A Single Coil TEM field metal detector with simultaneous transmit and recieve.

[moodz]

.In the simulations, let's not forget the ohmic resistance of the coil L1 Children of God , the constant current flows through it

whether in simulations or real circuits ... resistance is does not affect current sources ... we just dissipate more power.
A 1.0000 amp current source is a 1.0000 amp current source through 1 ohm or 100 ohms.

[Aziz]

Hi all,

it seems, that Paul has found his magic choke L1 too. It is really magic and a great invention.

We are of course very much interested in it's performance in the field. I am sure, Paul is working on this.

With an IB coil (TX, RX), a two independent channel version can be realised to get maximum of target/ground information.
Aziz

[Riss]

.In the simulations, let's not forget the ohmic resistance of the coil L1 Children of God , the constant current flows through it

[moodz]

I am thinking CSTEM. Current Summing Transient Electromagnetic Mono or CSTEMM if pedantic.

The current summing node is the central feature that allows simultaneous transmit / rx and target signal extraction.

The bipolar signal, the constant current pulses are all prior art features ie there is nothing new here.

In ZVT prior art the coil current is regulated in the transmit coil ( be sensing and feedback loops ) and connected to voltage sources which "short out" target signals.
Here we use a forcing function driven by a current source not a voltage source .. the tx coil has no choice but to follow along.
The current summing node is fed by a high impedance DC current source.
Since the impedance at this node is high a target voltage can be detected when the coil is connected from ground to the node by the switching method.
We dont have to worry about removing the switching pulses and flyback because the switching method ( in this case a bridge ) is between the node and the coil.

The famous "tilt" on the transmit pulses is proportional to the impedance of the current source. Use a better current source and you will get less tilt.
If the impedance of the current source is too low the DC voltage from the current source supply "leaks" through the current source and creates an LR charging curve with the TX coil during ON time .. this is the tilt.
The impedance of the current source needs to be higher than say 100K for tilt cancellation. ( actually it does not cancel ... it reduces )
Any form of loop control would kill our target responses.
I get less than 16 microamps of tilt across 100 microseconds -- in fact its becoming hard to measure.

[Carl-NC]

No, it's a mono-loop version of CCPI, which normally require IB or, minimally, a separate RX coil.

[Altra]

Probably yes. The output of this new scheme* hits a hard limit during the transition due to the diode.

*Paul, you need a name for this scheme. Mono-Reflective? Psueudo-Mono? Mono a mono?

Okay, I was thinking this was the PI version of the GoFind. Mono-loop IB...

[Carl-NC]

Probably yes. The output of this new scheme* hits a hard limit during the transition due to the diode.

*Paul, you need a name for this scheme. Mono-Reflective? Psueudo-Mono? Mono a mono?

[Altra]

Thanks, would we need an IB coil to measure the transition?

[Carl-NC]

During the flat-top, there is no X, only R. Just like a traditional PI. However, you could try to demodulate the transition signal which does include X and R.

[Altra]

Can we demodulate X and R components with this circuit?

[moodz]

I did some sims on this circuit and it looks really good. I set the voltage (2.6V) to produce a +/- 1A drive current:

The preamp ground is (by necessity) referenced to VB=2.6V. Offhand, I see no reason why the circuit cannot be rearranged to place the current source & output on the bottom so that everything is referenced to ground.

Paul gets another A+... clever, practical, simple, and efficient. The nice thing is, I think I can easily modify the CCPI driver in AMX to use this.

Thanks Carl for this analysis. You got almost the same experimental damping values spot on ( I used 10uF and 140 ohms ).

[Aziz]

Yep, this works exactly the same, but now the output is ground-referenced:

Yep, this is also very nice. A+ for you too.

[Carl-NC]

Offhand, I see no reason why the circuit cannot be rearranged to place the current source & output on the bottom so that everything is referenced to ground.

Yep, this works exactly the same, but now the output is ground-referenced:

[Carl-NC]

I did some sims on this circuit and it looks really good. I set the voltage (2.6V) to produce a +/- 1A drive current:



Coil is 300uH with 200pF and 0.5Ω. Transition is the expected half-cosine:



Here is the flat-top region of the current pulse:



Looks like 0.5mA/A for a 100us pulse. I suspect this entirely controlled by L1, which is 10mH in my case. Much better than the uncorrected Minelab transmitter.

Experimentally I found that R1=150 provided good damping and C1=10u minimized droop:



I then fed this to a modest G=10 preamp:



Here is the preamp with a 1us target imposed:



Here is my schematic:



The preamp ground is (by necessity) referenced to VB=2.6V. Offhand, I see no reason why the circuit cannot be rearranged to place the current source & output on the bottom so that everything is referenced to ground.

Paul gets another A+... clever, practical, simple, and efficient. The nice thing is, I think I can easily modify the CCPI driver in AMX to use this.

[moodz]

Moodz, skimming through the white paper I noticed some inventor name errors in the reference section. Here are four corrections, I didn't check the others (5)Douglas L. Johnson (7)David Emery (9)David E Johnson, (10)David E Johnson. Not nit picking but if this becomes legal it's good to be accurate.

You have a good eye for detail Altra ! I cut the generated list from an AI summary ... my bad and apologies to all those guys.

Here is an updated document ... whitepaper_v1.pdf