Category Archives: Hardware

Arduino and LCD2004 and PCF8574 and I2C – part 1

See also: part 1.

NOTE: This article was originally written two years ago, and meant to be part of a series. I never got around to writing Part 2, so I am just publishing this initial part by itself. If there is interest, I will continue the series. My Github actually shows the rest of the work I did for my “full” and “small” version of the drive code for this LCD.

Recently, my day job presented me an opportunity to play with a small 20×4 LCD display that hooked up via I2C. The module was an LCD2004. The 20 is the number of columns and the 04 is the number of rows. The LCD1602 would be a 16×2 display.

While I have found many “tutorials” about these displays, virtually all of them just teach you how to download a premade library and use library functions. Since I was going to be implementing code for an in-house project, and did not have room for a full library of functions I would not be using, I really needed to know how the device worked. Hopefully this article may help others who need (or just want) to do what I did.

LCD2004 / LCD1602 / etc.

These LCD modules use a parallel interface and require eleven I/O pins. The pinout on the LCD looks like this:

1 - VSS (Ground)
2 - VDD (5V)
3 - V0 (LCD Contrast)
4 - RS (Register Select: LOW=Instruction, HIGH=Data)
5 - RW (Read/Write: LOW=Write, HIGH=Read)
6 - E (Enable)
7 - D0 (Data lines, 8-bit)
8 - D1
9 - D2
10 - D3
11 - D4
12 - D5
13 - D6
14 - D7
15 - A/LED+ (Backlight Power)
16 - K/LED- (Backlight Ground)

A few of the pins are listed by different names based on whoever created the data sheet or hardware. On my LCD2004 module, pins 15 and 16 are listed as A and K, but I now know they are just power lines for the backlight.

If you have something like an Arduino with enough available I/O pins, you can wire the display up directly to pins. You should be able to hook up power (5V to VDD, Ground to VSS, and probably some power to the backlight and maybe something to control contrast), and then connect the eight data lines (D0-D7) to eight available digital I/O pins on the Arduino.

The LCD module has a simple set of instruction bytes. You set the I/O pins (HIGH and LOW, each to represent a bit in a byte), along with the RS (register select) and RW (read/write) pins, then you toggle the E (Enable) pin HIGH to tell the LCD it can read the I/O pins. After a moment, you toggle E back to LOW.

The data sheets give timing requirements for various instructions. If I read it correctly, it looks like the E pin needs to be active for a minimum of 150 nanoseconds for the LCD to read the pins.

Here is a very cool YouTube video by Ian Ward that shows how the LCD works without using a CPU. He uses just buttons and dip switches. I found it quite helpful in understanding how to read and write to the LCD.

If you don’t have 11 I/O pins, you need a different solution.

Ian Ward’s excellent LCD2004 video.

A few pins short of a strike…

If you do not have eleven I/O pins available, the LCD can operate in a 4-bit mode, needing only four pins for data. You send the upper four bits of a byte using the E toggle, followed by the lower 4-bits of the byte. This is obviously twice as slow, but allows the part to be used when I/O pins are limited.

If you don’t have 7 I/O pins, you need a different solution.

PCF8574: I2C to I/O

If you do not have seven I/O pins available, you can use the PCF8574 chip. This chip acts as an I2C to I/O pin interface. You write a byte to the chip and it will toggle the eight I/O pins based on the bits in the byte. Send a zero, and all pins are set LOW. Send a 255 (0xff) and all pins are set HIGH.

Using a chip like this, you can now use the 2-wire I2C interface to communicate with the LCD module–provided it is wired up and configured to operate in 4-bit mode (four pins for data, three pins for RS, RW and E, and the spare pin can be used to toggle the backlight on and off).

Low-cost LCD controller boards are made that contain this chip and have pins for hooking up to I2C, and other pins for plugging directly to the LCD module. For just a few dollars you can buy an LCD module already soldered on to the PCF8574 board and just hook it up to 5V, Ground, I2C Data and I2C Clock and start talking to it.

If you know how.

I did not know how, so I thought I’d document what I have learned so far.

What I have learned so far.

The PCF8574 modules I have all seem to be wired the same. There is a row of 16-pins that aligns with the 16 pins of the LCD module.

PCF8574 module.

One LCD I have just had the board soldered directly on to the LCD.

LCD2004 with the PCD8574 module soldered on.

Another kit came with separate boards and modules, requiring me to do the soldering since the LCD did not have a header attached.

PCF8574 module and LCD1602, soldering required.

If you are going to experiment with these, just get one that’s already soldered together or make sure the LCD has a header that the board can plug in to. At least if you are like me. My soldering skills are … not optimal.

The eight I/O pins of the PCF modules I have are connected to the LCD pins as follows:

1 - to RS
2 - to RW
3 - to E
4 - to Backlight On/Off
5 - D4
6 - D5
7 - D6
8 - D7

If I were to send an I2C byte to this module with a value of 8 (that would be bit 3 set, with bits numbers 0 to 7), that would toggle the LCD backlight on. Sending a 0 would turn it off.

That was the first thing I was able to do. Here is an Arduino sketch that will toggle that pin on and off, making the backlight blink:

// PCF8574 connected to LCD2004/LCD1602/etc.
#include <Wire.h>

void setup() {
  // put your setup code here, to run once:
  Wire.begin ();
}

void loop() {
  // put your main code here, to run repeatedly:
  Wire.beginTransmission (39); // I2C address
  Wire.write (8); // Backlight on
  Wire.endTransmission ();

  delay (500);

  Wire.beginTransmission (39); // I2C address
  Wire.write (0); // Backlight off
  Wire.endTransmission ();

  delay (500);
}

Once I understood which bit went to which LCD pin, I could then start figuring out how to talk to the LCD.

One of the first things I did was create some #defines representing each bit:

#define BIT(b)      (1<<(b))

#define DB7_BIT     BIT(7) // High nibble for 4-bit mode. 
#define DB6_BIT     BIT(6) //
#define DB5_BIT     BIT(5) //
#define DB4_BIT     BIT(4) //
#define BL_BIT      BIT(3) // Backlight (0=Off, 1=On)
#define E_BIT       BIT(2) // Enable (0=Disable, 1=Enable)
#define RW_BIT      BIT(1) // Read/Write (0=Write, 1=Read)
#define RS_BIT      BIT(0) // Register Select (0=Instruction, 1=Data)

We’ll use this later when building our own bytes to send out.

Here is a datasheet for the LCD2004 module. Communicating with an LCD1602 is identical except for how many lines you have and where they exist in screen memory:

https://cdn-shop.adafruit.com/datasheets/TC2004A-01.pdf

I actually started with an LCD1602 datasheet and had it all working before I understood what “1602” meant a different sized display than whatI had ;-)

Sending a byte

As you can see from the above sample code, to send an I2C byte on the Arduino, you have to include the Wire library (for I2C) and initialize it in Setup:

#include <Wire.h>

void setup() {
  // put your setup code here, to run once:
  Wire.begin ();
}

Then you use a few lines of code to write the byte out to the I2C address of the PCF8574 module. The address is 39 by default, but there are solder pads on these boards that let you change it to a few other addresses.

Wire.beginTransmission (39); // I2C address
Wire.write (8); // Backlight on
Wire.endTransmission ();

Communicating with the LCD module requires a few more steps. First, you have to figure out which pins you want set on the LCD, then you write out a byte that represents them. The “E” pin must be set (1) to tell the LCD to look at the data pins.

After a tiny pause, you write out the value again but with the E pin bit unset (0).

That’s all there is to it! The rest is just understanding what pins you need to set for what command.

