Electronic-Nulled Loops

[Altra]

Thanks for posting Mario! It will make a good study.

The digipot center taps labeled red and bla go to the preamp pcb in the coil. Where one goes to a series cap and the other to a series resistor. These are both connected to the rx coil entering the preamp. A small micro controller could probably replace most of the comparators and op amps.

[Mario-j23]

I've already done it, but I didn't recognize all the elements :-) [ATTACH]n418028[/ATTACH]

[dbanner]

I think I might be able to draw a layout of the auto null board based on the photos.
A schematic can then be drawn. It should be easy to discern what chips were used.
The component values aren't important. Just to grasp how the circuit is doing it's job.

[ivconic]

Yes too bad

Someone else on the forum has access to one. They posted about a month ago.
search Nautilus DMC IIB

[pito]

Use the CODE tag for code ("#" button in the editor bar):

void setup() {
pinMode( 9 , OUTPUT ); // Arduino Pin 9 = OCR1A
pinMode( 10 , OUTPUT ); // Arduino Pin 10 = OCR1B
}
​

Patent #2 = Needs program modification

// This code demonstrates how to generate two output signals
// with variable phase shift between them using an AVR Timer

// The output shows up on Arduino pin 9 and 10 (these are connected to the Timer1 A and B outputs respectively)

// More AVR Timer Tricks at http://josh.com

void setup() {

pinMode( 9 , OUTPUT ); // Arduino Pin 9 = OCR1A
pinMode( 10 , OUTPUT ); // Arduino Pin 10 = OCR1B

}


// prescaler of 1 will get us 4MHz - 488Hz
// User a higher prescaler for lower frequencies

#define PRESCALER 1
#define PRESCALER_BITS 0x01

// Output phase shifted waveforms on Arduino Pins 9 & 10
// half_period = 1/2 the period of both waveform in clock ticks (16Mhz clock default) note that the larger this number the lower the freq annd the more phase shift resolution you get
// shift = number of ticks delay between leading waveform and trailing. Must be < half_period
// invert_b- should we initially invert B? 1=A leads B, 0=B leads A
// To be completely out of phase, set shift=0 and invert_b=1

void startWaveforms( uint16_t half_period , uint16_t shift , byte invert_b ) {

TCCR1B = 0; // Turn off counter if it is on. Makes sure nothing happens while we are getting things set up.
// Leaves us in Normal mode, where we can do a Force Compare Match.

// First we need to get the phases of the two outputs set up correctly

// Set regs so when we force a compare it will match. This is the only way to set the state of the outputs on this chip.

//-------------------------------
// OCR2A = 127; //50% duty cycle
//OCR2B = 64; //about 25% duty cycle
OCR1A = 110;
OCR1B = 0;
TCNT1 = 0;


if (invert_b) {
TCCR1A = _BV( COM1A1 ) | _BV( COM1B1 ) | _BV( COM1B0 ); // Output A=0 and B=1 on compare match (which we will force presently)
} else {
TCCR1A = _BV( COM1A1 ) | _BV( COM1B1 ); // Output A=0 and B=0 on compare match (which we will force presently)
}

// Force a compare.
// If polarity=0, then make A=1, B=1
// If polarity=1, then make A=1, B=0
TCCR1C = _BV( FOC1A ) | _BV( FOC1B );

// Now that polarity is established, we can set things ready to run

// Both outputs into toggle mode. The output will toggle each time there is a compare match.
// Since they both toggle once per period, they will stay in whatever polarity they start in.
TCCR1A = _BV( COM1A0 ) | _BV( COM1B0 );

// OCR1A = 0;
OCR1A = 110;
OCR1B = shift;

// Set the counter to roll over on the first step. This will force a match when it rolls to 0.
TCNT1 = 0xffff;

// In CTC Waveform Generation Mode
// TCNT counts up to ICR1 then resets back to 0

ICR1 = half_period - 1; // We subtract 1 becuase when we get to ICR1 then on the next tick we go back to 0, so will make `period` ticks per cycle total.

// Turn on timer now
// Mode=CTC, clock=clkio/1 (no prescaling)
// Note: you could use a prescaller here for lower frequencies

TCCR1B = _BV( WGM13) | _BV( WGM12) | _BV( CS10 );

}

// Stop output, make both outputs low.

void stopWaveforms() {

TCCR1A = _BV( COM1A1 ) | _BV( COM1B1 ) ; // Output A=0 and B=0 on next compare match

}

// Min frequency is 16mhz / 65536 ticks per timer cyclke /2 timer cycles per waveform cycle ~= 122 hz

// freq_hz is frequency in hertz from 122 to 4,000,000
// shift_deg is phase shift between output A and output B in degrees (values outside -180 to +180 are normalized)

// Note that at frequencies above 44KHz you will loose precision on phase shift. At 4Mhz, everything is rounded to just 0 or +/-90 degrees.
// Note that all timing happens in 1/16Mhz increments, so frequencies that do not land on integer multipules of that will be aproximiate.
// Use startWaveforms() directly for more precise control.

