My AGD analog detector work

[AhmatAhmatov1959]

Hello
After watching the film, I decided to build your household appliance. Currently, I am at the final stage of the PCB design. I am a bit worried, however, that it reacts to small Fe. Can you give me some advice? In the Netherlands, I use GMAXX-2, the range for small coins ends at 30 cm. I would like to build a detector that would have a longer range. Maybe you have some documentation in your collection.

Hello. Did you assemble this detector? How did the tests go? Thank you for your reply.

[Auto-Mation-Assist]

I have built a good number of these detectors and they all worked fine for detecting very small gold. I use the dig all method but usually disregard real loud return signals that are likely not what I'm looking for. One of the problems I ran into using the CMOS VCO and running it a idle frequency of 300Hz was that any substantial detection voltage of the incorrect polarity could drive the VCO frequency down so low that it could not be heard or stop outputting a tone. To eliminate this a resistor was added and installed that prevented this from occurring. This addition appears to have produced a issue when using the detector design in a way that I normally do not use.

If you look at any of the old circuit diagrams that I posted you will find the resistor to the far right of DSC connector (Discrimination Control), It was a 100K, 1206 size resistor in those early versions. The resistor could be R23, R24,or R97 depending on development version. This resistor needs to be removed to be able to properly use the DSC adjustment control on the front panel to null or reduce FE high level signals. I have verified this in my own AGD24.2 version.

Present work in progress:

I have been busy working getting 2026 version designed and have both its initial circuit boards lay out completed and verified but holding on in case I overlooked something. The 2026 version will offer a lot more flexibly due its build in STM32 processor. The STM32 C code that I worked on for the 2025.1 version can be used to allow for initial checkout after re-assigning some processor pins in the code. I hope to have the initial circuit boards here by start of December and assembly starting several weeks later if all parts are on hand. I have not yet decided on a PCB manufacturer.

[AhmatAhmatov1959]

[QUOTE=Auto-Mation-Assist;n441569]I have built a good number of these detectors and they all worked fine for detecting very small gold. I use the dig all method but usually disregard real loud return signals that are likely not what I'm looking for. One of the problems I ran into using the CMOS VCO and running it a idle frequency of 300Hz was that any substantial detection voltage of the incorrect polarity could drive the VCO frequency down so low that it could not be heard or stop outputting a tone. To eliminate this a resistor was added and installed that prevented this from occurring. This addition appears to have produced a issue when using the detector design in a way that I normally do not use.

If you look at any of the old circuit diagrams that I posted you will find the resistor to the far right of DSC connector (Discrimination Control), It was a 100K, 1206 size resistor in those early versions. The resistor could be R23, R24,or R97 depending on development version. This resistor needs to be removed to be able to properly use the DSC adjustment control on the front panel to null or reduce FE high level signals. I have verified this in my own AGD24.2 version.

Present work in progress:

I have been busy working getting 2026 version designed and have both its initial circuit boards lay out completed and verified but holding on in case I overlooked something. The 2026 version will offer a lot more flexibly due its build in STM32 processor. The STM32 C code that I worked on for the 2025.1 version can be used to allow for initial checkout after re-assigning some processor pins in the code. I hope to have the initial circuit boards here by start of December and assembly starting several weeks later if all parts are on hand. I have not yet decided on a PCB manufacturer.

Hello Auto-Mation-Assist . I made four detectors, but none of them were suitable for gold prospecting. I have a request and a question for you. Please share the AGD24 printed circuit board. It's possible to use bipolar pulses in this detector, since there was no response to ferrite or ground with bipolar pulses. Thank you for your reply. Sorry for my English.

[Altra]

It's possible to use bipolar pulses in this detector, since there was no response to ferrite or ground with bipolar pulses. .

Hello, This statement is not correct. Bipolar PI's respond to ground minerals the same as mono polar PI's. Bipolar PI's do eliminate the earth's magnet field without additional samples

[AhmatAhmatov1959]

Hello, This statement is not correct. Bipolar PI's respond to ground minerals the same as mono polar PI's. Bipolar PI's do eliminate the earth's magnet field without additional samples

I have limited experience making detectors. I made a Bipolar PI MD with a GB and DD coil. Even with a mono coil, it didn't respond to a 100-gram ferrite core. It didn't respond to red brick or soil, which my other metal detectors responded to. It gave a signal to a 5-gram copper coin 20 cm behind a red brick.I don't have any hot stones yet, so I don't know. The effect of bipolar pulses seems sufficient to warrant their use.

