trace, but nope — it’s error messages! You need to dig through your trusty language manual, and on page 132 it has a table to explain:


(and yes, they’re in all-caps. Mixed case? Mixed feelings!)

So whenever Python throws a tantrum (or as it calls it, an exception) and wails at length about its problems, remember PDP-8 BASIC: Two letters + a line number. That’s all.

PoorFish, v2

On FontLibrary: PoorFish

Local copy: PoorFish.zip

Full Language Support: Afrikaans, Baltic, Basic Latin, Catalan, Central European, Dutch, Esperanto, Euro, Turkish, Western European. Terrible kerning comes free.

(original version from 2010: PoorFish)

The Pocket DEC Pretender (PDP) Zero

PDP (Pocket DEC Pretender) Zero: lettering came out a bit more, um,  artisanal than I’d hoped …

Digital (aka DEC) used to make some very solid minicomputers back when a minicomputer was fridge-sized and people were still building nuclear power stations to be controlled by them. The Raspberry Pi Zero is a very mini computer indeed, and in USB gadget mode running SimH it makes a nice little emulation platform.

The case is from Thingiverse: One Piece Raspberry Pi Zero + Camera Case (with GPIO) by Superrei, but with the DEC PDP logo in relief on the top.

DEC minis were famous for their arrays of blinkenlights. The Pocket DEC Pretender, not so much: it has one tiny green light that flickers a bit now and again:

PDP (Pocket DEC Pretender) Zero: case open, very few blinkenlights

But it’s a genuinely useful (for my values of useful) emulation platform. Here it is pretending to be a PDP-8, running BASIC under OS-8:

PDP (Pocket DEC Pretender) Zero: PDP-8 BASIC!

(background in case pictures woven in Toronto by Deftly Weft)

TYPE BANG: First Person Shooter, 1975 style

— from the source code of an early (1975) time-shared system version of The Oregon Trail, as documented in On the Trail of the Oregon Trail.

Library Hand – Disjoint

LibHandDis — Based on scans of “Library Hand – Disjoint”, described in Dana’s A Library Primer, with some modifications.

Major changes from scan:

  • As the scan only covered A-Z, a-z, 0-9 and ‘&’, I had to make the rest up.
  • Many of the descenders had to be shortened to fit with modern typography conventions.
  • Kerning is much tighter than Dana’s guidelines suggest.

(idea for this came via MetaFilter, This question of library handwriting is an exceedingly practical one)

Local copy: LibHandDis.zip.

a font for the person you’re just dotty about

LoveMatrix is a lo-fi dot matrix font made of ♥♥♥s. It’s a seasonally-adjusted version of my mnicmp font.

Local copy: LoveMatrix.zip

awaiting surgery

Seems I have two Macintosh PowerBook 180s: one has a dead colour screen but seems to boot fine, and the other (the one screenshotted above; yay ⌘+Shift+3 and enough room on the boot floppy …) has a lovely greyscale screen but a dead hard drive. I suspect we’re going to have to do a head transplant.

And no, I’m not having ¼-century 68030+68882 wish fulfillment one bit

Apple II on Raspberry Pi

C’mon let’s all die of dysentery on the Oregon Trail!

Building and installing the linapple-pie Apple IIe emulator is relatively easy on the Raspberry Pi:

sudo apt install libcurl4-openssl-dev libzip-dev zlib1g-dev libsdl1.2-dev libsdl-gfx1.2-dev libsdl-image1.2-dev libsdl-sound1.2-dev build-essential git
git clone https://github.com/dabonetn/linapple-pie.git
cd linapple-pie/src
sudo make install

This also works on an x86_64 Ubuntu machine. It does also install on a PocketCHIP (even if it takes a really long time) but I can’t get the display resolution to fit correctly.

Raspberry Pi Zero: avoiding random network addresses in Gadget Mode

The Raspberry Pi Zero can be set up to appear as one of several USB OTG “gadgets” if you plug it into another computer. The most popular setting seems to be the virtual network gadget that turns your Zero into a computer on the end of your USB cable. Andrew Mulholland’s guide Raspberry Pi Zero – Programming over USB! (Part 2) (along with his super-short simple guide) seems to be the definitive source on how to set these modes up.

One problem, though, is that the Zero would show up on different network addresses every time it was restarted. The changing addresses made ssh access no fun at all. A suggestion on the Raspberry Pi forum helped me come up with a solution. On the Raspberry Pi Zero, run this command once:

echo 'options g_ether host_addr='$(dmesg | awk '/: HOST MAC/{print $NF}')' dev_addr='$(dmesg | awk '/: MAC/{print $NF}') | sudo tee /etc/modprobe.d/g_ether.conf

This will set the USB port’s hardware addresses to a fixed value, and you should always get a connection on the same IP address if it’s available.

