Rorohiko Workflow Resources Wiki

This is my collection of interesting tidbits I collected while playing around with Arduino.

It is hard to get an in-depth understanding of Arduino and the various CPUs involved. The standard IDE involves a lot of 'magic' and there is not much help when you want to get a better understanding of how the magic works.

I had to do a lot of digging to figure out how to easily do assembler programming on the Arduino; below the results of my information quest.

This Arduino IDE is written in Java, and is merely a wrapper around the GNU toolchain for the AVR CPU range. The underlying toolchain has support for mixing assembler and C++ programs, but that functionality is not exposed through the IDE.

Only very minor changes are needed to the IDE source code to make it all work. I went through the exercise on Mac OS X and on Linux; things will be similar on Windows.

Before you can perform the procedure below, you need to have a proper Java development environment set up on your computer. You need to be able to call the Java compiler and the ant build tool from the command line.

These things were already set up on my Mac and Ubuntu machines when I started my quest, so I did not do any research into where they came from - not sure whether they're installed after a standard install or not. If not, they're fairly easy to come by.

On Mac, I imagine you might need to download and install Xcode from Apple (available on the App Store), or use fink or MacPorts. You might also need to download and install a Java JDK from Oracle.

On Linux, you might need to use a few apt-get or yum commands to fetch the necessary tools.

Step 1: Get the source code for the Arduino IDE. At the time of this writing, the source code is available in a git repository located at:

https://github.com/arduino/Arduino

Download the .zip archive with the IDE source code.

Expand the archive - you should end up with a directory called 'Arduino-master'.

First, go find the file called 'Sketch.java' in

wherever_you_put_it/Arduino-master/app/src/processing/app/Sketch.java

The problem is that this code does not recognize “S” (uppercase S) as a valid file name extension. So, you need to adjust the code to accept “S” in addition to “c”, “h”, and so on.

At the time of this writing, the first of the areas to change are around line 1455 (this is version 1.02 of the Arduino IDE).

We add a lower case “s” here - as far as I can tell, the file name being tested has already been converted to lower case, so the extension to look for is “s” even though the actual file name has an upper case “S”.

...
    // 3. then loop over the code[] and save each .java file

    for (SketchCode sc : code) {
      if (sc.isExtension("c") || sc.isExtension("cpp") || sc.isExtension("h")) {
        // no pre-processing services necessary for java files
...

needs to be come

...
    // 3. then loop over the code[] and save each .java file

    for (SketchCode sc : code) {
      if (sc.isExtension("c") || sc.isExtension("cpp") || sc.isExtension("h") || sc.isExtension("s")) {
        // no pre-processing services necessary for java files
...

and around line 1871

...
 /**
   * Returns a String[] array of proper extensions.
   */
  public String[] getExtensions() {
    return new String[] { "ino", "pde", "c", "cpp", "h" };
  }
...

needs to become

...
 /**
   * Returns a String[] array of proper extensions.
   */
  public String[] getExtensions() {
    return new String[] { "ino", "pde", "c", "cpp", "h", "s" };
  }
...

Then start a command-line session and navigate into the build directory, and build the patched IDE

cd wherever_you_put_it/Arduino-master/build
ant build
ant dist

For the version number, I entered '0102asm' - i.e. the source code repository I got from the GitHub was version 1.02, and I added 'asm' to the version number to remind me that I am not running a standard IDE.

After that, I ended up with a patched distribution in

wherever_you_put_it/Arduino-master/build/macosx/arduino-0102asm-macosx.zip

I then decompressed this (which gave me Arduino.app) and then moved this patched IDE into my /Applications folder instead of the 'official' downloadable version.

On Linux it works pretty much the same way; I haven't tried it on Windows, but I imagine it to be similar too.

The mpide IDE used for the Uno32 (which is Arduino-like, but uses a 32-bit PICX32 processor instead of an 8-bit AVR processor) is based on the Arduino IDE, but modified to support the Uno32. You can make similar changes to this source code too.

As a first exercise, I rebuilt the Blink example in assembler.

Here's how to do it: create a new sketch. In the sketch folder, add two text files: one called 'asmtest.h' and another called 'asmtest.S'.