Instructions versus Data

The LCD module uses a Register Select pin (RS) to tell it if the 8-bits of I/O represents an Instruction, or Data.

  • Instruction – If you set the 8 I/O pins and have RS off (0) then toggle the Enable pin on and off, the LCD receives those 8 I/O pins as an Instruction.
  • Data – If you set the 8 I/O pins and have RS on (1) then toggle the Enable pin on and off, the LCD received those 8 I/O pins as a Data byte.

Reading and Writing

In addition to sending Instructions or Data to the LCD, you can also read Data back. This tutorial will not cover that, but it’s basically the same process except you set the Read/Write pin to 1 and then pulse the E pin high/low and then you can read the pins that will be set by the LCD.

Initialize the LCD to 4-bit mode

Since only 4 of the PCF8574 I/O pins are used for data, the first thing that must be done is to initialize the LCD module to 4-bit mode. This is done by using the Function Set instruction.

Function set is described as the following:

RS  RW  DB7 DB6 DB5 DB4 DB3 DB2 DB1 DB0
--- --- --- --- --- --- --- --- --- ---
0 0 0 0 1 DL N F x x

Above, RS is the Register Select pin, RW is the Read/Write pin, and DB7-DB0 are the eight I/O pins. For Function Set, pins DB7-DB5 are “001” representing the Function Select instruction. After that, the pins are used for settings of Function Select:

  • DB4 is Data Length select bit. (DL)
  • DB3 is Number of Lines select bit
  • DB2 is Font select bit

When we are using the PCF8574 module, it ONLY gives us access to DB7-DB4, so it is very smart that they chose to make the DL setting one of those four bits. We have no way to access the pins for N or F until we toggle the LCD in to 4-bit data length mode.

If we were using all 8 I/O pins, we’d set them like this to go in to 4-bit mode:

RS  RW  DB7 DB6 DB5 DB4 DB3 DB2 DB1 DB0 E
--- --- --- --- --- --- --- --- --- --- ---
0   0   0   0   1   1   0   0   0   0   1   <- Enable pin ON
(pause)
0   0   0   0   1   1   0   0   0   0   0   <- Enable pin OFF

…EXCEPT, the LCD won’t listen to us until we initialize it by sending a series of Instructions first. Here is what the LCD expects to wake it up:

RS  RW  DB7 DB6 DB5 DB4 DB3 DB2 DB1 DB0 E
--- --- --- --- --- --- --- --- --- --- ---
0   0   0   0   1   1   0   0   0   0   1   <- Enable pin ON
(pause)
0   0   0   0   1   1   0   0   0   0   0   <- Enable pin OFF

(wait at least 4.1 ms)

0   0   0   0   1   1   0   0   0   0   1   <- Enable pin ON
(pause)
0   0   0   0   1   1   0   0   0   0   0   <- Enable pin OFF

(wait at least 100 us)

0   0   0   0   1   1   0   0   0   0   1   <- Enable pin ON
(pause)
0   0   0   0   1   1   0   0   0   0   0   <- Enable pin OFF

That sequence will initialize the LCD so we can send it commands. After that, we can use Function Set to change it to 4-bit mode (DB4 as 0 for 4-bit mode):

RS  RW  DB7 DB6 DB5 DB4 DB3 DB2 DB1 DB0 E
--- --- --- --- --- --- --- --- --- --- ---
0   0   0   0   1   0   0   0   0   0   1   <- Enable pin ON
(pause)
0   0   0   0   1   0   0   0   0   0   0   <- Enable pin OFF

If we used all 8 I/O pins directly, we could also set Font and Number of lines at the same time after the three initializing writes. BUT, since we are using the PCD8547 and only have access to the top four bits (DB7-DB4), we must put the LCD in to 4-bit mode first. More details on how we use that in a moment.

If I wanted to initialize the LCD, I would just need to translate the I/O pins into the bits of a PCF8574 byte. For the first three initialization writes, it would look like this:

LCD pins:                                    PCD8574 pins:

RS  RW  DB7 DB6 DB5 DB4 DB3 DB2 DB1 DB0 E      DB7 DB6 DB5 DB4 BL E  RW RS
--- --- --- --- --- --- --- --- --- --- --- =  --- --- --- --- -- -- -- --
0   0   0   0   1   1   0   0   0   0   1      0   0   1   1   0  1  0  0
(pause)
0   0   0   0   1   1   0   0   0   0   0      0   0   1   1   0  0  0  0

And that would turn in these two I2C writes:

Wire.beginTransmission(PCF8574_ADDRESS);

Wire.write(0b00110100); // DB7 DB6 DB5 DB4 BL E RW RS
delayMicroseconds(1);
Wire.write(0b00110000); // DB7 DB6 DB5 DB4 BL E RW RS
delayMicroseconds(37);
Wire.endTransimssion();

You could manually do the intiializtion writes like that, with the required sleeps between each one.

But, instead, I wrote a function that will send a 4-bit value out to the upper four I/O pins (DB7-DB4):

// [7  6  5  4  3  2  1  0 ]
// [D7 D6 D5 D4 BL -E RW RS]
void LCDWriteInstructionNibble(uint8_t nibble)
{
    uint8_t dataByte = BL_BIT | (nibble << 4);

    Wire.beginTransmission(PCF8574_ADDRESS);

    Wire.write(E_BIT | dataByte);
    delayMicroseconds(1);

    Wire.write(dataByte);
    delayMicroseconds(37);

    Wire.endTransmission();
}

ABove, you see only need to pass in the bit pattern for DB7 DB6 DB5 DB4. This routine will set the Backlight Bit (it doesn’t have to, but I didn’t want the screen to blank out when sending these instructions), and then write the byte out with the E pin set, pause, then write it out again with E off.

Thus, my initialization can now look like this:

// Initialize all pins off and give it time to settle.
Wire.beginTransmission(PCF8574_ADDRESS);
Wire.write(0x0);
Wire.endTransmission();

delayMicroseconds(50000);

// [7  6  5  4  3  2  1  0 ]
// [D7 D6 D5 D4 BL -E RW RS]
LCDWriteInstructionNibble(0b0011);
delay(5); // min 4.1 ms

LCDWriteInstructionNibble(0b0011);
delayMicroseconds(110); // min 100 us

LCDWriteInstructionNibble(0b0011);
delayMicroseconds(110); // min 100 us

// Set interface to 4-bit mode.
LCDWriteInstructionNibble(0b0010);

That looks much more obvious, and reduces the amount of lines we need to look at since the function will do the two writes (E on, E off) for us.

Sending 8-bits in a 4-bit world

Now that the LCD is in 4-bit mode, it will expect those four I/O pins set twice — the first time for the upper 4-bits of a byte, and then the second time for the lower 4-bits. We could, of course, do this manually as well by figuring all this out and building the raw bytes outselves.

But that makes my head hurt and is too much work.

Instead, I created a second function that will send an 8-bit value 4-bits at a time:

void LCDWriteByte(uint8_t rsBit, uint8_t dataByte)
{
    uint8_t newByte;

    Wire.beginTransmission(PCF8574_ADDRESS);

    newByte = BL_BIT | rsBit | (dataByte & 0xf0);
    Wire.write(E_BIT | newByte);
    delayMicroseconds(1);
    Wire.write(newByte);
    delayMicroseconds(37);

    newByte = BL_BIT | rsBit | (dataByte << 4);
    Wire.write(E_BIT | newByte);
    delayMicroseconds(1);
    Wire.write(newByte);
    delayMicroseconds(37);

    Wire.endTransmission();
}

You’ll notice I pass in the Register Select bit, which can either be 0 (for an Instruction) or 1 (for data). That’s jumping ahead a bit, but it makes sense later.

