THE PERFECT GROUND BALANCE

[Midas]

Is that true? I haven't tested an autotransformer coil for Earth field, but I would expect it to need bipolar pulsing as well.

I'd also be very surprised. The differential coil isn't that fundamentally different from the center tap RX design in Tinkerer's design. The only difference is the increased coupling between each half of the coil. Its still a basically a coil of wire and passing it through a magnetic field is going to create a negative potential at one end and positive one at the other. ergo no cancelling.

[moodz]

...I told you there was a trick to the timing and sampling .... the coil will generate an EMF in response to moving within a magnetic field and this EMF will be practically constant from pulse to pulse during transmit and recieve unless you can swing the coil across the field at more than 100 m/s and probably needs to be much faster. The diff pi pulses at 2 - 20 khz so you would need to swing mighty fast to cause an EMF delta in around 500 to 50 microseconds window. The earth field potential is like adding a very low frequency spectrum to a high frequency spectrum then using simple diode detector which would of course detect the energy of the whole spectrum.

The diff PI direct demodulator does not detect the low frequency because it is demodulating only the band of higher frequency target signals ( like a radio ) In fact the earth field EMF would have to vary at naer the carrier ( eg 10 Khz ) rate to fall into the target bandwidth. The target energy being demodulated is synchronously locked to the TX ( carrier ) rate.

You are right though ... the diff coil by itself is no different to in its response to motion through a magnetic field.

moodz

[Carl-NC]

...I told you there was a trick to the timing and sampling .... the coil will generate an EMF in response to moving within a magnetic field and this EMF will be practically constant from pulse to pulse during transmit and recieve unless you can swing the coil across the field at more than 100 m/s and probably needs to be much faster. The diff pi pulses at 2 - 20 khz so you would need to swing mighty fast to cause an EMF delta in around 500 to 50 microseconds window. The earth field potential is like adding a very low frequency spectrum to a high frequency spectrum then using simple diode detector which would of course detect the energy of the whole spectrum.

The diff PI direct demodulator does not detect the low frequency because it is demodulating only the band of higher frequency target signals ( like a radio ) In fact the earth field EMF would have to vary at naer the carrier ( eg 10 Khz ) rate to fall into the target bandwidth. The target energy being demodulated is synchronously locked to the TX ( carrier ) rate.

You are right though ... the diff coil by itself is no different to in its response to motion through a magnetic field.

moodz

Has this been covered elsewhere? Cause I'm not getting it. Earth field response can easily fall in the same realm as a target response, especially at the ends of the swing where the coil reverses direction. I don't see how merely sampling faster makes any difference as the target response will also be practically constant from pulse to pulse.

[Davor]

Somehow I'm in favour of Carl's position here. There is nothing really that prevents slowly changing field from entering a Rx and playing with a ground reference. With unipolar pulses it seem plausible.
However, monocoil and a single battery supply is worth a challenge.

[moodz]

...all I can say is

Has this been covered elsewhere? Cause I'm not getting it. Earth field response can easily fall in the same realm as a target response, especially at the ends of the swing where the coil reverses direction. I don't see how merely sampling faster makes any difference as the target response will also be practically constant from pulse to pulse.

I think this has been discussed in the past ....somewhere ....all I can say is that the method ...as it could have some commercial value... relies on this fact ....the earth field seems to provide a target signal ...but this is a "target" which does not absorb energy .... the energy is being supplied by you swinging the coil. ....once you start treating it as being outside the system ( ie energy input = energy output - losses ) then the problem is somewhat simpler.

[Midas]

I think this has been discussed in the past ....somewhere ....all I can say is that the method ...as it could have some commercial value... relies on this fact ....the earth field seems to provide a target signal ...but this is a "target" which does not absorb energy .... the energy is being supplied by you swinging the coil. ....once you start treating it as being outside the system ( ie energy input = energy output - losses ) then the problem is somewhat simpler.

Are you sure your method is actually as good as bi-polar pulsing ? In principal just taking a comparison sample after a long wait, after any true target signal has dissipated is probably the simplest solution. I guess the disadvantage is potentially a limit on how fast you can pulse and\or unintended cancelling of long TC signals. I've got no idea how significant those disadvantages are in practice..

[moodz]

Are you sure your method is actually as good as bi-polar pulsing ? In principal just taking a comparison sample after a long wait, after any true target signal has dissipated is probably the simplest solution. I guess the disadvantage is potentially a limit on how fast you can pulse and\or unintended cancelling of long TC signals. I've got no idea how significant those disadvantages are in practice..