In asmtest.h put:

/*
 * Global register variables.
 */
#ifdef __ASSEMBLER__

/* Assembler-only stuff */

#else  /* !ASSEMBLER */

/* C-only stuff */

#include <stdint.h>

extern "C" uint8_t led(uint8_t);
extern "C" uint8_t asminit(uint8_t);

#endif /* ASSEMBLER */

This defines the assembler routines in such a way that they can be called from a C/C++ program. To avoid issues with C++ name mangling, I defined the functions as extern “C” - this tells the C/C++ compiler that the underlying function is using pure C calling conventions as opposed to C++ calling conventions, and hence does not need name mangling.

In the file asmtest.S we get the assembler code:

#include "avr/io.h"
#include "asmtest.h"

; Define the function asminit()
.global asminit
asminit:
sbi  4,5; 4 = DDRB (0x24 - 0x20). Bit 5 = pin 13
ret

; Define the function led()
.global led ; The assembly function must be declared as global
led:
cpi r24, 0x00 ; Parameter passed by caller in r24
breq turnoff
sbi 5, 5; 5 = PORTB (0x25 - 0x20). Bit 5 = pin 13
ret
turnoff:
cbi 5, 5; 5 = PORTB (0x25 - 0x20). Bit 5 = pin 13
ret

The code took me a bit of fishing through the instruction set for the AVR processor. PORTB is in location 0x25, but when using the sbi (set bit immediate) or cbi (clear bit immediate) instructions you need to subtract 0x20 from that. Bit 5 of the PORTB byte corresponds to pin 13 of the Arduino board.

Finally the main sketch code in the .ino file is:

#include "asmtest.h"

void setup()
{
  asminit(0);
}

void loop()
{
    led(0);
    delay(1000);
    led(1);
    delay(1000);
}

In other words, I still define the setup() and loop() functions, and these then call into my assembler functions.

This first test is only 'half-assembler' - we still have some C/C++ backbone, but the difference in size is significant already. A standard 'Blink' sketch compiles to 1084 bytes. My assembler version is only 582 bytes.

Assembler is most often not the right language to do things in, as the most expensive resource is often the developer's time, and doing things in assembler is slower, more error-prone, and less efficient than C. However, when memory or CPU time is tight, using assembler can result in substantial space savings and speed improvements.

My next exercise was rebuilding the BlinkWithoutDelay example, my first attempt reducing code size from 1028 bytes to 580 bytes. The assembler part is the same as in the Blink example.

asmtest.h:

/*
 * Global register variables.
 */
#ifdef __ASSEMBLER__

/* Assembler-only stuff */

#else  /* !ASSEMBLER */

/* C-only stuff */

#include <stdint.h>

extern "C" uint8_t led(uint8_t);
extern "C" uint8_t asminit(uint8_t);

#endif /* ASSEMBLER */

asmtest.S:

#include "avr/io.h"
#include "asmtest.h"

.global asminit
asminit:
sbi  4,5; 4 = DDRB (0x24 - 0x20). Bit 5 = pin 13
ret

.global led ; The assembly function must be declared as global
led:
cpi r24, 0x01 ; Parameter passed by caller in r24
breq turnoff
sbi 5, 5; 5 = PORTB (0x25 - 0x20). Bit 5 = pin 13
ret
turnoff:
cbi 5, 5; 5 = PORTB (0x25 - 0x20). Bit 5 = pin 13
ret

sketch.ino:

#include "asmtest.h"

int x = 0;
int on = 1;

void setup()
{
  asminit(0);
}

void loop()
{
    int y;
    if ((y = millis()) - x > 0)
    {
      x = y + 1000;
      on = -on;
      led(on);
    } 
}

The next step I tried was to build a version of BlinkWithoutDelay that only uses assembler. This is what I came up with:

Make a new sketch, and empty the complete .ino file. You still need to have it, but it should be empty.

