Ground Balance Theory

[Aziz]

I hope someday you can actually make a working model of something that works in the REAL world that is not some computer idea. So good luck in making a working detector that is better than anything out here now.

PJ, you probably don't know it. But I have found the holy grail of VRM science.
Candy couldn't. *LOL*

Now let's look at the patent offices, whether someone will need taring & feathering.

Aziz

[PATCHES JUNIOR]

Hi all,

ya wanna see my new magnificiant, excellent, super-duper VRM soil modelling function at work? I have taken Thomas's data and made the best fit VRM model functions.
My fair rule: You get what you give.
Thanks Thomas for providing the measurement data.

See the pictures and the Excel file (all zipped).
[ATTACH]27178[/ATTACH]
F$#K!, the Excel file is too large for upload (~1.07 MB). Admin please! I'm getting mad again.
Where are my blue pills now? Doctor!!!!!

Cheers,
Aziz

PS: To forum and patent trolls:
Eat this now: G(t) = a*( (t+p)^b - 1/(t+p+w) )

I hope someday you can actually make a working model of something that works in the REAL world that is not some computer idea. So good luck in making a working detector that is better than anything out here now.

[Aziz]

Another Try: Now Really Busted!

Hi all,

ya wanna see my new magnificiant, excellent, super-duper VRM soil modelling function at work? I have taken Thomas's data and made the best fit VRM model functions.
My fair rule: You get what you give.
Thanks Thomas for providing the measurement data.

See the pictures and the Excel file (all zipped).
Thomas's-Decay-Curve-V2-pics.zip
F$#K!, the Excel file is too large for upload (~1.07 MB). Admin please! I'm getting mad again.
Where are my blue pills now? Doctor!!!!!

Cheers,
Aziz

PS: To forum and patent trolls:
Eat this now: G(t) = a*( (t+p)^b - 1/(t+p+w) )

[PiTec]

More measuremets

Hi,

I made more measurements to show how the TX pulse widths affect the exponent b. The previous measurements were made with a 50 µs TX on pulse, followed by a 2 µs TX off (flyback) pulse, i.e. quite typical for a PI detector. The flyback pulse is the one that actually induces usable eddy currents in resistive targets and magnetizes the soil. The TX on pulse is counterproductive, so separating it from the flyback pulse reduces or even eliminates the weakening effect. See my post from 24 Jan 2013:

http://www.geotech1.com/forums/showt...732#post164732

Here are 2 sets of measurements with a TX on pulse length of 500 µs. 450 µs of these are constant current, so the flyback pulse is effectively the only pulse that kicks the targets. The first measurements were made with a 2 µs flyback pulse as before. The exponents are closer to -1 now:



With an increased flyback pulse length of 10 µs particles with longer time constants react stronger, and the exponents are getting quite close to -1. This seems to be a timing that is more similar to what Eric uses in his tests.



BTW, why do inline attachments like the two above do no longer show up in the post? They used to do some time ago, now I have click on each attachment to see it. 'Show images' is checked in the general settings.

Thomas

[Aziz]

Ok guys,

we have found a much better VRM response model for non-conducting soils now.
Forget what I have written about.
More on the other forum.
More later here.
This is it:
G(t) = a*( (t+p)^b - 1/(t+p+w) )

Stay tuned..

Cheers,
Aziz

[Aziz]

Please explain why it is screwed. It is you that owe an explanation. Agreed that the exponent changes slightly due to the particle size distribution, but most other effects are down to electronic errors, particularly at early times. Noise at late times is a possible problem but that can be removed by filtering.

Eric.

Hi Eric,

we can exclude electronic errors with simple single TC target responses. In the log/lin plot, it must be a straight line. If not, the non-linearity can be obtained and compensated. In the measurements we did with single TC target responses, the non-linearity in the early timings can be excluded. Noise can be excluded with oversampling too (averaging over many many measurements).

We do not face with exponents of b = -1 in a typical PI configuration. These facts are based on many many different real field test measurements. I have processed a new set of measurement data today and they show the same fact.

I hope Robby_H or Rafferty can explain it. *LOL*

Aziz

[Ferric Toes]

Please explain why it is screwed. It is you that owe an explanation. Agreed that the exponent changes slightly due to the particle size distribution, but most other effects are down to electronic errors, particularly at early times. Noise at late times is a possible problem but that can be removed by filtering.

Eric.

[Aziz]

I think the 1/t law has got totally screwed in a typical PI usage finally.
The PI gurus owes us the explanation, when the 1/t law gets valid (if ever and the required conditions).

Aziz

[Aziz]

Here are new measurements with a delay of only 7 µs (i.e. 50 µs TX on, 2 µs TX off, then 7 µs delay). This is shortest time possible with the current setup. Horizontal range: 5 to 205 µs, with the curves starting at 7µs. Vertical range: 0 mV (lin) resp. 5 mV (log) to 3000 mV (1500 mV for volcanic rock).

Excel file: [ATTACH]27105[/ATTACH]

Notes:

• There is no time offset, i.e. p=0
• The dynamic range is higher (1:100 for the MV soil sample)
• The hematite curve is not exactly straight in the log-log graph. There is a slight kink at about 30 µs – no idea why.
• I had to use aluminium foil instead of the gold ring to simulate the viscosity decay. The resulting curve does not fit exactly, but I only spent a few minutes with adjusting the positions of the ring and the foil, so this could be improved.

Unfortunately, I do not have any Oz samples ...

Thomas

[ATTACH]27104[/ATTACH]

Thanks Thomas for your valuable effort. The exponent b is almost in the same ball park for Oz soils.

See the attachment for the short TX pulses of the above measurements in Oz.
MicksDataV2-Test2-ShortPulse-4pics.zip

Cheers,
Aziz

[PiTec]

New measurements with 7 µs delay

Here are new measurements with a delay of only 7 µs (i.e. 50 µs TX on, 2 µs TX off, then 7 µs delay). This is shortest time possible with the current setup. Horizontal range: 5 to 205 µs, with the curves starting at 7µs. Vertical range: 0 mV (lin) resp. 5 mV (log) to 3000 mV (1500 mV for volcanic rock).

Excel file: decay_curves_03a.rar

Notes:

• There is no time offset, i.e. p=0
• The dynamic range is higher (1:100 for the MV soil sample)
• The hematite curve is not exactly straight in the log-log graph. There is a slight kink at about 30 µs – no idea why.
• I had to use aluminium foil instead of the gold ring to simulate the viscosity decay. The resulting curve does not fit exactly, but I only spent a few minutes with adjusting the positions of the ring and the foil, so this could be improved.

Unfortunately, I do not have any Oz samples ...

Thomas