Bipolar PI design

[h9361]

Yes, that method will also work.

Thank you.

[Carl-NC]

Yes, that method will also work.

[h9361]

This is where pot values get converted into timing values:

Bool ADC_Run(void)
{
if(ADC_GetPot1() || ADC_GetPot2()) // If either pot has changed value
{
Settings.TgtDelay = ( 50 + Pot1Value) << 1; // Min = 50 (12.5us), max = 113 (28.25us)
Settings.TxWidth = ( 50 + Pot2Value) << 3; // Min = 50us, max = 305us
Settings.TxPeriod = (200 + (Pot2Value << 2)) << 3; // Min = 200us, max = 1220us
return True;
}
return False;
}​

The problem is, a minimum TxPeriod of 50us (20kHz) requires reducing everything else to fit. You would first need to drop the TxWidth down to, say, 10-15us and go from there.

Hello Dear Carl
I have a question about low/high tone for discrimination. Please see attached, we can also use such GB, subtraction operation because i am using such method for my PI and it is working very nice. But you prefer to use a different method for your Bipolar-PI.
My main question is that, do you think the high/low Disc can also work on the GB method?

[eduardo1979]

The diagram is in the forum. You have to read the thread from the beginning and stop at my last post.

[h9361]

Hi Ivconic.
I've recorded a new video.
Watch it at the link below.

https://www.youtube.com/watch?v=TKm-Gim7iC8&t=3s

My coil is 11 inches for GPX, and I've installed a 330 Ohm damping resistor. If the damping resistor isn't adequate, the detector won't discriminate. Pay close attention to the TX and RX cable connections and don't reverse them, because discrimination won't work either. Look carefully at the schematic. On mine, I marked them with numbers 1 and 2, and they should match. The frequency is 2 kHz and the delay is 10 uS. Carl's original circuit also discriminates between high- and low-conductivity metals by sound. Watch Carl's video and you'll see. The 74HC4053 isn't the problem, as you describe. The noise you hear in the speaker at low volume is due to the audio driver. I assume it needs to be improved a bit, but that's not the problem with the detector. The cable connecting the coil to the detector must be a stereo audio cable. I haven't protected the coil yet because I'd like to run more tests.

Hello Dear eduardo1979
I watched your video, can you send me your schematic?
Thank you

[vasilink]

Hello everyone,

I'm new to Geotech1 and this is one of my first technical posts — I appreciate your patience. I'd like to share a simple bridge driver circuit for a PI metal detector, built using bipolar transistors. It’s not necessarily optimal, but it’s easy to assemble and uses widely available components. I’ve attached the schematic below.

My goal is to build a deep-seeking detector focused on large targets. The pulse duration is planned at 250 µs, with a repetition rate of 300 Hz. I’m planning to sample the decay signal entirely in software, with a delay of 20–50 µs after the pulse and a sampling window of about 200 µs. I’m not using analog sample-and-hold — the full decay waveform will be digitized and processed in the microcontroller.

Controller: RP2040, powered at 3 V. The signal will be processed strictly within the 0–3 V range, matching the ADC input limits of the RP2040.

So far, I’ve implemented the bridge driver and basic preamplifier. Now I’m looking for advice on the next stage — amplifying the decay signal before ADC:

• What gain distribution across analog stages would be optimal to avoid clipping within the 0–3 V range?
• Which op-amps would you recommend for low-noise amplification in this context?
• Any suggestions for filtering or shaping the decay signal before digitization?

After ADC, I plan to implement:

• Integral accumulation and averaging to improve SNR.
• Digital amplification and processing.
• And most importantly, target discrimination based on the difference in decay responses under opposite pulse polarities — since ferromagnetic targets tend to produce asymmetric responses, unlike non-ferrous metals.

I’d be grateful for any suggestions, schematics, or links to relevant discussions. Also, feel free to critique the bridge driver — I’m open to improvements.

Thanks for reading!

[pechkata]

Hi Orbit,
Yes, but you must have in mind that Smart Pulse Hybrid uses IB coils and this project uses more simple mono coil with very good results. The difference in the coils is very important for DIY projects.

Timing Bipolar 3 TC4426.rar

[Carl-NC]

This is where pot values get converted into timing values:

Bool ADC_Run(void)
{
  if(ADC_GetPot1() || ADC_GetPot2())    // If either pot has changed value
  {
    Settings.TgtDelay = ( 50 + Pot1Value) << 1; // Min = 50 (12.5us), max = 113 (28.25us)
    Settings.TxWidth  = ( 50 + Pot2Value) << 3; // Min = 50us, max = 305us
    Settings.TxPeriod = (200 + (Pot2Value << 2)) << 3; // Min = 200us, max = 1220us
    return True;
  }
  return False;
}​

The problem is, a minimum TxPeriod of 50us (20kHz) requires reducing everything else to fit. You would first need to drop the TxWidth down to, say, 10-15us and go from there.

[algan]

Hi Carl, is it possible to change the bipolar code so that the frequency starts at 600 hertz, as it does now, but reaches 15-20 kilohertz? There must be some limit, but what do I need to change to increase it? Thank you.

[ivconic]

So we've done wonderful number one... it's time for wonderful number two...

[ivconic]

[ivconic]

[eduardo1979]

Hi eduardo1979,
Let to assembly my schematic in "Bipolar PI MD with GB and DD coil" and to make serious tests in real different ground conditions. All other comments are only words.

Hello Detectorist #1. You're right. We'll work on your design and conduct the necessary tests in real-world conditions.

[Carl-NC]

Generally, a higher pulse frequency helps as it integrates more samples. Eric's Gold Quest ran at 10kHz. Just watch the power consumption, reduce the pulse width as needed.

[algan]

I don't know if increasing the frequency would help for smaller goals, like starting at 4 kilohertz and then going up to 8 kilohertz, but let Carl say if that's possible and what effect it would have.