Then also create a file called 'asmtest.S' in the same directory as the (empty) .ino file:

#include "avr/io.h"

#define yl r28
#define yh r29

.global setup
setup:
  sbi  _SFR_IO_ADDR(DDRB), DDB5 ; Bit 5 = pin 13
  ret

// const long delay = 1000;
#define delay 1000 // ms

.global loop
loop:

  push yl
  push yh

  call millis   ; call millis(): 4-byte return value in r25...r22
  
  // Use Y as a pointer to fetch the next time to switch the LED
  ldi yl, lo8(nextSwitchAfterMillis) 
  ldi yh, hi8(nextSwitchAfterMillis)
  
  ld r18, y+
  ld r19, y+
  ld r20, y+
  ld r21, y+
  
  ld r17, y // ledStatus comes immediately after lastMillis, so we can use y
  
  // Compare nextSwitchAfterMillis with value returned by millis()
  sub r18, r22
  sbc r19, r23
  sbc r20, r24
  sbc r21, r25
  
  brcc tooEarly ; carry is set if r18...r21(nextSwitchAfterMillis) < r22...r25(millis())

  // Toggle LED state: 0 -> 1, 1 -> 0  
  inc r17
  andi r17, 1
  
  // Store ledStatus for next time. y still points at its memory location
  st y, r17
  
  // set LED state 
  brne turnoff
  
  cbi _SFR_IO_ADDR(PORTB), PORTB5; Bit 5 = pin 13
  rjmp ledSwitched  
  
turnoff:
  sbi _SFR_IO_ADDR(PORTB), PORTB5; Bit 5 = pin 13
  
 
ledSwitched:
  // Add long delay; to result of call to millis()
  ldi r17, lo8(delay)
  add r22, r17
  ldi r17, hi8(delay)
  adc r23, r17
  ldi r17, hlo8(delay)
  adc r24, r17
  ldi r17, hhi8(delay)
  adc r25, r17
  
  // Store this as the next point in time when we need to toggle the LED
  st -y, r25
  st -y, r24
  st -y, r23
  st -y, r22
  
tooEarly: 
  pop yh
  pop yl
  ret

.data 

nextSwitchAfterMillis:
.long 0

ledStatus:
.byte 0

A few tidbits:

- I changed the sbi 4,5 and similar to something like sbi _SFR_IO_ADDR(DDRB), DDB5 using predefined symbols that are defined in the “avr/io.h” include file, so the assembler code better expresses what it does. Underneath, it's still exactly the same - so there is no cost to doing this, but the code becomes more self-explanatory.

- I defined the setup() and loop() functions in assembler instead of in C. The Arduino 'wrapper' that is automatically compiled and linked in together with my code defines both setup() and loop() as 'extern “C”' routines, so the Arduino 'runtime' will find these routines, even though they're defined in assembler instead of C.

- I am calling millis() from assembler. This routine returns a 4-byte long; the assembler routine uses this long for comparison and for calculations. The millis() routine uses r25..r22 to return the long value, which are the standard AVR calling conventions.

- I am using the Y register (composed of r29..r28) as a 'pointer' into memory, using post-increment and pre-decrement to access a sequence of 5 bytes. 4 bytes are used for the millis value when the LED will be toggled, and another byte to contain the LED's current state in bit 0.

- I learned the hard way you need to save and restore the contents of r29..r28 when you clobber them. Hence the push… and pop… of yh/yl at the start and end of the loop routine.

- This version of the routine takes 576 bytes, so it only saved 4 extra bytes from the previous 'hybrid C++/asm' version.

- I also tried compiling a sketch with an empty loop() and setup() (both composed of just a 'ret' assembler instruction). Such a sketch takes 466 bytes. The two 'ret' instructions are 4 bytes, so the Arduino minimal 'runtime' weighs in at 462 bytes. That means that my last BlinkWithoutDelay needs about 576 - 466 = 110 bytes.

I'm working my way through the tutorials at www.mindkits.co.nz:

http://www.mindkits.co.nz/tutorials

as I ordered their ready-made tutorial stuff. I haven't progressed very far yet, as I keep getting sidetracked on all kinds of interesting thing, like assembler-programming tricks.

My next subject is an assembler version for part of Tutorial#0, where you wire up 8 LEDs to pins 2-9 of the Arduino and then make the LEDs light up in sequence, from left to right, then back from right to left, and so on.

Look for Exercise 0.1 on

http://www.mindkits.co.nz/tutorials/arduino_tutorials/Tutorial-0

First I did a C version, which is pretty simple. I have a variable named 'step' which is either +1 (when going 'up') or -1 (when going 'down'). Start at the first LED, add the step, until you hit the 'upper' LED, then change the sign on the step (+1 becomes -1, -1 becomes +1), continue until you hit the 'lower' LED, change the sign on the step, and so on…

This initial version never returns from the loop routine - which is not as tidy, but it works. Total size of compiled sketch: 1150 bytes.