How my Raspberry Pi Zero appears on my Ubuntu machine

Update: For some reason, this seemed to stop working, and I was getting the old random addresses again. I was resisting putting more stuff in /boot/cmdline.txt, but it seems to me it’s more reliable than what I proposed. So if your g_ether.conf looked like:

options g_ether host_addr=66:e0:94:40:44:74 dev_addr=de:82:66:79:e2:66

right at the end of your /boot/cmdline.txt (which is all one line, btw), add

 g_ether.host_addr=66:e0:94:40:44:74 g_ether.dev_addr=de:82:66:79:e2:66

You’ll probably also want to share the IPv4 connection so your Raspberry Pi Zero can use the network of the device it’s attached to:

Emulating a Mac IIx on Raspberry Pi

Yup, I have a failing for 512×384 screens …

Running Mini vMac on a Raspberry Pi is hardly news. But maybe running it as a colour Mac II is. The screen size I’ve chosen is closer to a Color Classic, for no other reason that I like it.

To build a Mac II-capable version of Mini vMac, you’ll need the Alpha source code. You’ll also need a working Mini vMac setup, as it uses a 68k Mac program to set up the source. Pretty much any basic setup and bootable disk will run this okay:

Mini vMac building on an emulated Mac Classic booting from the System 7 Network Access floppy image (no, I couldn’t boot from Classic’s hidden boot ROM disk)

The build options I chose are:

-t larm -ccs -m II -hres 512 -vres 384 -depth 3 -mf 2 -magnify 1 -em-cpu 2 -mem 8M

I’ve chosen to swap the Ctrl key with the Command (⌘) key, as most non-Mac keyboards work better with this.

The build program will export a file out/minivmac-3.5.0-larm.tar that you can unpack into the full source code. It’s a really simple build, and fast, too.

Now you’ll need a Mac IIx ROM image (which I’m not supposed to help you find, but it’s an easy search) and OS image disks from the Mini vMac System Software page. Have fun!

Fun things you learn from old computers …

The program on the left is running on the decimal interpreter, the one on the right the regular one

Microsoft used to supply two versions of its BASIC for Macintosh. One used decimal mathematics for precise tallying of small amounts. The other used the more familiar floating point mathematics, rounding errors and all. I don’t know what floating point library Microsoft used for sure — perhaps Motorola’s 32-bit Fast Floating Point system — but it introduces rounding errors pretty quickly. Modern routines don’t start displaying oddly until after 15 decimal places.

Consider this simple program:

10 LET x=36/10
20 LET a$="## #.#"
30 FOR n%=1 TO 18
40 PRINT USING a$; n%; x
50 LET a$=a$+"#"
60 NEXT n%
70 END

Along with the number of decimal places, it should print 3.6, 3.60, 3.600, 3.6000, … with an increasing line of zeroes after the 3.6. Bas makes a valiant but typical attempt:

 1 3.6
 2 3.60
 3 3.600
 4 3.6000
 5 3.60000
 6 3.600000
 7 3.6000000
 8 3.60000000
 9 3.600000000
10 3.6000000000
11 3.60000000000
12 3.600000000000
13 3.6000000000000
14 3.60000000000000
15 3.600000000000000
16 3.6000000000000001
17 3.60000000000000009
18 3.600000000000000089

Oddly enough, good old Locomotive BASIC on the Amstrad CPC does it more correctly somehow:

So the variables, they vary.

cbmbasic for BASIC Computer Games

Immersive it ain’t, but you have to remember it was 1978 …

cbmbasic is pretty cool. It’s a portable C rendition of the Commodore 64’s ROM BASIC interpreter. While not the spiffiest version of the language, it does allow some very old code to run — such as the games from David H. Ahl’s book BASIC Computer Games.

Here are all the programs automatically converted to cbmbasic’s tokenized format: cbmbasic-Ahl-BASIC_Games. They seem to run, but some might fail. Notes on sources of the text files and conversion methods are in the archive. Have fun!

Mac Classic II rev 2 re-capping

Since the 68kMLA wiki page on Capacitor Replacement doesn’t have it, and also doesn’t seem to be accepting new edits, here’s what you need to replace the leaky capacitors on a revision 2 Macintosh Classic II motherboard (part 820-0326-B):

  • 2× 1 µF, 50 V rating (C9, C15)
  • 3× 47 µF, 16 V rating (C3, C4, C13)
  • 12× 10 µF, 16 V rating (C5, C6, C7, C8, C10, C11, C12, C14, C21, C74, C79, C106)

You’ll probably also be looking for a 3.6 V ‘½ AA’/14250 lithium battery too, as if it hasn’t leaked in the 25 years since your Classic II was made it’ll be completely flat. These can be a bit pricey and hard to find. I got one at Sayal for nearly $10.

The revision 2 Classic II board is immediately identifiable by having only two ROM sockets where the first revision has four ROMs.