Actually I wont be able to get a quantitative evaluation till I finalise the 24 bit ADCs and samplers on the UNIPI.. with at least 6 x 24 bit channels to play with I can resume the front end trials ... however Aziz and others, including myself have posted bipolar frontends ( some evolution has occurred in exact circuits since then )..

http://www.geotech1.com/forums/showp...&postcount=175

moodz

[Aziz]

Hi all,

another good but very power efficient bipolar pulse is the triangular AC current pulse.

The LC series resonant tank can be modified by setting the frequency and the tank capacitor such, that the coil current is a more or less triangular current. Well, the capacitor C is just blocking the DC current.

A simple H-bridge driver does the job quite well (50% pulse duty). The current pulse is a wide band pulse (odd harmonics) and can be operated with single power supply.

Due to it's high efficiency you can easily power the tank circuit with high voltage to overcome the high impedance of the LC tank.

This is very old but works quite well.

Cheers,
Aziz

[Aziz]

BTW,

the topic "the perfect ground balace" is somewhat misleading. You all won't get it unless you change the processing means into the digital DSP world (doing software math calculations).

With analog electronics calculation means, you will get only a simple approximation (way off the perfect solution). Your simple analog math is often limitted to:
- simple addition / subtraction of a literal (constant)
- simple addition / subtraction of a variables (demodulated response signals)
- simple division by a literal (simple voltage divider)
- simple integration (or differential integration, which is phase sensitive)

Beyond this, you need more complex circuits to implement a particular math calculation (division, multiplier, log, exp, min, max, abs, sqrt, ..) .

When will you ever change into the real DSP world?

Cheers,
Aziz

[moodz]

BTW,

the topic "the perfect ground balace" is somewhat misleading. You all won't get it unless you change the processing means into the digital DSP world (doing software math calculations).

With analog electronics calculation means, you will get only a simple approximation (way off the perfect solution). Your simple analog math is often limitted to:
- simple addition / subtraction of a literal (constant)
- simple addition / subtraction of a variables (demodulated response signals)
- simple division by a literal (simple voltage divider)
- simple integration (or differential integration, which is phase sensitive)

Beyond this, you need more complex circuits to implement a particular math calculation (division, multiplier, log, exp, min, max, abs, sqrt, ..) .

When will you ever change into the real DSP world?

Cheers,
Aziz

If someone buys me the parts I will do it now

[Aziz]

If someone buys me the parts I will do it now

You don't need expensive parts for it.
Simple PI controller, sound card, PC/Laptop/Netbook/Tablet.
There you have it the DSP world.

Aziz

[moodz]

You don't need expensive parts for it.
Simple PI controller, sound card, PC/Laptop/Netbook/Tablet.
There you have it the DSP world.

Aziz

That would work fine except for the crappy and variable quality and expensive sound cards ... after I finish UNIPI I will make a precision USB to 24 bit converter box for PC with many input and output channels and synchronisation control for precise timing.

moodz

[Tinkerer]

Does the IB compensate for Earth field?

The bi-polar TX makes sense.

Is there another way?

Let's look at it from a different angle, do it on the coil.

Induction balance.

With the IB coil we have on the outside the TX coil and on the inside the RX and BU coil.
I (+TX) + I(-BU) = I (RX) 0.001

If the H TX = H RX, then TX and RX see the same I, generated by the Earth's field.

With the BU coil, the RX now sees I Earth's field /1000

Now, if somebody would be so kind and write this into a neat Mathematical formula?

Tinkerer

[Carl-NC]

Let's look at it from a different angle, do it on the coil.

A coaxial coil (RX+/TX/RX-) would do it. Carl V. Nelson has a couple of interesting coil patents that might also be effective.

[Davor]

another good but very power efficient bipolar pulse is the triangular AC current pulse.

Which is not unlike step voltage after all. No recovery time, and no delay, hence less juice required for detection.

I am toying with a matchless detector by rev Scarborough. OK, it is a toy, but it has a potential. All metal readings are related to mutual inductance between search coil and a target, and in case of "beat balance" it is merely acting as a free running PWM. Perhaps the PWM principle will yield a better PI after all.

BTW, there is an arrangement that uses a capacitor and analog switches that yields an opamp with over 120dB CMMR. Having PI sample as a windowed pulse, this seem as a perfect candidate, and it is not complicated at all. Any interest?