Factors effecting detection range of VLF coil

[americanelliott]

I am designing a metal-detector for a senior project in electrical engineering. For the coil, I went with the VLF DD design in the Inside the Metal Detector book. We have an oscillator circuit that delivers 500mA at 10kHz (square wave) to the coil. The detection circuit amplifies the receive signal to a max of 3.3V. When we test the coils through an o-scope with metal, we are getting noticeable results with large pieces of metal about 10-18 cm from the coils, and anything beyond that is unnoticeable.
We are unsure as to what factors effect detection range and how we can possibly modify our coils or circuit to improve the detection range. The coils will be easier to modify since the circuit is a PCB and major changes to it are make-shift since we can't change the layout of the PCB any more. Any help is appreciated.

[Gwil]

Can you post a schematic of the detector as it is? Without that, I don't think anyone will be able to help you.

[Qiaozhi]

I am designing a metal-detector for a senior project in electrical engineering. For the coil, I went with the VLF DD design in the Inside the Metal Detector book. We have an oscillator circuit that delivers 500mA at 10kHz (square wave) to the coil. The detection circuit amplifies the receive signal to a max of 3.3V. When we test the coils through an o-scope with metal, we are getting noticeable results with large pieces of metal about 10-18 cm from the coils, and anything beyond that is unnoticeable.
We are unsure as to what factors effect detection range and how we can possibly modify our coils or circuit to improve the detection range. The coils will be easier to modify since the circuit is a PCB and major changes to it are make-shift since we can't change the layout of the PCB any more. Any help is appreciated.

It will be difficult to see small changes on the oscilloscope by simply looking at the preamp output. Have a look at the Raptor project in Chapter 9 of ITMD. You will see that there is a lot more circuitry required beyond the preamp in order to extract and boost the miniscule signal from a metal target.

[Skippy]

Raw signals from the coil are very small. As a vague guide, a US 25c 'quarter' may give an output change of 2 millivolts when it's 25mm / 1 inch from the coil.
A similar-sized iron/steel item ( eg. a washer ) would give more response. Good commercial detectors will trigger on targets producing signal changes of 5 microvolts. For example the 25c coin at 250mm / 10" distance.

[americanelliott]

Here are the relevant schematics for the oscillator and detection circuits. Our signal is processed by an MSP430F5507 MCU. I will need to see if it can detect changes in a signal that are that small.

[Skippy]

You only have a pre-amp gain of 10 , less in practice as your LPF on it's output has an f_c = 10 kHz, so you're attenuating some of the fundamental as well as the harmonics.
Increasing gain to 100 up to 500 would be a good start. Do this with a second amplifier, don't have one high-gain stage. If your coil null output is below 5 mV peak-peak, your ADC should cope with that.

[americanelliott]

You only have a pre-amp gain of 10 , less in practice as your LPF on it's output has an f_c = 10 kHz, so you're attenuating some of the fundamental as well as the harmonics.
Increasing gain to 100 up to 500 would be a good start. Do this with a second amplifier, don't have one high-gain stage. If your coil null output is below 5 mV peak-peak, your ADC should cope with that.

Thank you. We will have to find a creative way to implement a second amplification stage since we can't make any more changes to the PCB layout. Do you think its possible to make this change by soldering a new op amp using a technique similar to one here? https://www.circuitrework.com/guides/6-3.html

[waltr]

Any of those creative methods can be used.
Another way is to build a board with the added circuits that then plugs into the pre-amp socket picking up power, input and output.
There are some two stage pre-amps in this Geotech forum that plug into the original single stage pre-amp socket.

Another method worth looking into deals with your ADC sampling. There are a couple of different ways.

One is sample at a high rate over a 'sampling' period, similar to what an analog PI detector does during target sampling, and simply ADD all the samples together. This increases the effective bit resolution and decreases Noise.

Another is Adding ADC samples over several Pulse cycles. Do the math of how many pulse cycles happen during a pass of the coil over a target to give an idea of number of cycles to add.

[Qiaozhi]

Another method worth looking into deals with your ADC sampling. There are a couple of different ways.

One is sample at a high rate over a 'sampling' period, similar to what an analog PI detector does during target sampling, and simply ADD all the samples together. This increases the effective bit resolution and decreases Noise.

Another is Adding ADC samples over several Pulse cycles. Do the math of how many pulse cycles happen during a pass of the coil over a target to give an idea of number of cycles to add.

Simply taking many samples and averaging does not increase the ADC resolution. The only way to do that is by oversampling and decimation, and the attached application note gives an excellent description of the process.

[waltr]

Simply taking many samples and averaging does not increase the ADC resolution. The only way to do that is by oversampling and decimation, and the attached application note gives an excellent description of the process.

Well I never said "averaging" but your point is right on.
I was trying to give the OP some ideas as this seems to be a student engineering project.

Very good paper on ADC oversampling that is worth studying.

[Qiaozhi]

Very good paper on ADC oversampling that is worth studying.

Yes, I've used that in a battery monitoring application. It works really well.