Minelab 8000

[Aziz]

Hi all,

there are obviously some restrictions and limitations in the coil geometry definition. Dilate and contract will not work if the coil has concave and convex structure at the same time.
Best example is the omega coil shown below.


At the bottom part, the connections from the inner to the outher coil are wrong.
This will lead to little inaccuracy in the numeric calculations.

But there is a work-around possible. Define each path individually and build the bundle.

[Aziz]

Anyway,

we could continue with the coil comparisons at this current coil software stage. The coil geometry description language gives us still enough possibilities to work with.
I have maybe implemented 60-70 % of the features till now. And bug fixing and fine tuning could be required too.

So what's next? Proof of the geometric-mean-rule?

Aziz

[Aziz]

Have a look.
A completely irregular coil bundle.
I am testing the coil software.



Nuclear fusion coil should be possible to model.

[Aziz]

Hi all,

I had a good idea today and have implemented it already.

A single coil geometry definition file can contain multiple definitions now. RX coils, TX coils, Fusion coils, Toilet coils, etc...
All realized with goto and label command.

//start
//..
//commands...

goto DoRX // goto DoRX label

//...all command not taken into account

DoRX: // label definition
// Make the RX coil now
//commands..
end

One can select the coil model name (label) in the dialog box of the coil settings. So I can select from one definition file each coil individually.

I was really fed up with lot's of external coil definition files...

[peterr3]

All of this is very interesting, but the fact is that the 8000 with its 18 coil is unusable over a lot of lateritic vrm soil in Western Australia and some Victorian soils as well. Even when run at its lowest sensitivity. Throw in some conductive soil components and the 8000 becomes an audio nightmare! These are soils the the7000 run in locate patch mode will cope with as will the gpx. Also, I am not aware of any 8000 uses finding a deep nugget that the 7000 or gpx could not detect. Most of the outrageous 8000 depth claims are due to hopeless pinpointing as the nugget was in the hole side wall and falls to the bottom the hole when it widened. Those with a 7000 and using big X CC coils will see not reason to acquire an 8000. Also, how does one detect with the 8000 when you can't carry a pick, mobile phone, pin pointer, gps or a PLB on your person?IMO the 8000 is overhyped and overpriced and has been released with the wrong coil for very severe soil types!

[ivconic]

Your remarks seem very logical to me.
Big companies run out of ideas and just recycle the same thing all over again.

[Aziz]

..
Most of the outrageous 8000 depth claims are due to hopeless pinpointing as the nugget was in the hole side wall and falls to the bottom the hole when it widened. Those with a 7000 and using big X CC coils will see not reason to acquire an 8000. Also, how does one detect with the 8000 when you can't carry a pick, mobile phone, pin pointer, gps or a PLB on your person?IMO the 8000 is overhyped and overpriced and has been released with the wrong coil for very severe soil types!

Yep, this seems to be plausible. Very plausible.

GPZ8000 is a good machine.
All the gold prospectors have to dig big and deep holes which makes them busy for a while.
And the GPZ8000 seems to be very very noisy. A lot of true gold signals won't even be digged out.
I you don't gonna get mad by the sound it is making.
After a while, you lose the focus to the real target signals.
GPZ8000 is a good machine.
It keeps big deep gold in the ground.

[Aziz]

Hi all,

since I have upgraded the coil software, I can model rectangular bundle coils with round corners now.

How to get more detection depth?
Let's cheat a bit.

Instead of a round coil, a square coil with same outer dimension can be considered to be "same size" when detecting. At least in point of view of swinging the coil.

The square coil has a bigger geometric "mean diameter" than a round coil.
It's near, intermediate and far field zone distance is greater than a round coil. And hence, the detection depth should be better.

Should we make more square/rectangular coils in the future?
How much gain would bring the super cheat?

Aziz

[Aziz]

The AI says, that the square coil would have appr. 7.9 % more equivalent radius R compared to a round coil with radius R.
(Equivalent inductance criteria).