I can then pass in a full instruction, like sending Function set to include the bits I couldn’t set during initialization when the LCD was in 8-bit mode and I didn’t have access to DB3-DB0. My LCDInit() routine set the LCD to 4-bit mode, and then uses this to send out the rest of the initialization:

// Function Set
// [0  0  1  DL N  F  0  0 ]
// DL: 1=8-Bit, 0=4-Bit
//  N: 1=2 Line, 0=1 Line
//  F: 1=5x10, 0=5x8
//             [--001DNF00]
LCDWriteByte(0, 0b00101000); // RS=0, Function Set

// Display On
// [0  0  0  0  1  D  C  B ]
// D: Display
// C: Cursor
// B: Blink
//             [--00001DCB]
LCDWriteByte(0, 0b00001100); // RS=0, Display On

// Display Clear
// [0  0  0  0  0  0  0  1 ]
LCDWriteByte(0, 0b00000001);
delayMicroseconds(3);  // 1.18ms - 2.16ms

// Entry Mode Set
// [0  0  0  0  0  1  ID S ]
// ID: 1=Increment, 0=Decrement
//  S: 1=Shift based on ID (1=Left, 0=Right)
//             [--000001IS]
LCDWriteByte(0, 0b00000110);

To make things even more clear, I then created a wrapper function for writing an Instruction that has RS at 0, and another for writing Data that has RS at 1:

void LCDWriteInstructionByte(uint8_t instruction)
{
    LCDWriteByte(0, instruction);
}


/*--------------------------------------------------------------------------*/
// Write with RS bit 1.
void LCDWriteDataByte(uint8_t data)
{
    LCDWriteByte(RS_BIT, data);
}

That lets the code look a bit cleaner, and means the user doesn’t have to know about using RS_BIT or 0 … just write Instruction or Data like this:

// Function Set
// [0 0 1 DL N F 0 0 ]
// DL: 1=8-Bit, 0=4-Bit
// N: 1=2 Line, 0=1 Line
// F: 1=5x10, 0=5x8
// [--001DNF00]
LCDWriteInstructionByte(0b00101000);

// Display On
// [0 0 0 0 1 D C B ]
// D: Display
// C: Cursor
// B: Blink
// [--00001DCB]
LCDWriteInstructionByte(0b00001100);

// Display Clear
// [--00000001]
LCDWriteInstructionByte(0b00000001);
delayMicroseconds(3); // 1.18ms - 2.16ms

// Entry Mode Set
// [0 0 0 0 0 1 ID S ]
// ID: 1=Increment, 0=Decrement
// S: 1=Shift based on ID (1=Left, 0=Right)
// [--000001IS]
LCDWriteInstructionByte(0b00000110);

The Display Clear instruction is 00000001. There are no other bits that need to be set, so I can clear the screen by doing “LCDWriteInstructionByte (0b00000001)” or simply “LCDWriteInstructionByte(1)”;

Ultimately, I’d probably create #defines for the different instructions, and the settable bits inside of them, allowing me to build a byte like this:

uint8_t functionSetByte = FUNCTION_SET | DL_BIT | N_BIT | F_BIT;

FUNCTION_SET would represent the bit pattern 0b00100000, and the DL_BIT would be BIT(4), N_BIT would be BIT(3) and F_BIT would be BIT(2). Fleshing out all of those defines and then making wrapper functions would be trivial.

But in my case, I only needed a few, so if you wanted to make something that did that, you could:

void LCDFunctionSet (uint8_t dlBit, uint8_t nBit, uint8_t fBit)
{
   uint8_t instruction = FUNCTION_SET;
   if (dlBit)
   {
      instruction |= DL_BIT;
   }

   if (nBit)
   {
      instruction |= N_BIT;
   }

   if (fBit)
   {
      instruction |= F_BIT;
   }

   LCDWriteInstructionByte (instruction);
}

This type of thing can allow your code to spiral out of control as you create functions to set bits in things like “Display On/Off Control” and then write wrapper functions like “LCDDisplayON()”, “LCDBlinkOn()” and so on.

But we won’t be going there. I’m just showing you the basic framework.

Now what?

With the basic steps to Initialize to 4-Bit Mode, then send out commands, the rest is pretty simple. If you want to write out bytes to be displayed on the screen, you just write out a byte with the Register Select bit set (for Data, instead of Instruction). The byte appears at whatever location the LCD has for the cursor position. Simple!

At the very least, you need a Clear Screen function:

void LCDClear(void)
{
    // Display Clear
    // [0  0  0  0  0  0  0  1 ]
    LCDWriteInstructionByte(0b00000001);
    delay(3); // 1.18ms - 2.16ms
}

…and a function to write out data bytes, such as raw data (point to the data, give it the size):

void LCDWriteData(uint8_t *dataPtr, uint8_t dataSize)
{
    for (int idx = 0; idx < dataSize; idx++)
    {
        LCDWriteDataByte(dataPtr[idx]);
    }
}

…and maybe a nicer function that deals with C strings:

void LCDWriteDataString(char *message)
{
    LCDWriteData((uint8_t *)message, strlen(message));
}

Wrap, wrap, wrap we go…

The last thing I implemented was a thing that sets the X/Y position of where text will go. This is tricky because the display doesn’t match the memory inside the screen. Internally my LCD2004 just has a buffer of screen memory that maps to the LCD somehow.

The LCD data is not organized as multiple lines of 20 characters (or 16). Instead, it is just a buffer of screen memory that is mapped to the display. In the case of the LCD2004, the screen is basically 128 bytes of memory, with the FIRST line being bytes 0-19, the SECOND line being bytes 64-83, the THIRD line being bytes 20-39, and the FOURTH line being bytes 84-103.

Consider this example:

+--------------------+
|ABCDEFGHIJKLMNOPQRST| (bytes 0-19)
|abcdefghijklmnopqrst| (bytes 64-83)
|UVWXYZ              | (bytes 20-39)
|uvwxyz              | (bytes 84-103)
+--------------------+

If you were to start at memory offset 0 (top left of the display) and write 80 bytes of data (thinking you’d get 20, 20, 20 and 20 bytes on the display), that wouldn’t happen ;-) You’d see some of your data did not show up since it was writing out in the memory that is not mapped in to the display. (You can also use that memory for data storage, but I did not implement any READ routines in this code — yet.)

If you actually did start at offset 0 (the first byte of screen memory) and wrote a series of characters from 32 (space) to 127 (whatever that is), it would look like this:

Above, you can see the first line continues on line #3, and then after the end of line 3 (…EFG” we don’t see any characters until we get to the apostrophe which displays on line 2. Behind the scenes, memory looks like this:

0  [ !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_] 63
64 [`abcdefghijklmnopqrstuvwxyz{|}--                                ] 127

That is the “2 line” mode of the LCD. But, the physical display treats that data like this:

0  [ !"#$%&'()*+,-./0123] 19
64 [`abcdefghijklmnopqrs] 83
20 [456789:;<=>?@ABCDEFG] 39 
84 [tuvwxyz{|}--        ] 103

not visible:

40 [HIJKLMNOPQRSTUVWXYZ[\]^_] 63
104[                        ] 127

Clear as mud? Good.

