A closer look at the AF108 bi-polar drive circuit

[Altra]

No response?

I noticed you have not published full details on your mag?

[ivconic]

No response?

[ivconic]

... here is a video of the scheme I used https://www.youtube.com/watch?v=HKHWsdSx6Bw

Give schematic&pcb HERE!

[Altra]

Tried your timings with Excel. With ground slope -1 and -1.3. With bipolar Tx adjust GEB delay until target sample and GEB sample are the same. With unipolar Tx adjust delay until[int with(GEB)] is minimum. I have been setting GEB delay about 4us and GEB sample about 100us. Adjust target delay 6 to 20us, then adjust target sample to adjust ground balance. Not saying it's a better way. Either delay or sample time could be adjusted to ground balance.

Nice spread sheet Green. There are numerous combination that would work as long as you can get the teeter totter to balance. If you didn't see it already, here is a video of the scheme I used https://www.youtube.com/watch?v=HKHWsdSx6Bw

[green]

I'll post a simple diagram tomorrow. What I have been doing 100us Tx, Sample delay 10us, sample1 10us, GB Adjust(delay) 0 to 50uS, GB sample 30uS. This repeats for the negative cycle. Signal = (Sample1 + (-sample1)) - (GB + (-GB)). The differential integrator takes care of the sign on the negative samples by inverting them. Not sure if this is the best way, probably better schemes. If you change the sample delay the GB needs re-adjusting.

Tried your timings with Excel. With ground slope -1 and -1.3. With bipolar Tx adjust GEB delay until target sample and GEB sample are the same. With unipolar Tx adjust delay until[int with(GEB)] is minimum. I have been setting GEB delay about 4us and GEB sample about 100us. Adjust target delay 6 to 20us, then adjust target sample to adjust ground balance. Not saying it's a better way. Either delay or sample time could be adjusted to ground balance.

[Carl-NC]

Carl is there any type of noise reduction with bipolar?

Yes, subtractive sampling forms a comb filter. But this also happens with the late EFE sample in a monopolar design.

[6666]

I'll post a simple diagram tomorrow. What I have been doing 100us Tx, Sample delay 10us, sample1 10us, GB Adjust(delay) 0 to 50uS, GB sample 30uS. This repeats for the negative cycle. Signal = (Sample1 + (-sample1)) - (GB + (-GB)). The differential integrator takes care of the sign on the negative samples by inverting them. Not sure if this is the best way, probably better schemes. If you change the sample delay the GB needs re-adjusting.

Thank you greatly appreciated.

[Altra]

Is there a simple timing / sample drawing that shows how the ground balance is achieved with bipolar pulsing ? thanks

I'll post a simple diagram tomorrow. What I have been doing 100us Tx, Sample delay 10us, sample1 10us, GB Adjust(delay) 0 to 50uS, GB sample 30uS. This repeats for the negative cycle. Signal = (Sample1 + (-sample1)) - (GB + (-GB)). The differential integrator takes care of the sign on the negative samples by inverting them. Not sure if this is the best way, probably better schemes. If you change the sample delay the GB needs re-adjusting.

[6666]

I noticed you can get equal signals on hi and low conductors when not using ground balance. Normal pi's the late ef sample subtracts from the early sample and targets that are high conductors like silver give less response than gold. A bipolar can operate with a single early sample per phase. PI's by nature ignore light mineralization, so in some locations you can increase the range of different targets at equal depths. Ground balance is easier to accomplish, since the earth field is not in the equation.

Is there a simple timing / sample drawing that shows how the ground balance is achieved with bipolar pulsing ? thanks

[ivconic]

Since there is no late EFE sample, you can run a much higher pulse rate.

That's good!
Problem is in schematic.
I am not sure is it without further mistakes and errors, the one i am redrawing from.
I only have found that CD4081 was renamed to "4069", so far...

[Altra]

Carl is there any type of noise reduction with bipolar?

[Carl-NC]

Apart from what's explained; are there any other benefits from this or not?

Since there is no late EFE sample, you can run a much higher pulse rate.

[Altra]

This answers the question i was about to post.
Apart from what's explained; are there any other benefits from this or not?

I noticed you can get equal signals on hi and low conductors when not using ground balance. Normal pi's the late ef sample subtracts from the early sample and targets that are high conductors like silver give less response than gold. A bipolar can operate with a single early sample per phase. PI's by nature ignore light mineralization, so in some locations you can increase the range of different targets at equal depths. Ground balance is easier to accomplish, since the earth field is not in the equation.

[ivconic]

The '108 is a mine detector. A critical requirement is that it generates ZERO net magnetic field to avoid triggering a magnetic mine. So it pulses a positive field followed by a negative field. A side benefit is the cancelation of EFE.

This answers the question i was about to post.
Apart from what's explained; are there any other benefits from this or not?

[ivconic]

I was lazy to finish it, only PS stage is missing.
Also one "4069" from "original" schematic is actually a CD4081.