Isn't that great?

Simple cheat -> more depth!

[Aziz]

Hi all,

and what consequences have the square coils during the swinging?
And how the square coil is rotated to the swinging direction?
0 degree, 45 degree (max diagonal), or the optimal 4-pin-point coverage degree?

More ground coverage?
More pin-pointing regions (at the corners), evenly distributed to the ground coverage depending on the rotating angle of the square coil (up to 4 pin-point regions possible)?

You see, the question is not trivial.

[Aziz]

The AI says, if rotating the square coil by approx. 26.6 degree,
we would have an evently distributed pin-pointing coverage (4 pin-point regions).



This orientation would pickup small targets as well. You won't miss the target anymore, if you swing a large square coil.


This is a rocket science now. Nobody has thinked of it till now.

Aziz

[Aziz]

BTW,

you don't need my coil software anymore.

Just ask the AI for a comparison of target response between
a round mono coil (diameter D) and
a square mono coil (side lenght D)
for a specific 1-turn round target
at different concentric co-axial z positions.

You specify the search coil (L=300 µH, coil bundle thickness 15 mm, coil current 1A, frequency 1 MHz or lower) and the target size (let's say 2 cm diameter, 1 mm wire thickness).
And ask for a z-sweep position from z=0 to z=1m, each 10 cm height step (or finer).

That's all. You will get very reasonable results comparable with my coil software results.

Who's going to make the AI prompt?
I'm sooooooo tired...

[moodz]

Ran your prompt, Aziz — and your 7.9% holds up.

Setup (your spec): round coil dia D vs square coil side D, both wound to
L = 300 uH, 15 mm bundle, 1 A. Target a shorted 1-turn loop, 20 mm dia /
1 mm wire. Concentric on-axis z-sweep, 0 to 1 m. Mono coil, so the target
response goes as B-squared.

To hit the same 300 uH the round needs ~24.6 turns, the square ~22.4 — the
square encloses 27% more area per turn, so it gets to 300 uH with 9% fewer
turns. Net magnetic moment: square +15.8%. Expressed as your "equivalent
radius" (m scales as R^2), that's +7.6% — basically your 7.9%, the gap is
just inductance-formula constants. So the AI and your coil software land in
the same place.

Two things the "simple cheat" framing skips though:

1) +7.9% radius is NOT +7.9% depth. A mono coil's response falls off as
1/z^6, so the +15.8% moment = +34% signal at depth, but only ~+5% actual
detection depth once you take the sixth root. Real, but small — reading the
7.9% as a depth figure overstates it by about half.

2) There's a near-field crossover at z ~ 0.70 D (~17 cm for a 250 mm coil).
Shallower than that, for a small co-axial target the ROUND coil wins — up to
+49% right at the surface — because it stacks all its turns at one radius
and gives a tighter, higher central field. The square only pulls ahead past
the crossover. So it's specifically a deep-target cheat; on shallow small
gold it's a slight penalty.

Bottom line: square buys a few % more depth on deep targets, paid for with
some near-surface punch on small ones. Not a free lunch, but a genuine (if
modest) deep-target edge — and your "equivalent radius" intuition was right.

Plot attached: |dZ| vs depth for both coils, plus the round/square ratio
showing the crossover.

(All for a concentric co-axial target — off-axis / swing geometry is a
separate question, which is where your 26.6-degree pinpoint idea comes in.)

[Aziz]

Hi Paul,

thank you for your AI prompt.
Google AI is making so much mistakes... I should run my coil software..

Now the big suprize is comparing against big deep gold (larger target sizes, lets say 4-10 cm diameter target).
The square coil definitely wins!

This is really nice.

[Aziz]

Hey Minelab,

we need a square shaped toilet seat coils now. And rotated by 27 degree angle.
Is this possible?
My best mate Donald is asking me. So I am asking you.