All you need to know is that the visible screen doesn’t match LCD memory, so when creating a “set cursor position” that translates X and Y to an offset of memory, it has to have a lookup table, like this one:

void LCDSetOffset(uint8_t offset)
{
    // DDRAM AD SET
    LCDWriteInstructionByte(0b10000000 | offset);
    delayMicroseconds(40);
}

void LCDSetXY(uint8_t x, uint8_t y)
{
    uint8_t offset = 0;

    if (y == 1)
    {
        offset = 64; // 0x40
    }
    else if (y == 2)
    {
        offset = 20; // 0x14
    }
    else if (y == 3)
    {
        offset = 84; // 0x54
    }
    else
    {
        // Offset will be 0.
    }

    offset = offset + x;

    LCDSetOffset(offset);
}

You will see I created a function that sends the “Set Offset” instruction (memory location 0 to 127, I think) and then a “Set X/Y” function that translates columns and rows to an offset.

With all that said, here are the routines I cam up with. Check my GitHub for the latest versions:

https://github.com/allenhuffman/PCF8547_LCD2004

The LCDTest.ino program also demonstrates how you can easily send an Instruction to load character data, and then send that data using the LCDWriteData functions.

I plan to revisit this with more details on how all that works, but wanted to share what I had so far.

Until next time…


LCD2004_PCF8574.h

#ifndef LCD2004_PCF8547_H
#define LCD2004_PCF8547_H

#include <stdint.h>

#define PCF8574_ADDRESS   0x27 // 39 (7-bit address)
//#define PCF8574_ADDRESS   (0x27*2) // 39 (8-bit address)

#if !defined(BIT)
#define BIT(b)      (1<<(b))
#endif

// PCF8574 8-bit pins map to the following LCD2004 pins:
//
// [7  6  5  4  3  2  1  0 ]
// |D7 D6 D5 D4 BL -E RW RS
//
// NOTE: This is hard-coded to assume the upper four bits are
// the 4 data lines to the LCD, in that order.

typedef enum
{
    DB7_BIT = BIT(7), // High nibble for 4-bit mode. 
    DB6_BIT = BIT(6), //
    DB5_BIT = BIT(5), //
    DB4_BIT = BIT(4), //
    BL_BIT  = BIT(3), // Backlight (0=Off, 1=On)
    E_BIT   = BIT(2), // Enable (0=Disable, 1=Enable)
    RW_BIT  = BIT(1), // Read/Write (0=Write, 1=Read)
    RS_BIT  = BIT(0)  // Register Select (0=Instruction, 1=Data)
} LCDBitEnum;

// #define DB7_BIT     BIT(7) // High nibble for 4-bit mode. 
// #define DB6_BIT     BIT(6) //
// #define DB5_BIT     BIT(5) //
// #define DB4_BIT     BIT(4) //
// #define BL_BIT      BIT(3) // Backlight (0=Off, 1=On)
// #define E_BIT       BIT(2) // Enable (0=Disable, 1=Enable)
// #define RW_BIT      BIT(1) // Read/Write (0=Write, 1=Read)
// #define RS_BIT      BIT(0) // Register Select (0=Instruction, 1=Data)

// Function Prototypes
bool IsLCDEnabled (void);

bool LCDInit (void);
void LCDTerm (void);

void LCDWriteInstructionNibble (uint8_t nibble);

void LCDWriteByte (uint8_t rsBit, uint8_t value);
void LCDWriteInstructionByte (uint8_t instruction);
void LCDWriteDataByte (uint8_t data);

void LCDWriteData (uint8_t *dataPtr, uint8_t dataSize);
void LCDWriteDataString (uint8_t x, uint8_t y, char *message);

void LCDSetOffset (uint8_t offset);
void LCDSetXY (uint8_t x, uint8_t y);

void LCDClear (void);

void LCDWaitForBusyFlag (void);

#endif   /* LCD2004_PCF8547_H */

// End of LCD2004_PCF8547.h

LCD2004_PCF8547.ino

#ifndef LCD2004_PCF8547_C
#define LCD2004_PCF8547_C

/*--------------------------------------------------------------------------*/
// Include Files
/*--------------------------------------------------------------------------*/
#include <stdint.h>

#include <Wire.h>

#include "LCD2004_PCF8547.h"


/*--------------------------------------------------------------------------*/
// Variables
/*--------------------------------------------------------------------------*/
static bool S_IsLCDEnabled = false;


/*--------------------------------------------------------------------------*/
// Functions
/*--------------------------------------------------------------------------*/
bool IsLCDEnabled(void)
{
    return S_IsLCDEnabled;
}


/*--------------------------------------------------------------------------*/
// Initialize the LCD2004.
/*--------------------------------------------------------------------------*/
//    4-Bit data mode
//    2 Line display
//    Display On
//    Display Clear
//    Left-to-Right display mode.
bool LCDInit(void)
{
    int ack = 0;

    Wire.begin();

    // Set all PCF8547 I/O pins LOW.
    Wire.beginTransmission(PCF8574_ADDRESS);
    Wire.write(0x0);
    ack = Wire.endTransmission();

    if (ack != 0)
    {
        return false;
    }
    delayMicroseconds(50000);

    // [7  6  5  4  3  2  1  0 ]
    // [D7 D6 D5 D4 BL -E RW RS]
    LCDWriteInstructionNibble(0b0011);
    delay(5); // min 4.1 ms

    LCDWriteInstructionNibble(0b0011);
    delayMicroseconds(110); // min 100 us

    LCDWriteInstructionNibble(0b0011);
    delayMicroseconds(110); // min 100 us

    // Set interface to 4-bit mode.
    LCDWriteInstructionNibble(0b0010);

    // Function Set
    // [0  0  1  DL N  F  0  0 ]
    // DL: 1=8-Bit, 0=4-Bit
    //  N: 1=2 Line, 0=1 Line
    //  F: 1=5x10, 0=5x8
    //                      [--001DNF00]
    LCDWriteInstructionByte(0b00101000);

    // Display On
    // [0  0  0  0  1  D  C  B ]
    // D: Display
    // C: Cursor
    // B: Blink
    //                      [--00001DCB]
    LCDWriteInstructionByte(0b00001100);

    // Display Clear
    // [0  0  0  0  0  0  0  1 ]
    LCDWriteInstructionByte(0b00000001);
    delayMicroseconds(3);  // 1.18ms - 2.16ms

    // Entry Mode Set
    // [0  0  0  0  0  1  ID S ]
    // ID: 1=Increment, 0=Decrement
    //  S: 1=Shift based on ID (1=Left, 0=Right)
    //                      [--000001IS]
    LCDWriteInstructionByte(0b00000110);

    S_IsLCDEnabled = true;

    return true;
}


/*--------------------------------------------------------------------------*/
// Disable LCD screen.
/*--------------------------------------------------------------------------*/
void LCDTerm(void)
{
    S_IsLCDEnabled = false;
}


/*--------------------------------------------------------------------------*/
// Write out a 4-bit value.
/*--------------------------------------------------------------------------*/
// [7  6  5  4  3  2  1  0 ]
// [D7 D6 D5 D4 BL -E RW RS]
void LCDWriteInstructionNibble(uint8_t nibble)
{
    uint8_t dataByte = BL_BIT | (nibble << 4);

    Wire.beginTransmission(PCF8574_ADDRESS);

    Wire.write(E_BIT | dataByte);
    delayMicroseconds(1);

    Wire.write(dataByte);
    delayMicroseconds(37);

    Wire.endTransmission();
}


/*--------------------------------------------------------------------------*/
// Write a byte out, 4-bits at a time. The rsBit will determine
// if it is an Instruction (rsBit=0) or Data byte (rsBit=1).
/*--------------------------------------------------------------------------*/
void LCDWriteByte(uint8_t rsBit, uint8_t dataByte)
{
    uint8_t newByte;