/*
  Blink
  Turns on an LED on for one second, then off for one second, repeatedly.
 
  This example code is in the public domain.
 */
 
// Pin 13 has an LED connected on most Arduino boards.
// give it a name:
#define ledFrom  2
#define ledTo 9
#define ledJump (ledTo - ledFrom)

char cur;
char step;
char destLed;

// the setup routine runs once when you press reset:
void setup() {                
  // initialize the digital pin as an output.
  for (char i = ledFrom; i <= ledTo; i++)
  {
    pinMode(i, OUTPUT);
  }
  cur = ledFrom;
  step = 1;
  destLed = ledTo;
}

// the loop routine runs over and over again forever:
void loop() {
  if (cur == destLed)
  {
    if (step > 0)
    {
      destLed -= ledJump;
    }
    else
    {
      destLed += ledJump;
    }
    step = -step;
  }  
  digitalWrite(cur, LOW);   // turn the LED on (HIGH is the voltage level)
  cur += step;
  digitalWrite(cur, HIGH);   // turn the LED on (HIGH is the voltage level)
  delay(100);
}

Then I set out to do as assembler-only version. In this version, the .ino file remains empty, and all you do is add a file 'knightrider.S' to the sketch on a second tab. You need to use the patched IDE for this - standard IDEs won't recognize the .S file.

#include "avr/io.h"

#define yl r28
#define yh r29

.global setup
setup:
  push yl
  push yh
  
  ldi yl, lo8(DDRD)
  ldi yh, hi8(DDRD)
  ld r25, y // DDRD: bit 2 = pin 2 .. bit 7 = pin 7
  ori r25, 0xFC
  st y, r25
  ld r25, y // DDRB: bit 0 = pin 8, bit 1 = pin 9
  ldi yl, lo8(DDRB)
  //ldi yh, hi8(DDRB) // is same, no need to reload
  ori r25, 0x03
  st y, r25
  
  pop yh
  pop yl  
  
  ret

// const long delay = 1000;
#define delay 50 // ms

.global loop
loop:

  push yl
  push yh

  call millis   ; call millis(): 4-byte return value in r25...r22
  
  // Use Y as a pointer to fetch the next time to switch the LED
  ldi yl, lo8(nextSwitchAfterMillis) 
  ldi yh, hi8(nextSwitchAfterMillis)
  
  ld r18, y+
  ld r19, y+
  ld r20, y+
  ld r21, y+
  
  // ldi yl, lo8(shifters) not needed; y is already correct 
  // ldi yh, hi8(shifters)
  ld r17, y+ 
  ld r16, y
  
  // Compare nextSwitchAfterMillis with value returned by millis()
  cp r22, r18
  cpc r23, r19
  cpc r24, r20
  cpc r25, r21
  
  brcs tooEarly ; carry is clear if r18...r21(nextSwitchAfterMillis) < r22...r25(millis())

  // Rotate left & right; bit travels up then down through r17 and r16
  // Carry is clear, no need for CLC
  rol r17
  sbrc r16,0
  ori r17,1
  ror r16

  st y, r16
  st -y, r17

  // Combine left and right shifter
  or r17,r16
 
  // 6 high bits of port D
  ldi yl,lo8(PORTD)
  ldi yh,hi8(PORTD)
  mov r16, r17
  lsl r16 // Zeroes 2 low bits of r16
  lsl r16
  st y,r16
  
  // 2 low bits of port B
  ldi yl,lo8(PORTB)
  //ldi yh,hi8(PORTB) // is already same
  ldi r16,6
 shift6:
  lsr r17 // Zeroes high bits of r17
  dec r16
  brne shift6
  st y,r17
  
   // Add delay to result of call to millis()
  ldi r17, lo8(delay)
  add r22, r17
  ldi r17, hi8(delay)
  adc r23, r17
  ldi r17, hlo8(delay)
  adc r24, r17
  ldi r17, hhi8(delay)
  adc r25, r17
  
  // Store this as the next point in time when we need to toggle the LED
  ldi yl, lo8(nextSwitchAfterMillis) 
  ldi yh, hi8(nextSwitchAfterMillis)
  
  st y+, r22
  st y+, r23
  st y+, r24
  st y+, r25
  
tooEarly: 
  pop yh
  pop yl
  
  ret

.data 

nextSwitchAfterMillis:
.long 0

shifters:
.byte 1
.byte 0

This one weighs in at 630 bytes.