void startQuadrature( unsigned long freq_hz , int shift_deg ) {

// This assumes prescaler = 1. For lower freqnecies, use a larger prescaler.

unsigned long clocks_per_toggle = (F_CPU / freq_hz) / 2; // /2 becuase it takes 2 toggles to make a full wave

// Normalize into 0-360 degrees

while ( shift_deg < 0 ) {
shift_deg += 360;
}

while ( shift_deg >= 360 ) {
shift_deg -= 360;
}

// Normalize a shift of more than 180 degress

byte invert = 0 ;

if (shift_deg >= 180 ) {

invert = 1 ; // Invert direction
shift_deg -= 180;

}

unsigned long offset_clocks;

offset_clocks = (clocks_per_toggle * shift_deg) / 180UL; // Do multiplication first to save precision

startWaveforms( clocks_per_toggle , offset_clocks , invert );

}

void stopQuadrature() {
stopWaveforms();
}



void loop() {

// 10KHz, A and B exactly out of phase
startQuadrature( 10000 , 180 ); //(-180 same result)
delay(10);
// stopQuadrature();
//delay(1000);

}​

[Altra]

Pity you didn't trace it out!

Yes too bad

Someone else on the forum has access to one. They posted about a month ago.
search Nautilus DMC IIB

[Carl-NC]

Use the CODE tag for code ("#" button in the editor bar):

void setup() {
  pinMode( 9 , OUTPUT );    // Arduino Pin 9  = OCR1A
  pinMode( 10 , OUTPUT );   // Arduino Pin 10 = OCR1B
}
​

[pito]

Patent #2 = Needs program modification




// This code demonstrates how to generate two output signals
// with variable phase shift between them using an AVR Timer

// The output shows up on Arduino pin 9 and 10 (these are connected to the Timer1 A and B outputs respectively)

// More AVR Timer Tricks at http://josh.com

void setup() {

pinMode( 9 , OUTPUT ); // Arduino Pin 9 = OCR1A
pinMode( 10 , OUTPUT ); // Arduino Pin 10 = OCR1B

}


// prescaler of 1 will get us 4MHz - 488Hz
// User a higher prescaler for lower frequencies

#define PRESCALER 1
#define PRESCALER_BITS 0x01

// Output phase shifted waveforms on Arduino Pins 9 & 10
// half_period = 1/2 the period of both waveform in clock ticks (16Mhz clock default) note that the larger this number the lower the freq annd the more phase shift resolution you get
// shift = number of ticks delay between leading waveform and trailing. Must be < half_period
// invert_b- should we initially invert B? 1=A leads B, 0=B leads A
// To be completely out of phase, set shift=0 and invert_b=1

void startWaveforms( uint16_t half_period , uint16_t shift , byte invert_b ) {

TCCR1B = 0; // Turn off counter if it is on. Makes sure nothing happens while we are getting things set up.
// Leaves us in Normal mode, where we can do a Force Compare Match.

// First we need to get the phases of the two outputs set up correctly

// Set regs so when we force a compare it will match. This is the only way to set the state of the outputs on this chip.

//-------------------------------
// OCR2A = 127; //50% duty cycle
//OCR2B = 64; //about 25% duty cycle
OCR1A = 110;
OCR1B = 0;
TCNT1 = 0;


if (invert_b) {
TCCR1A = _BV( COM1A1 ) | _BV( COM1B1 ) | _BV( COM1B0 ); // Output A=0 and B=1 on compare match (which we will force presently)
} else {
TCCR1A = _BV( COM1A1 ) | _BV( COM1B1 ); // Output A=0 and B=0 on compare match (which we will force presently)
}

// Force a compare.
// If polarity=0, then make A=1, B=1
// If polarity=1, then make A=1, B=0
TCCR1C = _BV( FOC1A ) | _BV( FOC1B );

// Now that polarity is established, we can set things ready to run

// Both outputs into toggle mode. The output will toggle each time there is a compare match.
// Since they both toggle once per period, they will stay in whatever polarity they start in.
TCCR1A = _BV( COM1A0 ) | _BV( COM1B0 );

// OCR1A = 0;
OCR1A = 110;
OCR1B = shift;

// Set the counter to roll over on the first step. This will force a match when it rolls to 0.
TCNT1 = 0xffff;

// In CTC Waveform Generation Mode
// TCNT counts up to ICR1 then resets back to 0

ICR1 = half_period - 1; // We subtract 1 becuase when we get to ICR1 then on the next tick we go back to 0, so will make `period` ticks per cycle total.