    Wire.beginTransmission(PCF8574_ADDRESS);

    newByte = BL_BIT | rsBit | (dataByte & 0xf0);
    Wire.write(E_BIT | newByte);
    delayMicroseconds(1);
    Wire.write(newByte);
    delayMicroseconds(37);

    newByte = BL_BIT | rsBit | (dataByte << 4);
    Wire.write(E_BIT | newByte);
    delayMicroseconds(1);
    Wire.write(newByte);
    delayMicroseconds(37);

    Wire.endTransmission();
}


/*--------------------------------------------------------------------------*/
// Write with RS bit 0.
/*--------------------------------------------------------------------------*/
void LCDWriteInstructionByte(uint8_t instruction)
{
    LCDWriteByte(0, instruction);
}


/*--------------------------------------------------------------------------*/
// Write with RS bit 1.
/*--------------------------------------------------------------------------*/
void LCDWriteDataByte(uint8_t data)
{
    LCDWriteByte(RS_BIT, data);
}


/*--------------------------------------------------------------------------*/
// Write out one or more data bytes.
/*--------------------------------------------------------------------------*/
void LCDWriteData(uint8_t *dataPtr, uint8_t dataSize)
{
    for (int idx = 0; idx < dataSize; idx++)
    {
        LCDWriteDataByte(dataPtr[idx]);
    }
}


/*--------------------------------------------------------------------------*/
// Write out a NIL terminated C string.
/*--------------------------------------------------------------------------*/
void LCDWriteDataString(uint8_t x, uint8_t y, char *message)
{
    if (S_IsLCDEnabled == true)
    {
        LCDSetXY(x, y);
        LCDWriteData((uint8_t *)message, strlen(message));
    }
}


/*--------------------------------------------------------------------------*/
// Set LCD offset in screen memory.
/*--------------------------------------------------------------------------*/
void LCDSetOffset(uint8_t offset)
{
    // DDRAM AD SET
    LCDWriteInstructionByte(0b10000000 | offset);
    delayMicroseconds(40);
}


/*--------------------------------------------------------------------------*/
// Set LCD cursor position.
/*--------------------------------------------------------------------------*/
// LCD2004 is internally treated like a two line display of 64 characters
// each. The first internal line is bytes 0-63 and the second internal
// line is bytes 64-127.
//
// For the physical LCD2004 20x4 four-line display, the first 20 bytes of
// internal line 1 (0-19) is the display line 1. The second 20 bytes of
// internal line 1 (20-39) is the display line 3. The first 20 bytes
// of internal line 2 (64-83) is display line 2. The second 20 bytes
// of internal line 2 (84-103) is display line 4.
//
// Super easy and not confusing at all.
//
//                          +--------------------+
// Internal Line 1 (0-19)   |aaaaaaaaaaaaaaaaaaaa| Display line 1
// Internal Line 2 (64-83)  |bbbbbbbbbbbbbbbbbbbb| Display line 2
// Internal Line 1 (20-39)  |aaaaaaaaaaaaaaaaaaaa| Display line 3
// Internal Line 2 (84-103) |bbbbbbbbbbbbbbbbbbbb| Display line 4
//                          +--------------------+
//
// Because of this, we will use a simple translation to get between
// column (x) and row (y) to the actual offset of these two internal
// 64-byte lines.
//
void LCDSetXY(uint8_t x, uint8_t y)
{
    uint8_t offset = 0;

    if (y == 1)
    {
        offset = 64; // 0x40
    }
    else if (y == 2)
    {
        offset = 20; // 0x14
    }
    else if (y == 3)
    {
        offset = 84; // 0x54
    }
    else
    {
        // Offset will be 0.
    }

    offset = offset + x;

    LCDSetOffset(offset);
}


/*--------------------------------------------------------------------------*/
// Clear the LCD (and home the cursor position).
/*--------------------------------------------------------------------------*/
void LCDClear(void)
{
    if (S_IsLCDEnabled == true)
    {
        // Display Clear
        // [0  0  0  0  0  0  0  1 ]
        LCDWriteInstructionByte(0b00000001);
        delay(3); // 1.18ms - 2.16ms
    }
}


// NOT WORKING YET! I was experimenting with doing a READ operation, and
// seeing if I could poll the BF (busy flag) bit.
#if 0
void LCDWaitForBusyFlag (void)
{
  while (1)
  {
    // Toggle on READ mode.
    Wire.beginTransmission (PCF8574_ADDRESS);
    Wire.write (E_BIT | RW_BIT | DB7_BIT);
    Wire.endTransmission ();

    // Read I/O pins.
    Wire.requestFrom (PCF8574_ADDRESS, 1);
    if ((Wire.read() & DB7_BIT) == 0)
    {
      break;
    }

    // Else toggle off E, to start another ready.
    Wire.beginTransmission (PCF8574_ADDRESS);
    Wire.write (RW_BIT | DB7_BIT);
    Wire.endTransmission ();
  }
}
#endif // 0

#endif /* LCD2004_PCF8547_C */

// End of LCD2004_PCF8547.c

LCDTest.ino

// LCDTest.ino

// TI PCF8574 (or compatible)

#include <stdint.h>

#include <Wire.h>

#include "LCD2004_PCF8547.h"

#define LCD_WIDTH   20
#define LCD_HEIGHT  4

uint8_t data[] =
{
  // Pac Close
  0b01110,
  0b11111,
  0b11111,
  0b11111,
  0b11111,
  0b11111,
  0b11111,
  0b01110,

  // Pack Right
  0b01110,
  0b11111,
  0b11110,
  0b11100,
  0b11100,
  0b11110,
  0b11111,
  0b01110,

  // Ghost 1
  0b01110,
  0b11111,
  0b11111,
  0b10101,
  0b11111,
  0b11111,
  0b11111,
  0b10101,

  // Ghost 2
  0b01110,
  0b11111,
  0b11111,
  0b10101,
  0b11111,
  0b11111,
  0b11111,
  0b01010,

  // Dot
  0b00000,
  0b00000,
  0b00000,
  0b01100,
  0b01100,
  0b00000,
  0b00000,
  0b00000,

  // Maze Left
  0b00000,
  0b01111,
  0b10000,
  0b10000,
  0b10000,
  0b10000,
  0b01111,
  0b00000,

  // Maze Center
  0b00000,
  0b11111,
  0b00000,
  0b00000,
  0b00000,
  0b00000,
  0b11111,
  0b00000,

  // Maze Right
  0b00000,
  0b11110,
  0b00001,
  0b00001,
  0b00001,
  0b00001,
  0b11110,
  0b00000,
};

void setup()
{
  Serial.begin(9600);
  Wire.begin(); // I2C Master

  Serial.print ("sizeof(data) = ");
  Serial.println (sizeof(data));

  LCDInit ();

  delay(500);

  // Set CGRAM Address
  uint8_t offset = 0;
  LCDWriteInstructionByte (0b01000000 | offset);
  delayMicroseconds(40);
  LCDWriteData (&data[0], sizeof(data));

  LCDClear ();
}


void loop()
{
  delay (1000);

  // LCD memory locations 0-127.
  // for (int idx=0; idx<128; idx++)
  // {
  //   LCDSetOffset (idx);
  //   LCDWriteDataByte (idx+32);
  // }

  LCDWriteDataString (0, 0, "8 Programmable Chars");

  LCDWriteDataString (1, 2, "0:");
  LCDWriteDataByte (0);
  LCDWriteDataString (6, 2, "1:\1  2:\2  3:\3");