The trick I used is to have two 8-bit values, where one is being rotated right-to-left, and the other is being rotated left-to-right. The rol and ror instructions rotate through the carry flag.

I.e. whatever bit 'falls out' of the register being rotated falls into the carry flag, and the original contents of the previous carry flag are rotated 'into' the register. rol and ror are effectively 9-bit rotations, where the Carry flag is the 9th bit.

The two 8-bit values are kept in r17 and r16; looking at the two 8-bit values you'd see successive states like shown below, because of the rol r17 and ror r16 instructions:

Step 1:

r17:00000001
r16:00000000

Step 2:

r17:00000010
r16:00000000

Step 3:

r17:00000100
r16:00000000

Step 4:

r17:00001000
r16:00000000

...

Step 8:

r17:10000000
r16:00000000

Step 9: (the 1 bit rotates 'out' of r17 into the carry, then from the carry into the topmost bit of r16)

r17:00000000
r16:10000000

Step 10:

r17:00000000
r16:01000000

Step 11:

r17:00000000
r16:00100000

...

Step 16:

r17:00000000
r16:00000001

Step 17: The 1 bit rotates 'out' of r16 into the carry, and disappears from view. However, before
that happens, the sbrc r16,0 instruction tests for the '1' bit being in the bit-position 0 of r16,
and if it is, executes an ori r17, 1 - effectively re-instating the bit into r17, ready for another
up-down round.

r17:00000001
r16:00000000

So, we now have this single bit, marching endlessly around the two 8-bit values. Round and round it goes.

The second trick is to then make the 'logical or' of these two values. That gives us a single-byte end-result where the bit seems to go back and forth all the time. That's the value we will use to drive our LEDs.

Before we can use this 8-bit value, we need to do some more shifting, because the LED's are driven from two different ports: 6 LEDs are driven by bit 2-7 of PORTD, and 2 LED are driven by bit 0-1 of PORTB. That's what all the lsl and lsr stuff is about.

The rest of the code is very similar to our previous BlinkWithoutDelay as far as calculating millis and so on - so this version returns properly from the 'loop' routine each time it is called; it does not get 'stuck'.

Now, I thought I should be able to do better, so I rewired the LEDs in such a fashion that I could drop those extra shifts.

Led 1 = pin 8 PORTB bit 0
Led 2 = pin 9 PORTB bit 1
--
Led 3 = pin 2 PORTD bit 2
Led 4 = pin 3 PORTD bit 3
Led 5 = pin 4 PORTD bit 4
Led 6 = pin 5 PORTD bit 5
Led 7 = pin 6 PORTD bit 6
Led 8 = pin 7 PORTD bit 7

This way, I do not need to shift the bits around any more - the bottommost two bits of PORTB are driven by bits 0 and 1 of my calculated value, and the topmost 6 bits of PORTD can be driven by bits 2-7, without needing any additional shifts after the rol/ror trick.

One of my goals is to get a 'feel' for how much it costs to use C instead of assembly. In order to have a useful comparison, I decided to rewire my board, then try writing a C version first, this time using direct port manipulation instead of directWrite.

directWrite is a nice abstraction, but there is quite a memory and timing-overhead attached to it, and direct port manipulation in C generates much more compact code.

So here is my C version for the rewired setup, no assembler involved.

#include <Arduino.h>

void setup()
{
  DDRD = DDRD | 0xFC; // Top 6 bits as outputs
  DDRB = DDRB | 0x03; // Bottom 2 bits as outputs
}

#define delay 50

long nextSwitchAfterMillis;
byte shiftUp = 0x01;
byte shiftDown = 0x00;

void loop()
{
  long curTime = millis();
  if (curTime > nextSwitchAfterMillis)
  {
    nextSwitchAfterMillis = curTime + delay;
    byte tempShift = shiftUp;
    shiftUp = (tempShift << 1) | (shiftDown & 1);
    shiftDown = (tempShift & 0x80) | (shiftDown >> 1);
    tempShift = shiftUp | shiftDown;
    PORTD = tempShift & 0xFC;
    PORTB = tempShift & 0x03;
  }
}

This one compiles to 610 bytes, and that surprised me - that is way smaller than I expected!

