Bipolar PI design

[Carl-NC]

Hi Carl, here the two impulses to are high, and for example if one is low and the other is high, then the picture will be the same as in the book, but won't the bugs then disappear, have you tested it this way? Thanks.

Fig 23-11 only applies to the differential coil driver (Fig 23-10), not the H-bridge driver. The H-bridge driver produces flyback only in one direction.

[detecto]

Where can i buy latest book? @Carl?

[Carl-NC]

Amazon for paperback. If you want hardback you might find it on Amazon or you can order it through a retail bookseller.

[pechkata]

It is very difficult to find anything in the book that is not very interesting.
But I immediately decided on the Bipolar PI design as the first thing to focus on.
There are several reasons for that.
First, I like the idea of a "non-invasive" detection method because of the possibility of adapting this project for demining purposes.
Second, this approach is only the tip of the iceberg of possibilities for further development of PI detectors. That is my humble opinion, there are probably many who will disagree with me.
I have tried to participate in the AMX thread a long time ago but felt deprived of the vast knowledge that the other participants in the thread undoubtedly have.
Then I realized that I still have a lot to learn in order to be an equal interlocutor with others.
Luckily Carl's book is here, I have plenty of time to read and study. And now I'm slowly (after all, I'm quite old) starting to understand some things.
That's why I think Bipolar PI design is a good initial step and an introduction to some broader knowledge.
Carl offered very interesting solutions that "provoke" further thoughts.
There are two parts in particular that I really like. TX and audio. I wouldn't dwell on audio for now, because I think it is necessary to make it first and only then draw conclusions.
At this point I would focus on TX first.
Originally from the book Bipolar PI TX looks like this:

I am also not very aware of this matter - especially with these chips - peak.
But I want the readers of this forum to look into your words.

[Detectorist#1]

Why you want this? What is the purpose? This is only one not very good solution for controlling of 4053 switches. Further improvements are more better.

[pechkata]

Hello Detector#1.
I didn't mean the scheme - but the words.
I dragged the scheme with the post - I don't know how it is cut.
Meaning - how can you comment on something you didn't do.
Nothing personal - sorry.

[ivconic]

A good friend mailed me two PICs from a neighboring state... a long time ago!
They never arrived.
It happens with our local mob called "Post and Customs".
So I am forced to think about alternative solutions.
Porting code from PIC to AVR is an interesting task.
These days I am preparing to try it.
But I'm out of shape!
Because I have an inflamed nerve in my shoulder that produces sharp pain and interferes with my focus.
I flipped through the book again and looked at the source code that Carl provided.
I had a problem at first because Carl is a huge fan of abbreviations that don't resemble anything!
And the worst thing is when he mixes up those abbreviations and uses one in the text and the other on the schematic!
Oooooahahahaha!
Carl... another suggestion for "Errata"...
Right now I'm an "idiot" and I don't understand anything.
Here is the graph and timing.
As a dumb programmer I need a literal translation of the train of pulses into the correct order.
Source code comments didn't help me much!
Carl is "The Duke of Obfuscation"!
His works do not need additional copyright protection!




And here is the first attempt at porting the code:
(I checked on an Atmega328P, it actually works!?? But you need to organize the timings better...)

Warning!
The code is under development, NEVER try it with mosfets if you are not a fan of white smoke!

#include <Arduino.h>

#define TXP_PIN 9   // Transmit Positive Pulse (PB1)
#define TXN_PIN 10  // Transmit Negative Pulse (PB2)
#define ADC_POT A1  // Potentiometer for delay adjustment
#define TCLKP 2     // Demodulator control (PD2)
#define GCLKP 3     // Demodulator control (PD3)
#define TCLKN 4     // Demodulator control (PD4)
#define GCLKN 5     // Demodulator control (PD5)
#define BCLK 6      // Charge pump control (PD6)
#define LED_PIN 13  // Debug LED (PB5)

// Timing constants for 5kHz PPS
#define TX_PERIOD 200  // 200µs period for 5kHz PPS
#define TX_WIDTH 50    // 50µs TX pulse width
#define MAIN_DELAY 15  // 15µs delay before sampling
#define MAIN_WIDTH 10  // 10µs main sample width
#define GND_DELAY 5    // 5µs ground delay
#define GND_WIDTH 40   // 40µs ground width

volatile unsigned long pulseDelay = MAIN_DELAY; // Default delay after main pulse

void setup()
{
    DDRB |= (1 << PB1) | (1 << PB2) | (1 << PB5); // Set TXP, TXN, and LED as outputs
    DDRD |= (1 << PD2) | (1 << PD3) | (1 << PD4) | (1 << PD5) | (1 << PD6); // Set control pins as outputs
    