  LCDWriteDataString (1, 3, "4:\4  5:\5  6:\6  7:\7");

  delay (4000);

  LCDWriteDataString (0, 1, "\5\6\6\6\6\6\6\6\6\6\6\6\6\6\6\6\6\6\6\7");
  LCDWriteDataString (0, 2, "\4\4\4\4\4\4\4\4\4\4\4\4\4\4\4\4\4\4\4\4");
  LCDWriteDataString (0, 3, "\5\6\6\6\6\6\6\6\6\6\6\6\6\6\6\6\6\6\6\7");

  delay (2000);

  for (int x=0; x<24; x++)
  {
    if (x < 20)
    {
      LCDSetXY (x , 2);
      LCDWriteDataByte (x & 1);
    }

    if (x > 4)
    {
      LCDSetXY (x-4, 2);
      LCDWriteDataByte (2 + (x & 1));
    }

    delay(250);
    
    if (x < 20)
    {
      LCDSetXY (x, 2);
      LCDWriteDataByte (32);
    }

    if (x > 4)
    {
      LCDSetXY (x-4, 2);
      LCDWriteDataByte (32);
    }
  }

  delay(3000);

  LCDClear ();

  int x = 0;
  int y = 0;
  int xm = 1;
  int ym = 1;

  for (int idx=0; idx<40; idx++)
  {
    LCDWriteDataString (x, y, "Sub-Etha!");
    delay(250);
    LCDWriteDataString (x, y, "         ");

    x = x + xm;
    if ((x < 1) || (x >= LCD_WIDTH-8-1))
    {
      xm = -xm;
    }

    y = y + ym;
    if ((y < 1) || (y >= LCD_HEIGHT-1))
    {
      ym = -ym;
    }
  }

  delay(3000);
}

// End of LCDTest.ino

Data paranoia, Drobos and Synology NAS

I am in the slow process of upgrading five WD 6TB hard drives in my Synology DS1522+ NAS to Seagate 8TB drives. While folks I asked overwhelming say the larger drives (16TB, 20TB, etc.) “have not had any issues,” I am old school and do not want to make that large of a storage jump just yet. For those as data paranoid as I am, some tips:

  1. Do a low-level format (or “secure erase”) on each new drive first. This will write to every sector, and can catch issues. I have only caught one (or maybe two) drive with issues of the years, but the time spent is worth it to me. I’d rather catch a problem before I install and send them back for a replacement, rather than having an issue show up much later (possibly when the drive is out of warranty).
  2. Until someone can say that a “20TB” drive is as reliable as a smaller drive, upgrage to the next size up that gives you enough storage. The less dense the data, the safer it “should” be. Also, if a 6TB drive fails, the rebuild time to replace it will be significantly faster than replacing a 20TB drive. And, during the rebuilt time, your data is at risk. I run dual drive redundancy so during the 12 hours my NAS rebuilds, if I have a second drive fail, I am still okay… but if that happens I have no protection from a third failure. Doing 20 hours rebuilds creates a much larger window for data loss if something goes terribly wrong.
  3. And, of course, make sure anything important is on a backup drive (I use a standalone 10TB just to clone my most important “can’t live without” data), and have an offsite backup of that (I then have that entire drive backed up to a cloud backup service).

If my home burns down, I should at least be able to get back the 10TB of “can’t live without” data from my offsite backup.

Hardware Redundancy is Better

Sadly, Synology units are much more expensive than my Drobos were. My Drobo 5C was $300 retail. I had two that I paid maybe $250 each for. That let me have two 5-bay units for hardware redundancy. This means if a Drobo suddenly died, I still had a second unit with duplicate data (I would sync the two drives). Spending $500 for two 5-drive units was an easier investment than the $1400 it would take for me to buy two DS1522+.

Eventually I do plan to have a duplicate Synology unit. It doesn’t matter what features the device has if one morning it has died. I would have zero access to my data until the unit is replaced or repaired. Having backup hardware is what I prefer.

But I am data paranoid.

How about you? Are you unlucky, like I am, and have had drives “die suddenly” over the years? After that happens enough, the paranoia sets in. I can’t think of a time in the past decades where I didn’t have three backups of everything important ;-)

Wokwi online Arduino/ESP32 simulator

Oh, Wokwi! Where have you been all my life? Or at least where were you a few years ago when I was working on Arduino projects?

I initially started using this Sub-Etha Software blog for Arduino projects. I did crazy things like porting a Color BASIC program to Arduino C, and fun things like figuring out how to write Pac-Man using the Arduio TVOut library. Eventually, I merged my interests in the old Radio Shack Color Computer (CoCo) with Arduino (and ESP8266/ESP32) and experimented with a serial port sound player and WiFi modem.

Prototype “Sir Sound” sound module for the CoCo (or anything with a serial port, actually).

So … many … wires.

At the time, I was hoping to find some kind of Arduino emulator so I could write and test code without hooking up hardware. I found nothing.

But that seems to have changed. I just learned about Wokwi which allows one to “simulate IoT projects in your browser.” In a nutshell, it’s a website that has a code editor (which appears to be Microsoft Visual Studio Code), compiler, and virtual target hardware like Arduino and ESP32 devices. It even supports some add-on hardware, like buttons, LCD displays, LEDs and more.

Here’s a project someone made that simulates an Arduino hooked to a numeric keypad and LCD display:

And you can build and run it right there!

There is a library of devices that are supported, and you can add them to your project and wire them up to the computer’s I/O pins. For example, as I write this blog post, I opened up a starter project that is an Arduino and a two-line LCD display. I then added a pushbutton to it.

I could then move the button to where I wanted it, then click on the connectors and draw wire lines between it and I/O pins on the Arduino. By hooking one side to an I/O pin, and the other to ground, I could then modify the program to read that button and, for this example, increment a counter while the button is being held done.

It’s just that easy! I had no idea!

The files can be downloaded and used on real hardware, or you can make an account and log back in to continue working on them. (It has an unusual way to log in — it sends you an e-mail and you click a link to log in, rather than having a username and password. This seems to mean I cannot log in from any system that I don’t have my e-mail account configured on, but I do see options for using a Google or Github login.)

I hope you find this as useful as I already have.

Happy coding!

Arduino, ZigBee and motion sensor help needed.

For a future project, I need to make use of remote triggers. These could be motion sensors, beam sensors, pressure mats, etc.

The ZigBee standard seems to be the way to go since I can find cheap consumer motion sensors that run on batteries. There also seems to be ZigBee repeaters, which allow giving the distance I need simply by plugging them in from place to place to create a mesh network.

XBee might be another option, if cheap motion sensors and repeaters are also available.

The goal is to have a central location be able to read the motion sensor status for many sensors, that could be spread out beyond walls hundreds of feet away.

Any pointers to where I might get started would be appreciated. Ideally I’d drive this all by a low-cost Arduino since the device will be used in an area where power might not be stable (and I wouldn’t want to corrupt the Linux file system on a Raspberry Pi).

Thanks…

When there’s not enough room for sprintf…

Updates:

  • 2022-08-30 – Corrected a major bug in the Get8BitHexStringPtr() routine.

