field test unit no 001 "model T"

[Qiaozhi]

Thanks .... the detector is "very" sensitive to moving magnetic fields ... ie if the coil is moving it will detect the EMF ... if the coil is stationary it will detect a moving speaker magnet ( ferrite ... not particulary strong ) at just on 12 feet ... which I was quite impressed with ... because that would be a quite a small signal at the coil.

Even though the detector is responding to mag fields it is relatively easy to pick target responses ... the big plus is auto ground balance ... so what you gain on the swings you lose on the merrygoround LOL ... IMHO ... an earth field problem is much easier to solve than ground balance.

You can solve the earth field effect problem very easily by taking a second [much later] sample and subtracting it from the main sample. If set up correctly, the magnetic field from a large white board magnet can be completely cancelled.

[moodz]

When you have the coil stationary and a moving speaker magnet is sensed when 12 ft from the coil, but when the same magnet is stationary and the coil is moved you find the signal is no different whether the magnet is there or not, it is not that the coil is sensing the moving magnet, it is that the moving magnet is distorting the earth's magnetic field and the coil is within the distortion zone.

Trap for young players, Moodz.

Thanks... yeh I realise that .... the magnet at 12 feet is much weaker than the earths magnetic field however the "movement" of the field causes the detector to react. I will try your pipe experiment after I knock up the next circuit.

[moodz]

You can solve the earth field effect problem very easily by taking a second [much later] sample and subtracting it from the main sample. If set up correctly, the magnetic field from a large white board magnet can be completely cancelled.

Hey thanks Qiaozhi ... I am kinda aware of that but I appreciate your tip.
I made a couple of changes to the front end and the Mag problem is solved now ....

The shots below show the output ( yellow ) of the first stage with a 1 gram natural gold nugget ( ie not a ring or foil ) at 2cm from the coil.
The red trace is the TX OFF gate pulse edge.
I am getting 1 volt of deflection which is not bad IMHO .
This signal goes to the second stage where it is amplified again and DSPed etc etc etc

The coil is a mono, 0.4 ohm, 300uH

My "magic sauce' front end is damping this coil in around 5us ...

waving the magnet does not change this result.

[Aziz]

Wow! That's an impressive result.
It's the end of ... *LOL*
Cheers,
Aziz

[Qiaozhi]

I have just removed a large number (~14) nonsense and personally abusive posts from this thread. Please keep this type of behaviour out of Geotech!
This is a very interesting thread that is being dragged down into the gutter by a few individuals. If it continues then Geotech Admin will be forced to step in and start issuing infractions.

[UrbanFox]

Hey thanks Qiaozhi ... I am kinda aware of that but I appreciate your tip.
I made a couple of changes to the front end and the Mag problem is solved now ....

Moodz, the only changes I can see being made to the front end to reduce static magnetic field effects is to use capacitive coupling. Is this what you have done?

[moodz]

Moodz, the only changes I can see being made to the front end to reduce static magnetic field effects is to use capacitive coupling. Is this what you have done?

There is a far more elegant solution involving DSP. Though I did give the cap method a go in a previous design. I have not tried the late sampling method suggested because it can create holes in the target response profiles.

[UrbanFox]

There is a far more elegant solution involving DSP.

However, your statement was "I made a couple of changes to the front end and the Mag problem is solved now ....".

What advantage do you expect to get from DSP when you state you have already solved the problem?

[moodz]

I said " there is" not "there will be". ..... The changes to the frontend actually just pipe the mag signals to the dsp instead of trying to block them. If you can't solve a problem in one place ... Solve it in another.

[Qiaozhi]

There is a far more elegant solution involving DSP. Though I did give the cap method a go in a previous design. I have not tried the late sampling method suggested because it can create holes in the target response profiles.

The "holes" in the target response profiles will only be a problem if you make the secondary sample pulse too close to the main sample. A much later secondary sample will only affect targets with longer time constants, and I presume you're really focussing on targets with very short time constants (such as gold nuggets).

[dougAEGPF]

The "holes" in the target response profiles will only be a problem if you make the secondary sample pulse too close to the main sample. A much later secondary sample will only affect targets with longer time constants, and I presume you're really focussing on targets with very short time constants (such as gold nuggets).

Some multi ounce Oz gold nuggets can have very long Tc's!!! For Oz some of us want a detector that can handle the worst ground but not sacrifice depth on some bigger, solid nuggets with longer TC's as happens with some other detectors timings on some of these nuggets!
DougAEGPF

[Aziz]

Capacitive coupling: High-pass filter (-20 dB/decade = 1. order hp-filter)
Subtracting late sample: Mathematically equivalent with high-pass filter (-20 dB/decade = 1. order hp-filter)

Who dares to show us the mathematical proof? (UFox?)
(I won't do this )
Aziz

[Qiaozhi]

Some multi ounce Oz gold nuggets can have very long Tc's!!! For Oz some of us want a detector that can handle the worst ground but not sacrifice depth on some bigger, solid nuggets with longer TC's as happens with some other detectors timings on some of these nuggets!
DougAEGPF

The problem then is to find an acceptable compromise. Any technique designed to eliminate the worst ground will inevitably result in sacrificing depth for other targets.

[Qiaozhi]

Capacitive coupling: High-pass filter (-20 dB/decade = 1. order hp-filter)
Subtracting late sample: Mathematically equivalent with high-pass filter (-20 dB/decade = 1. order hp-filter)

Who dares to show us the mathematical proof? (UFox?)
(I won't do this )
Aziz

If the two techniques are mathematically equivalent, then (as I mentioned in my previous post) the ground rejection will require some compromise between gold targets of differing time constants. You cannot change the laws of physics.

Go on, don't be a spoil sport ... give us the proof.

[Tinkerer]

If the two techniques are mathematically equivalent, then (as I mentioned in my previous post) the ground rejection will require some compromise between gold targets of differing time constants. You cannot change the laws of physics.

Go on, don't be s spoil sport ... give us the proof.

A novel solution to the very long TC target decay, is to use the "forced decay curve", a method that accelerates the decay, so that the decay curve reaches 0 in a very much shorter time than it's natural TC.
This allows for much higher pulse repetition rates, without loosing any sensitivity to long target TC nuggets.

Tinkerer