Uni-PI - Universal Pulse Induction Metal Detector board for microcontrollers like Ardiuno, PIC, AVR.

[Waikiki_Sweep]

This project desperately need some, at least basic form of GB\disc control. Very little change is needed, two sample subtraction using same integrator. I have working hardware on table, analog, not uC controlled, will post more details, actually tempted to add this to minipulse project.

Yes, sure. That is why I am using differential integrator and two sampling switches schematic in Uni PI. With timing driven by microcontroller any complex sequences of target and ground samples can be realized in few clicks.
Also with using low noise opamp ISL28291, didital processing of signal, new sampling ideas there is a hope to improve depth, speed of reaction on targets and other parameters with using cheap and simple schematic.

[Qiaozhi]

This project desperately need some, at least basic form of GB\disc control. Very little change is needed, two sample subtraction using same integrator. I have working hardware on table, analog, not uC controlled, will post more details, actually tempted to add this to minipulse project.

Regarding the Minipulse project ..... sorry about the delay in posting the rest of the information. Hopefully it will get to the top of my To-Do List in a few days.

[Tepco]

This project desperately need some, at least basic form of GB\disc control. Very little change is needed, two sample subtraction using same integrator. I have working hardware on table, analog, not uC controlled, will post more details, actually tempted to add this to minipulse project.

I mentioned this circuit for GB, sorry for hand written quality but this is still on prototyping board. I guess I have to explain lot about, but in short: This is simple analog solution, can be easily added to almost any classic detector, but with MCU controlled timing can be even better. Instead of using two channels, here second sample subtraction is achieved using single integrator. Earlier I tried similar system using one integrator and GB adjustment achieved by varying preamp gain, but this was bipolar pulsing detector and differential 2 stage AC coupled amplifier, for classic detector this is out of question. Instead, I used integrator to produce subtraction. First sample closes S1 switch for 18uS (not quite critical, just convenient RC value) as in standard PI, integrator is connected via 10K resistor. After some delay, 15-20uS, another 18uS GB sample is generated, closing S3 and feeding signal to opposite integrator input, via 1.5K fixed resistor and 4k7 GB adjusting pot, so this ratio (10K\ 1.5K+adjustment) determine subtraction ratio. Also, this will screw up EF sample balance. To fix this, after 150uS inter sample delay, both S2 and S4 switches are closed to opposite ends of integrator, via same component values, compensating proportionally S1 and S3 contribution during first two samples. Integrator output will now be bipolar, going in one direction for short TC, in another for long TC, and remain fixed for certain TC, that of ground signal. Rest of circuit must be designed to accept this, in analog incarnation, full wave rectifier between SAT circuit and threshold circuit and VCO controlled from SAT output. This will produce high-low-high, or low-high-low tone for different objects. In case of MCU control, with signal digitized after SAT this is easy to be done in software. In this version timing is fixed and integrator gain changes, but gain can be held fixed and GB adjustment achieved by variable timing, more suitable for MCU design. Unlike two channel approach (requiring two channels to be as close as possible), here timing can be bit more “liberal”. Surprisingly, this appear to work just fine, at least on prototyping board, can be used to add GB to surf for example (in essence, just another 4093 is needed) but MCU control can be even better, nice platform for experiments.

[Qiaozhi]

Or you could try this method from mickstv ->
http://www.geotech1.com/forums/showt...ground+balance

[Tepco]

This is the same thing in essence, just method of realization may vary.

[Qiaozhi]

This is the same thing in essence, just method of realization may vary.

Yes, I know. It's just a different way to skin the proverbial cat.

[Tepco]

Now this is just a competition for simplest way to do that.

[6666]

VCO controlled from SAT output. This will produce high-low-high, or low-high-low tone for different objects.

That would be nice.

[Tepco]

Two notorious problems are always present with this approach. First one is, when object is first detected above threshold, it will be always detected as fast TC, no matter of what they are made of, because only first sample contribute. Only with stronger signal, at some 70-80% range, it can change sign to proper ID. Some sort of proportional audio seems to be best solution, not comparator driven high-low response. Like, 800Hz neutral for week signal at threshold level, then going up or down to say, 1600 or 300Hz.Output of SAT amplifier is just about right amplitude to drive 4046 VCO directly. Another problem is preamplifier overloading with strong enough signal, producing wrong audio response. Aside simple rising the coil, I made quick fix: timing is not synchronized from TX, but from RX amplifier output, at some 1V below point when it goes out of saturation, then entire timing sequence will shift from this point.

[Waikiki_Sweep]

Here is probable schematics of GB1 with using 14 pin microcontroller ATTiny84. Also VCO with microcontroller is relatively easy to make. I have tried VCO with some results already.

[6666]

I mentioned this circuit for GB, sorry for hand written quality but this is still on prototyping board. I guess I have to explain lot about, but in short: This is simple analog solution, can be easily added to almost any classic detector, but with MCU controlled timing can be even better. Instead of using two channels, here second sample subtraction is achieved using single integrator. Earlier I tried similar system using one integrator and GB adjustment achieved by varying preamp gain, but this was bipolar pulsing detector and differential 2 stage AC coupled amplifier, for classic detector this is out of question. Instead, I used integrator to produce subtraction. First sample closes S1 switch for 18uS (not quite critical, just convenient RC value) as in standard PI, integrator is connected via 10K resistor. After some delay, 15-20uS, another 18uS GB sample is generated, closing S3 and feeding signal to opposite integrator input, via 1.5K fixed resistor and 4k7 GB adjusting pot, so this ratio (10K\ 1.5K+adjustment) determine subtraction ratio. Also, this will screw up EF sample balance. To fix this, after 150uS inter sample delay, both S2 and S4 switches are closed to opposite ends of integrator, via same component values, compensating proportionally S1 and S3 contribution during first two samples. Integrator output will now be bipolar, going in one direction for short TC, in another for long TC, and remain fixed for certain TC, that of ground signal. Rest of circuit must be designed to accept this, in analog incarnation, full wave rectifier between SAT circuit and threshold circuit and VCO controlled from SAT output. This will produce high-low-high, or low-high-low tone for different objects. In case of MCU control, with signal digitized after SAT this is easy to be done in software. In this version timing is fixed and integrator gain changes, but gain can be held fixed and GB adjustment achieved by variable timing, more suitable for MCU design. Unlike two channel approach (requiring two channels to be as close as possible), here timing can be bit more “liberal”. Surprisingly, this appear to work just fine, at least on prototyping board, can be used to add GB to surf for example (in essence, just another 4093 is needed) but MCU control can be even better, nice platform for experiments.

Hi Tepco, did you get the gb to work ?
what is the tx pulse rate (pps) thanks

[6666]

I have programmed a small pic to do the timing, seems to work but I may have to go to faster pic as I'm not sure how super critical the 18uS timing is
Tepco from your GB circuit could you please tell me what target gives a short time constant and what target gives a long time constant ? thanks.

[WM6]

Hi 6666
Seems Tepco's last visit here was two years ago.

[6666]

Hi WM6
thanks looks like he left us , cheers.

[kt315]

ordinary case. somebody has give well good start and ended without result. ivitsa am I right?
where is you my friend ivconic... live or had dead )))