// Turn on timer now
// Mode=CTC, clock=clkio/1 (no prescaling)
// Note: you could use a prescaller here for lower frequencies

TCCR1B = _BV( WGM13) | _BV( WGM12) | _BV( CS10 );

}

// Stop output, make both outputs low.

void stopWaveforms() {

TCCR1A = _BV( COM1A1 ) | _BV( COM1B1 ) ; // Output A=0 and B=0 on next compare match

}

// Min frequency is 16mhz / 65536 ticks per timer cyclke /2 timer cycles per waveform cycle ~= 122 hz

// freq_hz is frequency in hertz from 122 to 4,000,000
// shift_deg is phase shift between output A and output B in degrees (values outside -180 to +180 are normalized)

// Note that at frequencies above 44KHz you will loose precision on phase shift. At 4Mhz, everything is rounded to just 0 or +/-90 degrees.
// Note that all timing happens in 1/16Mhz increments, so frequencies that do not land on integer multipules of that will be aproximiate.
// Use startWaveforms() directly for more precise control.

void startQuadrature( unsigned long freq_hz , int shift_deg ) {

// This assumes prescaler = 1. For lower freqnecies, use a larger prescaler.

unsigned long clocks_per_toggle = (F_CPU / freq_hz) / 2; // /2 becuase it takes 2 toggles to make a full wave

// Normalize into 0-360 degrees

while ( shift_deg < 0 ) {
shift_deg += 360;
}

while ( shift_deg >= 360 ) {
shift_deg -= 360;
}

// Normalize a shift of more than 180 degress

byte invert = 0 ;

if (shift_deg >= 180 ) {

invert = 1 ; // Invert direction
shift_deg -= 180;

}

unsigned long offset_clocks;

offset_clocks = (clocks_per_toggle * shift_deg) / 180UL; // Do multiplication first to save precision

startWaveforms( clocks_per_toggle , offset_clocks , invert );

}

void stopQuadrature() {
stopWaveforms();
}



void loop() {

// 10KHz, A and B exactly out of phase
startQuadrature( 10000 , 180 ); //(-180 same result)
delay(10);
// stopQuadrature();
//delay(1000);

}

[pito]

Patent #1 = Supper D




In my case ground cancelation 90%

[ivconic]

What I posted can be automated. The Nautilus 2b, Tyndall Electronics used two digipots to null the coil. No micro was involved. It used cmos up/down counters and a halfwave rectifier to sense the rx voltage. I had a 2b and started reversing the nulling pcb, it was a rats nest. I ended up selling the unit before making a schematic. The Nautilus did not use a opamp inverter. The tx oscillator was configured as a Hartley using a center tapped coil. The nulling pots were placed across the tx leads giving 180 degrees of adjustment Most of the posted schematics are incorrect in this section.


Another interesting coil setup is the Minelab Gofind.

Pity you didn't trace it out!

[Altra]

What I posted can be automated. The Nautilus 2b, Tyndall Electronics used two digipots to null the coil. No micro was involved. It used cmos up/down counters and a halfwave rectifier to sense the rx voltage. I had a 2b and started reversing the nulling pcb, it was a rats nest. I ended up selling the unit before making a schematic. The Nautilus did not use a opamp inverter. The tx oscillator was configured as a Hartley using a center tapped coil. The nulling pots were placed across the tx leads giving 180 degrees of adjustment Most of the posted schematics are incorrect in this section.


Another interesting coil setup is the Minelab Gofind.

[GeoMax]

This one is working very good, pot = 100k, R= 10k, C = 1nF for 8kHz, on the front I added a potentiometer for amplitude regulation.

Hi all,

google for "gyrator circuit"... there you will find more information on this circuit and what you can do with it...

[ivconic]

What would be the possible range of values for that R ?
Digipot is one solution, with some limitations.
The second thing that comes to mind is the mosfet in the role of variable resistance.
But it all depends on the possible value span for R.
The third (many won't like) is a series of resistors that are added/subtracted via digital switches.
I personally prefer a series of small transistors that simply turn R on/off in series.
...
Have any of you ever paid attention to the series of Russian DIY detectors such as Quant, Fortuna, Soha...
I did and I made Quant and Fortuna (or was it Soha, I can't remember).
I don't have the source code, only HEX, so I can't say what was done there.
But both have a procedure of "adaptation" to any coil, which reminds a lot of what Carl wrote.
The procedure/function is by calling from the service menu.
It starts at ~1kHz and goes up to 20kHz and measures the voltage on the RX coil. And finds the most suitable frequency for the given coil.
I really liked it and tested it extensively on multiple coils. It works almost flawlessly.
It always finds the best frequency for a given coil. And the detector works great afterwards.
Just the best solution for a multi-f detector.
Only, in that case, the "adaptation" process would have to happen very quickly and synchronously with the TX waveform.
Which is not a problem lately with the flood of very powerful processors and ADC circuits with affordable prices (affordable for professionals and not for amateurs).

[profesor]

Another method is arado120 detector.
A variable capacitor placed between tx and rx.
It was zeroing the coil against slips and deviations.

https://www.geotech1.com/forums/file...=366974.

[pito]

This one is working very good, pot = 100k, R= 10k, C = 1nF for 8kHz, on the front I added a potentiometer for amplitude regulation.