[Altra]

I made a Bipolar PI MD with a GB and DD coil. .

Okay, I thought you meant bipolar pulses automatically eliminate the ground minerals. Your circuit used GB elimination the same as would be used on mono polar pulse detector.

[AhmatAhmatov1959]

Okay, I thought you meant bipolar pulses automatically eliminate the ground minerals. Your circuit used GB elimination the same as would be used on mono polar pulse detector.

Hello Altra. Google Translate seems to be using artificial intelligence. For some reason, it cuts out words and sentences. And after editing, the meaning changes. I disabled the GB function in this detector. Using GB decreased the sensitivity. When I connected the DD coil, the ferrite reacted when I touched the coil.

[Detectorist#1]

It is not accidental that Carl mentioned in one of their last treads - all my new Pi designs are bipolar. This solves simultaneously many problems - the need of low noise (expensive) OpAmp for the first RX amplifier, the need of EFE samples from the control unit (and adding of additional RX parts for this) and the danger of firing of magnetic sensitive mines under the search coil.

[Carl-NC]

the need of low noise (expensive) OpAmp for the first RX amplifier,

Keep in mind that bipolar pulsing only cancels 1/f (flicker) noise. You still want to use a decent low-noise opamp.

[Auto-Mation-Assist]

I received the development circuit boards for version 26.0 of my AGD26-0 detector which consists of the back end signal handling board, and another circuit board that has all the front end circuitry and required regulated power supplies. The STM32G474 MCU and its power sources are on the back end signal handling board. All wiring that that normally route to the real panel have been eliminated since the coil connector and the Mono-DD coil switch are mounted directly to the front end board. This eliminates a problem with precision coil damping variations when inside or outside the enclosure due to changes in wire positions the changes in capacitance to surrounding parts and other issues that such wires inject.

As expected I have found a few things that I overlooked with version zero of my design after building and testing the back end processing board. These mostly have to do with routing the signals into the MPU analog to digital converters and the reference level of 2.048 volts to allow for 0.5mV raw signal detection with 12bits. The MPU code I have had running does use oversampling for its channels but I think I mat split the signals required to be processed into a plus 12 bit and minus 12 bit ADC channels. The design of AGD26 already has the required plus and minus signal splitters. This would be mostly adding some additional ADC drivers an coding work.

All the past versions of the AGD detectors have has a very fast 5 or 6 nano second comparator, in version 26 this will be the AD8611 which is faster than the one I have used in the past and was always used to provide a signal to the circuitry to start the receive gating after it determines at which point in the coils decay point is best to do so and then adding whatever the number of nano seconds is correct for the delay to activate the receive signal gates. That has been the main function of the decay comparator and not much happens is it does not give the proceed trigger. With the addition on the MPU it is now possible to also use it to determining the difference in the decay waveform caused by none-ferrous and ferrous material. These will cause the sample point in the decay waveform to change left for none-ferrous, consider it as being less damping, ferrous moving the trigger point right and as more damping. The left movement of the trigger point then looks like a under damped decay wave, while the ferrous right decay trigger point movement will increase decay time. When the output of the decay comparator is input into the MPU to can be compared to the TX pulse end time to see if the number of nano seconds deceased or increased and compared to normal idle time. The result can be used to potentially determine what type of metal is being detected. The direction movement for non-ferrous will be much smaller than the increased damping effect that ferrous material gives. This is all very nice and workable for it all depends on precision coil damping. This is were a few pf of capacitance and a small fraction of a ohm can make a difference in in ability to make such a detection.

In my next post I will show the circuitry simulations of actual circuits that the AGD26.1 is expected to end up with for active coil damping.

[Auto-Mation-Assist]

In this post I will show my results for coil decay tests in where I set the damping so that there was a 5mV level of under damped, negative going peaking for both a resistive load only and active loading like that planned for the AGD26.1 with two different active devices. The measurement point was the Signal-Out in the circuit diagram, and uses a resistive and diode clamp with an added 10pf pcb trace capacitance. The injected test signal ha a frequency of 1Mhz.

