PI backend Atmega328P and firmware

[Gunghouk]

Wrong address for the LCD, address is 0x27 for my 1602 display rather than 0x3F

main.c

line 160 i2c_lcd_init(0x27, 0, 1, 2, 4, 5, 6, 7, 3); // set the LCD address to 0x3F for a 16 chars and 2 line display

Code is running on an Uno

Also reduced I2C clock from 600kHz to 100kHz for more reliable operation.

twimaster.h

line 26 /* I2C clock in Hz */
line 27 #defineSCL_CLOCK100000L // PCF8574 can work from 100KHz to 400KHz max.

[Teleno]

Wrong address for the LCD, address is 0x27 for my 1602 display rather than 0x3F

Those are things each one has to know and modify accordingly. There's no universal address and there's no routine included to automatically find the right one. It may be added to the init code, of course.

Also reduced I2C clock from 600kHz to 100kHz for more reliable operation.

twimaster.h

line 26 /* I2C clock in Hz */
line 27 #defineSCL_CLOCK100000L // PCF8574 can work from 100KHz to 400KHz max.

Not recommended to go too low because the time budget of 1ms (for a 1KHz PI) can be exceeded and the whole thing becomes unreliable (ADC readings missed, etc).
Stay a 400kHz minimum.
The code is sending only 1 characer per transmit period to stay within the time budget.

[ivconic]

Very simple code to find the correct addresses of all "devices" on the I2C port. I always use that when I don't know the address.

#include <Wire.h>
void setup()
{
  Wire.begin();
  Serial.begin(9600);
  Serial.println("\nI2C Scanner");
}
void loop()
{
  byte error, address;
  int nDevices;
  Serial.println("Scanning...");
  nDevices = 0;
  for(address = 1; address < 127; address++ )
  {
    Wire.beginTransmission(address);
    error = Wire.endTransmission();
    if (error == 0)
    {
      Serial.print("I2C device found at address 0x");
      if (address<16)
        Serial.print("0");
      Serial.print(address,HEX);
      Serial.println("  !");
       nDevices++;
    }
    else if (error==4)
    {
      Serial.print("Unknow error at address 0x");
      if (address<16)
        Serial.print("0");
      Serial.println(address,HEX);
    }    
  }
  if (nDevices == 0)
    Serial.println("No I2C devices found\n");
  else
    Serial.println("done\n");
 
  delay(5000);          
}​

[Teleno]

main.c adapted to automatically find the address of the LCD.

void init() {
  if (use_lcd) {
    // find the I2C address of the LCD
    i2c_init();
    uint8_t address = 1;
    while (i2c_start(address, I2C_READ) == 1 && address < 128) address++;
    /* prepare LCD */
    i2c_lcd_init(address, 0, 1, 2, 4, 5, 6, 7, 3);  // set the found LCD addressfor a 16 chars and 2 line display​​ 

main.zip

[Teleno]

When doing changes to the code check the signals at PD7 and PB1 (OC1A) on the scope.

PB1 is the transmit period.
PD7 is high from the begin to the end of the main() loop.
If PD stays high for longer than the period at PB1 then you have a problem, the time budget has been exceeded.

Additionally, PB2 (OC1B) shows the ADC conversions, from the trigger event to the end of the interrupt. Actual sampling is delayed 2 x ADC_CLK cycles after the trigger (rising edge of PB2).

[Gunghouk]

Code runs on the Nano no problem but I've noticed a fixed 4uS difference between the actual 1st sample and the delay setting ie. 6uS for a setting of 10 uS which stays the same throughout the delay setting range. Is this the 2*ADC_CLK mentioned ?

[Teleno]

Code runs on the Nano no problem but I've noticed a fixed 4uS difference between the actual 1st sample and the delay setting ie. 6uS for a setting of 10 uS which stays the same throughout the delay setting range.

Yes. The ADC is auto triggered by compare match B of Timer 1. According to the datasheet, page 208,

When auto triggering is used, the prescaler is reset when the trigger event occurs. This assures a fixed delay from the trigger event to the start of conversion. In this mode, the sample-and-hold takes place two ADC clock cycles after the rising edge on the trigger source signal. Three additional CPU clock cycles are used for synchronization logic.

So the OC1B pin goes high when the match occurs and triggers the ADC, but the OC1B setting has to be 4us earlier than the actual sampling given by the desired delay in "Sens.delay"

ADC_CLK is 500KHz, 2us cycle, so the lag from the compare match has to be 4us.

ISR(ADC_vect) {
...
...
    case MODE_EF:
      OCR1B = pulseWidth + Sens.delay - 4 * MICROSECOND;​​ 

Actually it should be 4 * MICROSECOND + 3, to add the extra three main clock cycles, but let's not be that picky.

As I said before, you must take this into account when you check the signals on the scope,

Additionally, PB2 (OC1B) shows the ADC conversions, from the trigger event to the end of the interrupt. Actual sampling is delayed 2 x ADC_CLK cycles after the trigger (rising edge of PB2).

If using the 555 to drive the gate, add the group delay of the 555 which is about 27 main clock cycles (1.7 us)

OCR1B = pulseWidth + Sens.delay - 4 * MICROSECOND - 3 - 27; should account for the exact delay from the gate pulse pin to the actual gate pulse at the MOSFET.

[Gunghouk]

So displayed delay should actually have 4uS subtracted from it ?

[Teleno]

So displayed delay should actually have 4uS subtracted from it ?

The displayed delay is in the variable "Sens.delay" and is correct. The code subtracts 4us from it before setting OCR1B.

On the scope you add 4us to the rising edge of PB2 to know the actual moment of sampling, which should be the same as "Sens.delay".

Example: if you set the delay to 10us on the display, the signal PB2 on the scope rises at 6us, but the actual sampling takes place at 6us + 4us = 10us.

[Gunghouk]

[Teleno]

New version changes:
Maximum and minimum sound frequencies configurable in "log_sound.h" defines.

#define FREQ_LOW 400
#define FREQ_HIGH 1000​ 

Optimized do_log function (exec time from 60us down to 30us).
LogTable256 in PROGMEM to save RAM.

uint32_t do_log(uint32_t x) {
    if (x == 0) return 0;
    uint8_t n = 0;
    uint32_t temp, log_result;
    // Create an array pointer of uint8_t pointing to the bytes of x
    const uint8_t *bytes = (const uint8_t *)&x;
    // Find position of highest set bit using direct byte access
    if (bytes[3] != 0)  n = 24 + pgm_read_byte(&LogTable256[bytes[3]]);
    else if (bytes[2] != 0) n = 16 + pgm_read_byte(&LogTable256[bytes[2]]);
    else if (bytes[1] != 0) n = 8 + pgm_read_byte(&LogTable256[bytes[1]]);
    else n = pgm_read_byte(&LogTable256[bytes[0]]);
  
    // Simplified calculation for log_result
    temp = ((PRECISION > n) ? x << (PRECISION - n) : x >> (n - PRECISION)) - ONE;
    log_result = (n << PRECISION) + ((temp - ((temp * temp) >> (PRECISION + 1))) * LN2_INVERSE >> PRECISION);
    return log_result;
}​ 

interrupt routines and associated defines moved to a header file to declutter main.c

[Gunghouk]

Both main.c and main.zip the same ?

[Teleno]

Both main.c and main.zip the same ?

No, replace everything.

[Gunghouk]

I meant inside the zip are main.c and main.zip. Which should be used ?

[Teleno]

I meant inside the zip are main.c and main.zip. Which should be used ?

Delete main.zip