“Here we go again…”

Last week I ran out of ROM space in a work project. For each code addition, I have to do some size optimization elsewhere in the program. Some things I tried actually made the program larger. For example, we have some status bits that get set in two different structures. The code will do it like this:

shortStatus.faults |= FAULT_BIT;
longStatus.faults |= FAULT_BIT;

We have code like that in dozens of places. One of the things I had done earlier was to change that in to a function. This was primarily so I could have common code set fault bits (since each of the four different boards I work with had a different name for its status structures). It was also to reduce the amount of lines in the code and make what they were doing more clear (“clean code”).

void setFault (uint8_t faultBit)
{
    shortStatus.faults |= faultBit;
    longStatus.faults |= faultBit;
}

During a round of optimizing last week, I noticed that the overhead of calling that function was larger than just doing it manually. I could switch back and save a few bytes every time it was used, but since I still wanted to maintain “clean code”, I decided to make a macro instead of the function. Now I can still do:

setFault (FAULT_BIT);

…but under the hood it’s really doing a macro instead:

#define setFault(faultBit) { shortStatus.faults |= faultbit; longStatus.faults |= faultBit; }

Now I get what I wanted (a “function”) but retain the code size savings of in-lining each instance.

I also thought that doing something like this might be smaller:

shortStatus.faults |= FAULT_BIT;
longStatus.faults = shortStatus.faults;

…but from looking at the PIC24 assembly code, that’s much larger. I did end up using it in large blocks of code that conditionally decided which fault bit to set, and then I sync the long status at the end. As long as the overhead of “this = that” is less than the overhead of multiple inline instructions it was worth doing.

And keep in mind, this is because I am 100% out of ROM. Saving 4 bytes here, and 20 bytes there means the difference between being able to build or not.

Formatting Output

One of the reasons for the “code bloat” was adding support for an LCD display. The panel, an LCD2004, hooks up to I2C vie a PCF8574 I2C I/O chip. I wrote just the routines needed for the minimal functionality required: Initialize, Clear Screen, Position Cursor, and Write String.

The full libraries (there are many) for Arduino are so large by comparison, so often it makes more sense to spend the time to “roll your own” than port what someone else has already done. (This also means I do not have to worry about any licensing restrictions for using open source code.)

I created a simple function like:

LCDWriteDataString (0, 0, "This is my message.");

The two numbers are the X and Y (or Column and Row) of where to display the text on the 20×4 LCD screen.

But, I was quickly reminded that the PIC architecture doesn’t support passing constant string data due to “reasons”. (Harvard architecture, for those who know.)

To make it work, you had to do something like:

const char *msg = "This is my message";
LCDWriteDataString (0, 0, msg);

…or…

chr buffer[19];
memcpy (buffer, "This is my message");
LCDWriteDataString (0, 0, msg);

…or, using the CCS compiler tools, add this to make the compiler take care of it for you:

#device PASS_STRINGS=IN_RAM

Initially I did that so I could get on with the task at had, but as I ran out of ROM space, I revisited this to see which approach was smaller.

From looking at the assembly generated by the CCS compiler, I could tell that “PASS_STRINGS=IN_RAM” generated quite a bit of extra code. Passing in a constant string pointer was much smaller.

So that’s what I did. And development continued…

Then I ran out of ROM yet again. Since I had some strings that needed formatted output, I was using sprintf(). I knew that sprintf() was large, so I thought I could create my own that only did what I needed:

char buffer[21];
sprintf (buffer, "CF:%02x C:%02x T:%02x V:%02x", faults, current, temp, volts);
LCDWriteDataString (0, 0, buffer);

char buffer[21];
sprintf (buffer, "Fwd: %u", watts);
LCDWriteDataString (0, 1, buffer);

In my particular example, all I was doing is printing out an 8-bit value as HEX, and printing out a 16-bit value as a decimal number. I did not need any of the other baggage sprintf() was bringing when I started using it.

I came out with these quick and dirty routines:

char GetHexDigit(uint8_t nibble)
{
  char hexChar;

  nibble = (nibble & 0x0f);

  if (nibble <= 9)
  {
    hexChar = '0';
  }
  else
  {
    hexChar = 'A'-10;
  }

  return (hexChar + nibble);
}

char *Get8BitHexStringPtr (uint8_t value)
{
    static char hexString[3];

    hexString[0] = GetHexDigit(value >> 4);
    hexString[1] = GetHexDigit(value & 0x0f);
    hexString[2] = '\0'; // NIL terminate

    return hexString;
}

The above routine maintains a static character buffer of 3 bytes. Two for the HEX digits, and the third for a NIL terminator (0). I chose to do it this way rather than having the user pass in a buffer pointer since the more parameters you pass, the larger the function call gets. The downside is those 3 bytes of variable storage are reserved forever, so if I was also out of RAM, I might rethink this approach.

I can now use it like this:

const char *msgC = " C:"; // used by strcat()
const char *msgT = " T:"; // used by strcat()
const char *msgV = " V:"; // used by strcat()

char buffer[20];

strcpy (buffer, "CF:"); // allows constants
strcat (buffer, Get8BitHexStringPtr(faults));
strcat (buffer, msgC);
strcat (buffer, Get8BitHexStringPtr(current));
strcat (buffer, msgT);
strcat (buffer, Get8BitHexStringPtr(temp));
strcat (buffer, msgV);
strcat (buffer, Get8BitHexStringPtr(volts));

LCDWriteDataString (0, 1, buffer);

If you are wondering why I do a strcpy() with a constant string, then use const pointers for strcat(), that is due to a limitation of the compiler I am using. Their implementation of strcpy() specifically supports string constants. Their implementation of strcat() does NOT, requiring me to jump through more hoops to make this work.

Even with all that extra code, it still ends up being smaller than linking in sprintf().

And, for printing out a 16-bit value in decimal, I am sure there is a clever way to do that, but this is what I did:

char *Get16BitDecStringPtr (uint16_t value)
{
    static char decString[6];
    uint16_t temp = 10000;
    int pos = 0;

    memset (decString, '0', sizeof(decString));

    while (value > 0)
    {
        while (value >= temp)
        {
            decString[pos]++;
            value = value - temp;
        }

        pos++;
        temp = temp / 10;
    }

    decString[5] = '\0'; // NIL terminate

    return decString;
}

Since I know the value is limited to what 16-bits can old, I know the max value possible is 65535.

I initialize my five-digit string with “00000”. I start with a temporary value of 10000. If the users value is larger than that, I decrement it by that amount and increase the first digit in the string (so “0” goes to “1”). I repeat until the user value has been decremented to be less than 10000.

Then I divide that temporary value by 10, so 10000 becomes 1000. I move my position to the next character in the output string and the process repeats.

Eventually I’ve subtracted all the 10000s, 1000s, 100s, 10s and 1s that I can, leaving me with a string of five digits (“00000” to “65535”).

I am sure there is a better way, and I am open to it if it generates SMALLER code. :)

And that’s my tale of today… I needed some extra ROM space, so I got rid of sprintf() and rolled my own routines for the two specific types of output I needed.

But this is barely scratching the surface of the things I’ve been doing this week to save a few bytes here or there. I’d like to revisit this subject in the future.

Until next time…

Drobo “saves the day”?

I hate it when this happens… It looks like a 4TB drive in my 5-bay Drobo has gone out. Drobo cannot detect it. I have dual-drive redundancy enabled, so two drives can fail and I’d still have my data… Fortunately.

Drobo 5C showing a 4TB drive failure.

Hopefully, I won’t have two drives fail between now and the time my replacement drive arrives. :)

On the plus, drive prices have dropped since I bought these drives in 2019. I’ll begin the process of upgrading drives to 6TB models over coming months, money permitting.

16-bits don’t always add up.

Consider this simple program:

#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>

int main(int argc, char **argv)
{
    uint16_t    val1;
    uint16_t    val2;
    uint32_t    result;

    val1 = 40000;
    val2 = 50000;

    result = val1 + val2;

    printf ("%u + %u = %u\n", val1, val2, result);

    return EXIT_SUCCESS;
}

What will it print?