    Serial.begin(115200);
    Serial.println("Bipolar PI (5kHz PPS)");
    
    noInterrupts();
    TCCR1A = 0;
    TCCR1B = (1 << WGM12) | (1 << CS10); // CTC mode, no prescaler
    OCR1A = 16000; // Match value for 1ms (assuming 16MHz clock)
    TIMSK1 |= (1 << OCIE1A); // Enable Timer1 interrupt
    interrupts();
}

ISR(TIMER1_COMPA_vect)
{
    PINB |= (1 << PB5); // Toggle LED for debug
}

void generatePulse()
{
    PORTB |= (1 << PB1);  // TXP HIGH
    PORTB &= ~(1 << PB2); // TXN LOW
    delayMicroseconds(TX_WIDTH); // TX pulse width
    PORTB &= ~(1 << PB1); // TXP LOW
    PORTB |= (1 << PB2);  // TXN HIGH
    delayMicroseconds(TX_WIDTH);
    PORTB &= ~(1 << PB2); // TXN LOW
    delayMicroseconds(pulseDelay); // Delay before demodulator signals
}



void controlChargePump()
{
    PORTD |= (1 << PD6);  // BCLK HIGH
    delayMicroseconds(50);
    PORTD &= ~(1 << PD6); // BCLK LOW
}

void loop()
{
    generatePulse();
    controlDemodulator();
    controlChargePump();
    delayMicroseconds(TX_PERIOD - (2 * TX_WIDTH + pulseDelay + MAIN_WIDTH + GND_DELAY + GND_WIDTH)); // Adjust cycle timing for 5kHz PPS
}


void controlDemodulator() // 25uS delay is just for testing purposes, don't be confused by this...
{
    PORTD |= (1 << PD2);  // TCLKP HIGH
    delayMicroseconds(25);
    PORTD &= ~(1 << PD2); // TCLKP LOW
        delayMicroseconds(25);
    PORTD |= (1 << PD4);  // TCLKN HIGH
    delayMicroseconds(25);
    PORTD &= ~(1 << PD4); // TCLKN LOW
        delayMicroseconds(25);
    PORTD |= (1 << PD3);  // GCLKP HIGH
    delayMicroseconds(25);
    PORTD &= ~(1 << PD3); // GCLKP LOW
        delayMicroseconds(25);
    PORTD |= (1 << PD5);  // GCLKN HIGH
    delayMicroseconds(25);
    PORTD &= ~(1 << PD5); // GCLKN LOW
    //delayMicroseconds(25);
    
   delayMicroseconds(MAIN_WIDTH); // Demodulator pulse width
  
}

/*
 * Reserved
 * void controlDemodulator()
{
    PORTD |= (1 << PD2);  // TCLKP HIGH
    PORTD &= ~(1 << PD4); // TCLKN LOW
  
    PORTD &= ~(1 << PD3); // GCLKP LOW
    PORTD |= (1 << PD5);  // GCLKN HIGH
    
    delayMicroseconds(MAIN_WIDTH); // Demodulator pulse width
    
    PORTD &= ~(1 << PD2); // TCLKP LOW
    PORTD |= (1 << PD4);  // TCLKN HIGH

    PORTD |= (1 << PD3);  // GCLKP HIGH
    PORTD &= ~(1 << PD5); // GCLKN LOW
}
 */

 /*
 *  void generatePulse() {
    PORTB |= (1 << PB1);  // TXP HIGH
    PORTB &= ~(1 << PB2); // TXN LOW
    delayMicroseconds(TX_WIDTH); // TX pulse width
    
    PORTB &= ~(1 << PB1); // TXP LOW
    delayMicroseconds(3); // DEAD TIME (adjustable)

    PORTB |= (1 << PB2);  // TXN HIGH
    delayMicroseconds(TX_WIDTH);
    
    PORTB &= ~(1 << PB2); // TXN LOW
    delayMicroseconds(3); // DEAD TIME before next cycle
}

MOSFET Switching Characteristics
IRF9640 (P-Channel)

Turn-On Delay (td(on)): ~25 ns
Turn-Off Delay (td(off)): ~100 ns
Total Switching Time: ~125 ns
IRF740 (N-Channel)