I then did some digging, and checked out the assembly language output of the C-compiler, and as a result, learned a few new tricks.

First trick: the AVR instruction set has no 'ADCI' or 'ADI' instruction. My solution was to load the value into a register and then use ADC or ADD.

The C-compiler has a much better trick up it's sleeve: the AVR instruction set does have a SBCI and SBI instruction, so instead of adding a constant value, it simply subtracts the negative of the value, and no intermediate register is needed to hold the immediate values.

I also completely overlooked the IN and OUT instructions, and instead was using memory addressing to access DDRB, DDRD, PORTB, PORTD as memory locations. The C compiler instead uses IN and OUT, which again saved a few bytes.

While I was looking over the instruction set I also noticed a few more instructions that allowed me to save some more bytes: I found STS and LDS, which allow addressing via the Y register but with an offset applied.

So, I finally came up with this:

#include "avr/io.h"

//
// Adjusted wiring, which allows us to suppress some shift
// instructions
//
// Led 1 = pin 8
// Led 2 = pin 9
// Led 3 = pin 2
// Led 4 = pin 3
// Led 5 = pin 4
// Led 6 = pin 5
// Led 7 = pin 6
// Led 8 = pin 7
//

#define yl r28
#define yh r29

// const long delay = 100;
#define delay 100 // ms

.global setup
setup:
  in r25, _SFR_IO_ADDR(DDRD)
  ori r25, 0xFC
  out _SFR_IO_ADDR(DDRD), r25
  
  in r25, _SFR_IO_ADDR(DDRB)
  ori r25, 0x03
  out _SFR_IO_ADDR(DDRB), r25

  ret

.global loop
loop:

  push yl
  push yh

  call millis   ; call millis(): 4-byte return value in r25...r22
  
  // Use Y as a pointer to fetch the next time to switch the LED
  ldi yl, lo8(nextSwitchAfterMillis) 
  ldi yh, hi8(nextSwitchAfterMillis)
  
  ld r18, y
  ldd r19, y+1
  ldd r20, y+2
  ldd r21, y+3
  
  ldd r17, y+4
  ldd r16, y+5
  
  // Compare nextSwitchAfterMillis with value returned by millis()
  cp r22, r18
  cpc r23, r19
  cpc r24, r20
  cpc r25, r21
  
  brcs tooEarly ; carry is clear if r18...r21(nextSwitchAfterMillis) < r22...r25(millis())

  // Rotate left & right; bit travels up then down through r17 and r16
  // Carry is already clear because brcs was not taken, so no need for CLC
  rol r17
  sbrc r16,0
  ori r17,1
  ror r16

  // Update the memory storage with the new rotated values
  std y+4, r17
  std y+5, r16

  // Combine left and right shifter
  or r17, r16
  
  // 6 high bits of port D
  mov r16, r17
  andi r16, 0xFC
  out _SFR_IO_ADDR(PORTD), r16
 
  // 2 low bits of port B
  andi r17,0x03
  out _SFR_IO_ADDR(PORTB), r17
  
  // Add delay (subtract negative delay because there is no addi/adci)
  // to result of call to millis()
  subi r22, lo8(-delay)
  sbci r23, hi8(-delay)
  sbci r24, hlo8(-delay)
  sbci r25, hhi8(-delay)
      
  st y, r22
  std y+1, r23
  std y+2, r24
  std y+3, r25
  
tooEarly: 
  pop yh
  pop yl
  
  ret

.data 

nextSwitchAfterMillis:
.long 0

shifters:
.byte 1
.byte 0

This last version is 590 bytes - 20 bytes less than the C compiler.

Conclusion: the C compiler is doing a pretty good job, and the extra effort of writing things in assembler is probably rarely worth it. Subtracting the 462 bytes for the runtime overhead from both compiled sketch sizes, we're looking at saving a little over 10% in size for this particular exercise.

Nevertheless, sometimes 20 bytes can be the difference between 'it fits' and 'it does not fit', and, more importantly, I love tinkering, so I'll do some more assembler, just because I can!

Now, just for kicks, a tiny change - change the last listing so it ends as follows:

...
shifters:
.byte 1
.byte 0x80

Run it again. Is that cool, or what?

Interesting links:

http://bitly.com/bundles/zwettemaan/3