On my Windows PC, I see the following:

40000 + 50000 = 90000

…but if I convert the printf() and run the same code on an Arduino:

void setup() {
  // put your setup code here, to run once:
  Serial.begin(9600);

    uint16_t    val1;
    uint16_t    val2;
    uint32_t    result;

    val1 = 40000;
    val2 = 50000;

    result = val1 + val2;

    //printf ("%u + %u = %u\n", val1, val2, result);
    Serial.print(val1);
    Serial.print(" + ");
    Serial.print(val2);
    Serial.print(" = ");
    Serial.println(result);
}

void loop() {
  // put your main code here, to run repeatedly:

}

This gives me:

40000 + 50000 = 24464

…and this was the source of a bug I introduced and fixed at my day job recently.

Tha’s wrong, int’it?

I tend to write alot of code using the GCC compiler since I can work out and test the logic much quicker than repeatedly building and uploading to our target hardware. Because of that, I had “fully working” code that was incorrect for our 16-bit PIC24 processor.

In this case, the addition of “val1 + val2” is being done using native integer types. On the PC, those are 32-bit values. On the PIC24 (and Arduino, shown above), they are 16-bit values.

A 16-bit value can represent 65536 values in the range of 0-65535. If you were to have a value of 65535 and add 1 to it, on a 16-bit variable it would roll over and the result would be 0. In my example, 40000 + 50000 was rolling over 65535 and producing 24464 (which is 90000 – 65536).

You can see this happen using the Windows calculator. By default, it uses DWORD (double word – 32-bit) values. You can do the addition just fine:

You see that 40,000 + 50,000 results in 90,000, which is 0x15F90 in hex. That 0x1xxxx at the start is the rollover. If you switch the calculator in to WORD mode you see it gets truncated and the 0x1xxxx at the start goes away, leaving the 16-bit result:

Can we fix it?

The solution is very simple. In C, any time there is addition which might result in a value larger than the native int type (if you know it), you simply cast the two values being added to a larger data type, such as a 32-bit uint32_t:

void setup() {
  // put your setup code here, to run once:
  Serial.begin(9600);

    uint16_t    val1;
    uint16_t    val2;
    uint32_t    result;

    val1 = 40000;
    val2 = 50000;

    // Without casting (native int types):
    result = val1 + val2;

    //printf ("%u + %u = %u\n", val1, val2, result);
    Serial.print(val1);
    Serial.print(" + ");
    Serial.print(val2);
    Serial.print(" = ");
    Serial.println(result);

    // Wish casting:
    result = (uint32_t)val1 + (uint32_t)val2;

    Serial.print(val1);
    Serial.print(" + ");
    Serial.print(val2);
    Serial.print(" = ");
    Serial.println(result);
}

void loop() {
  // put your main code here, to run repeatedly:

}

Above, I added a second block of code that does the same add, but casting each of the val1 and val2 variables to 32-bit values. This ensures they will not roll over since even the max values of 65535 + 65535 will fit in a 32-bit variable.

The result:

40000 + 50000 = 24464
40000 + 50000 = 90000

Since I know adding any two 16-bit values can be larger than what a 16-bit value can hold (i.e., “1 + 1” is fine, as is “65000 + 535”, but larger values present a rollover problem), it is good practice to just always cast upwards. That way, the code works as intended, whether the native int of the compiler is 16-bits or 32-bits.

As my introduction of this bug “yet again” shows, it is a hard habit to get in to.

Until next time…

PiZ SuperCap safe shutdown for Raspberry Pi Zero

Last year, I learned of a new Kickstarter that solved a significant problem with using a Raspberry Pi for embedded “turn key” projects. The Raspberry Pi is a disk-based Linux system, using a microSD card in place of a hard drive. If you kill power to a Raspberry Pi without safely shutting down Linux first, file system corrupt can occur. I have seen this dozens of times over the years on my devices. When it happens, I just reformat the memory card and re-image it and continue.

But now we don’t have to — at least not if we are using the Pi Zero.

Abhinav Shukla created the PiZ SuperCap and launched it as a Kickstarter in 2021. The device is a small circuit board with a capacitor. It connects to a Pi Zero via the I/O header. Instead of plugging the USB power cable directly in to the Pi, you plug it in to the PiZ SuperCap board. This charges the capacitor then begins powering the Pi.

PiZ SuperCap by Abhinav Shukla

If power is disconnected, the capacitor has enough power to run the Pi Zero for a short amount of time (about 15 seconds). It also toggles a GPIO pin to indicate that power has been lost. By running a simple Python script on the Pi, you can now detect when power has been lost and, if it is not resumed in a set amount of seconds, safely shut down the Pi Zero by issuing a “shutdown” command.

I backed ten of these units, and I am glad I did. They work great! Now I can use a Pi Zero for any type of embedded project I want and just kill power when I want to shut down.

How it works

Here are some things to be aware of:

  1. When you first apply power, the Pi Zero will not immediately power up like you are used to. It must first charge the capacitor. The Pi Zero won’t actually start up until about 8 seconds after you apply power.
  2. When you turn off power, if no shutdown script is installed, the Pi won’t turn off until the capacity runs out. On my Pi Zero W (first version), I timed it at 75 seconds when just sitting at a text console login prompt. If you are running a graphical desktop or any programs using the CPU, it won’t last that long.
  3. The PiZ SuperCap will toggle GPIO pin 4 to indicate a power loss. To enable safe shutdown, you need some form of program or script that will monitor GPIO pin 4 and shut the system down when power goes away. The sample code provided uses a 9 second delay before deciding to power off, making the unit act like a mini UPS rather than shutting down on any temporary power blip. Clever.

Shutdown script

The designer provides some sample Python code on the project page, but here is a shorter one I came up with. I am not a Python programmer, so I have no idea if my technique is good. I just wanted it to block until it sees a power loss, and either shut down (after a delay) or continue:

supercap_shutdown.py

#!/usr/bin/python3

import RPi.GPIO as GPIO
import time
import os

GPIO.setmode(GPIO.BCM)
GPIO.setup(4, GPIO.IN)

while True:
        # Wait for power to drop.
        print ("Waiting for power loss...");
        GPIO.wait_for_edge(4, GPIO.FALLING)
        print ("Power loss")

        # Give user 9 seconds to restore power.
        print ("Waiting to see if power is restored...")
        time.sleep(9)

        if GPIO.input(4) == 0:
                print ("Power not restored.")
                os.system("sudo shutdown -h now")
                break;
        else:
                print ("Power restored.")

The above script has print statements in it so you can run it from the console to verify it is working. Those can be removed (or commented out) once you are sure it works.

You will need to do a “chmod +x supercap_shutdown.py” to make it executable.

If you want to make it run on startup, edit the /etc/rc.local file:

sudo pico /etc/rc.local

…and add this line at the end before “exit 0”:

# Pi Z SuperCap monitor script:
python /home/pi/supercap shutdown.py

You can then restart the system (“sudo restart now”) and when it reboots, the script should be running. Disconnect power and after 9 seconds (or whatever time you modify the script to use) it will issue a safe “shutdown” command.

3-D printed enclosure

I have created a very simple 3-D printer enclosure that holds a Raspberry Pi Zero and the PiZ SuperCap. Let me know if this is something you might want.

PiZ SuperCap by Abhinav Shukla in a custom 3-D printed enclosure by Sub-Etha Software.

Support the designer

If one of these might be of interest to you, consider backing the project at:

https://moonshotjr.com/moonfunded/project/piz-supercap/