Turn-On Delay (td(on)): ~10 ns
Turn-Off Delay (td(off)): ~80 ns
Total Switching Time: ~90 ns
-P-MOSFETs (IRF9640) turn off slower (~100 ns delay)
-N-MOSFETs (IRF740) turn off slightly faster (~80 ns delay)
*************
-1 µs dead time between TXP OFF and TXN ON
-1 µs dead time between TXN OFF and next cycle
-Prevents shoot-through & protects MOSFETs

**************************************************
void generatePulse() {
    PORTB |= (1 << PB1);  // TXP HIGH
    PORTB &= ~(1 << PB2); // TXN LOW
    delayMicroseconds(TX_WIDTH); // TX pulse width

    PORTB &= ~(1 << PB1); // TXP LOW
    delayMicroseconds(1); // DEAD TIME (~1 µs)

    PORTB |= (1 << PB2);  // TXN HIGH
    delayMicroseconds(TX_WIDTH);

    PORTB &= ~(1 << PB2); // TXN LOW
    delayMicroseconds(1); // DEAD TIME (~1 µs)
}

************************************************
 */​

[Carl-NC]

If you see inconsistencies, then point them out. I'll fix them. Meanwhile, somewhere you will need to use interrupts or you won't be able to read the pots with the ADC.

[ivconic]

Exactly.
I didn't even use pot at this stage.
So I reserved the IRQ and put in a trivial LED for now.
There is no inconsistency, but the abbreviations for the pulses do not match on the graph, in the text and in the schematic.
Although I understand English, sometimes it takes me time to decipher some abbreviations and understand them properly.
...
I made a mistake in math (I was never good at it) and instead of 5kHz I'm measuring 1/2.


Hmmm... we need Joop here!
Porting is not easy for me...

[algan]

Hello Ivkonich, is the code for this generator? It is interesting..

[pechkata]

Exactly.
I didn't even use pot at this stage.
So I reserved the IRQ and put in a trivial LED for now.
There is no inconsistency, but the abbreviations for the pulses do not match on the graph, in the text and in the schematic.
Although I understand English, sometimes it takes me time to decipher some abbreviations and understand them properly.
...
I made a mistake in math (I was never good at it) and instead of 5kHz I'm measuring 1/2.


Hmmm... we need Joop here!
Porting is not easy for me...

Hello ivconic
Which key will you use to control the samplers?

[ivconic]

Algan no, it is for this currrent Bipolar PI, the topic is about.
...
Pechkata,
...potentiometer, but in this present code is not used yet.
Code is not finished, it is under development.
You can only test it with Atmga328P WITHOUT mosfets.
I thought it would be a good idea to make a code suggestion, so other members could contribute.

[algan]

between TXP turning off and TXN turning on.
The MOSFETs (IRF9640 and IRF740) specification shows that the P-channel MOSFET turns off a little slower (~100ns), while the N-channel turns off faster (~80ns).
Suggestion: We can increase the dead time to 1-2µs to make sure there is no overlap.
TXP and TXN switching order
We need to make sure that first TXP (the P-channel MOSFET) turns off, wait for the dead time, then TXN (the N-channel MOSFET) turns on.
The same applies when switching from TXN back to TXP.

---

void generatePulse() {
PORTB |= (1 << PB1); // TXP HIGH (P-channel MOSFET on)
PORTB &= ~(1 << PB2); // TXN LOW (N-channel MOSFET off)
delayMicroseconds(TX_WIDTH); // TX pulse width

PORTB &= ~(1 << PB1); // TXP LOW (turn off P-channel first)
delayMicroseconds(2); // DEAD TIME (~2 µs) - increased for safety

PORTB |= (1 << PB2); // TXN HIGH (turn on N-channel MOSFET)
delayMicroseconds(TX_WIDTH);

PORTB &= ~(1 << PB2); // TXN LOW (turn off N-channel MOSFET)
delayMicroseconds(2); // DEAD TIME (~2 µs) before next cycle
}

[Op04]

Hi,
gpz7k run tx with irfi 4227, there's a reason for this right?

[Carl-NC]

between TXP turning off and TXN turning on.
The MOSFETs (IRF9640 and IRF740) specification shows that the P-channel MOSFET turns off a little slower (~100ns), while the N-channel turns off faster (~80ns).
Suggestion: We can increase the dead time to 1-2µs to make sure there is no overlap.
TXP and TXN switching order
We need to make sure that first TXP (the P-channel MOSFET) turns off, wait for the dead time, then TXN (the N-channel MOSFET) turns on.
The same applies when switching from TXN back to TXP.

---

If we're talking about the circuit in Fig 22-18 (post 169) then the NMOS switches are self-driven. TXP & TXN refer to the positive and negative TX pulses, not the PMOS and NMOS switches.