The resistive load only. Time to decay to the 5mV undershoot level from the end of the TX pulse:
Took 4.418us, Injected 1Mhz signal level recovery at 13.112uV

AGD26.1 circuit using LM7171. Time to decay to the 5mV undershoot level from the end of the TX pulse:
Took 2.684us, Injected 1Mhz signal level recovery at 10.166uV

AGD26.1 circuit using THS4631. Time to decay to the 5mV undershoot level from the end of the TX pulse:
Took 2.592us, Injected 1Mhz signal level recovery at 10.632uV

It is clear that a resistive only load when using a resistive and diode clamp into a RX front end gain stage presents a problem that is mostly due to the diodes becoming a open circuit near the most critical part of the receive signal. Increasing the value of the damping resistor will gain speed and get closer to active damping results but it will very rapidly increase the 5mv peak undershoot to a 100mv or more making the front end gain stages may not very happy. A low value resistor feeding the diode clamp will make this worse,while a high value resistor in creases noise and is really not very suitable with gain stages that have significant input offset currents.

I used the LM7171 for my initial tests for this application some time back but was never happy with it. Its output voltage limits is quite a be less than the rail voltages and it has significant input offset currents But it has the speed required for this application. The THS4631 appears to be a better choose with a decay time of 2.592us from the end of the 36us transmit pulse and that this speed decreased the signal loss by 0.466uV.

In the first case shown for the resistive damping load only the clamping diodes stopped conducting and caused a no load condition like a FET based front end and thus the signal level is higher.


The LM7171 5mV under damped test
Scope 5mV per division vertical, 1us per division horizontal.

T


The LM7171 Signal recovery test
Scope 5uv per division vertical, 1us per division horizontal.




The LM7171 Active damping adjusted for no over or under shoot. Decay to marker at 1mV level from 730 V coil volts in 2.631us
Scope 1mV per division vertical, 500ns per division horizontal.




Resistor R6 and trimmer are used to adjust damping. R6 fixed resistor value sets rough operating point, the trimmer R10 set damping, shown at 24.35% of 1000, so 243.5 ohms.
The active damping is in the green section. R9, R2, R37, R39 and the two clamping diodes D4A, D2B provide normal but not complete damping. C3 sims PCB trace capacitance.
C10 across coil at 50% normal, stepped 1pf or 2pf plus or minus to check decay detect none-ferrous or ferrous for decay comparator tests. Circuitry not shown.
1 Mhz test inject signal and C10 shown near top left.

[Auto-Mation-Assist]

I had a few more pictures that show the where the "positive" 1mV point was with the LM7171 and the 1Mhz test inject signal recovery.

Scope vertical 1mV per division, Horizontal 1us per division.

Now that I have good controlled damping we can convert the decay idle horizontal us counts at 2.685us and add or subtract the shift in position
of the marker, left or right with decay comparator. This will to help determine none-ferrous, Left movement of the yellow marker or ferrous,
right movement of the marker via conversion to in nano seconds difference from the end of the TX pulse.




In reference to the above picture:

Based on the position of the marker in relation the the curve in the decay show that we are in a good position for measuring very small amplitude
variations that may be caused my metallic objects as the coil is being swept back and forth if we were in the field. This would show up near the
bottom of the curve like the under or over damping peaks would but be very small and likely only visible on most test equipment after the signal
passes through the first gain stage. The use of test injects signal can help work around this and use that to verify proper RX gate timing.

----

The picture below shows the recovery of the 1Mhz test inject signal for the above picture.

Vertical scale is 5uv per division, and horizontal scale at 1us per division. Displayed signal level at markers is 9.954uV P/P.



End of post

[Detectorist#1]

Hello AMA,
With your projects, where every nanosecond and microvolt is under control, you are turning this work into art! The practical application of these high achievements should be discussed - that's my opinion. It will be interesting to see what some other participants in this discussion think.

[pito]

It will be interesting to see what some other participants in this discussion think.

For start I am thinking to build just transmitter with voltage in 500V range, can we have coil data amd Tx circuit ?

[Detectorist#1]

Hello pito,
For which project you ask - AGD or AMX? Here the discussion is for AGD project from AMA. The TX circuit is well described in the previous posts.