detection depth

[green]

The last distance vs amplitude curve I posted needs some description added to the chart.

calculated curves: each coil, same current, same inductance
changing coil current or different targets shift the curves up or down, slope remains the same

Better wording or other descriptions?

Reply #60 [that stage couples to the next stage, 10k to inverted input of next stage] remembered wrong, 100k not 10k.

[flex]

Hiello Green,

did you get any advantage of integrator 2 instead integrator 1 ? I think integrator 2 will have much bigger problems with temperature ?

[green]

Hiello Green,

did you get any advantage of integrator 2 instead integrator 1 ? I think integrator 2 will have much bigger problems with temperature ?

http://www.geotech1.com/forums/showt...433#post216433

Don't think I have a problem with temperature. I have added a 100 ohm pot(wiper to inverted input, ends to switches)to trim EF cancelling.

[green]

The last distance vs amplitude curve I posted needs some description added to the chart.

calculated curves: each coil, same current, same inductance
changing coil current or different targets shift the curves up or down, slope remains the same

Better wording or other descriptions?

Reply #60 [that stage couples to the next stage, 10k to inverted input of next stage] remembered wrong, 100k not 10k.

Two other distance vs amplitude charts I pasted to paint for reference. Neither chart including the one I posted defines if the curves are for same inductance or same number of turns. Aziz chart looks like it is for same inductance. I think the description for the other chart on how to use it suggests it's for same number of turns. I would like to add what is important to the chart so if someone wants to use it they don't need to read the thread. Could add more coil sizes to chart if anyone thinks it's needed.

[Funfinder]

Nice work but don't make the mistake and compare long wave induction with "pure" electricity-induction!

MDs are no AC-DC adapters!

Personally I guess that all these loop-coils might turn out as pretty stupid design after some much better method has been invented!
Yes, they are "bundling" or concentrating the field-lines somehow depending on how much loops the coil has but what's the big deal?

Pressing huge long-waves into tiny loop coils for sure is not any good idea!
There is not even much depth gain difference in using 1x5cm small ferrite-coils!
Where are all the cars with thick loop antennas on the roof?

At least you need to calculate the extreme AM-amplitudes also within these charts
because a huge amount of the overall energy gets lost through such small antennas.

[green]

Another attempt at the graph for a PI detector.

[green]

Question: is there a formula to calculate the time constant of a thin square target?
Did a test with some 1x1inch thin targets and one 1x1inch thicker target with some materials I had. The thin targets(centered on coil at 2inches) charted a straight line on a linear log chart and had about the same amplitude after 5usec delay, the thicker target(centered on coil at 2inches) is maybe 1/4 the amplitude at 5usec delay and doesn't chart a straight line. Thicker target laying on coil to give a higher signal strength was charted on the second graph also. Used 8inch figure eight IB coil. Charted change in amplifier out. Wasn't expecting targets B, C and D to chart as close as they did.

[Skippy]

I've done some maths on your readings, based on some data I posted summer 2017.

Targets can be modelled as a wire loop, with inductance and resistance. The resulting L/R value is the time-constant. The inductance doesn't appear to vary with wire diameter/width/thickness (not much, anyway), so it's just the resistance changes that have an effect. Doubling the thickness halves the resistance, doubling the time-constant. So TC is proportional to metal thickness. For convenience, if we use IACS conductivity values, the resistance is proportional to 1/conductivity, and TC will be proportional to conductivity.
* TC is proportional to IACS conductivity
* TC is proportional to metal thickness

The ally drinks can sample: I measured conductivity of this metal as 37% IACS. Thickness = 0.100mm. I measured TC = 6.8 usec for a 25mm square. This looks close to your latest graph, which suggests about 6 usec?

The copper-clad board: "One Ounce" copper is apparently 0.0347mm thick. IACS conductivity = 100%. So a calculated TC =6.8 us x (0.0347/0.100) x (100/37) = 6.38 usec.

The lead sheet: 0.016" thick = 0.406mm. Conductivity of lead = 8.4% IACS. So calculated TC = 6.8 us x (0.406/0.100) x (8.4/37) = 6.26 usec.

These are pretty close, as you observed, 94% and 92% of the ally can sample.

Measurement errors etc can easily swing these figures out. I bet copper-clad thickness is quite variable. And it's not a smooth sheet, it will follow the surface of the grp laminate. And drinks cans may have different alloys, with slightly different conductivities.

As for the large target, it's accepted that the bulk of the material causes more complex current decays (for PI machines especially). There's another thread on the forum about target modelling, I recall seeing multistage L R L R circuits as Spice models, but I can't remember where.

[green]

I've done some maths on your readings, based on some data I posted summer 2017.

Targets can be modelled as a wire loop, with inductance and resistance. The resulting L/R value is the time-constant. The inductance doesn't appear to vary with wire diameter/width/thickness (not much, anyway), so it's just the resistance changes that have an effect. Doubling the thickness halves the resistance, doubling the time-constant. So TC is proportional to metal thickness. For convenience, if we use IACS conductivity values, the resistance is proportional to 1/conductivity, and TC will be proportional to conductivity.
* TC is proportional to IACS conductivity
* TC is proportional to metal thickness

The ally drinks can sample: I measured conductivity of this metal as 37% IACS. Thickness = 0.100mm. I measured TC = 6.8 usec for a 25mm square. This looks close to your latest graph, which suggests about 6 usec?

The copper-clad board: "One Ounce" copper is apparently 0.0347mm thick. IACS conductivity = 100%. So a calculated TC =6.8 us x (0.0347/0.100) x (100/37) = 6.38 usec.

The lead sheet: 0.016" thick = 0.406mm. Conductivity of lead = 8.4% IACS. So calculated TC = 6.8 us x (0.406/0.100) x (8.4/37) = 6.26 usec.

These are pretty close, as you observed, 94% and 92% of the ally can sample.

Measurement errors etc can easily swing these figures out. I bet copper-clad thickness is quite variable. And it's not a smooth sheet, it will follow the surface of the grp laminate. And drinks cans may have different alloys, with slightly different conductivities.

As for the large target, it's accepted that the bulk of the material causes more complex current decays (for PI machines especially). There's another thread on the forum about target modelling, I recall seeing multistage L R L R circuits as Spice models, but I can't remember where.

Thanks. A couple things I've been thinking about lately. What I would need to change to detect a quarter at 24inches(reply #53) and what it would take to detect a 70 lb silver bar at 15feet http://www.geotech1.com/forums/showt...073#post238073 something Eric did 35 years ago. Don't have a 70 lb silver bar. How to estimate the time constant for the bar,

the reason for reply #67. A couple formulas based on your calculations: TC=45usec/inches^2*thickness(inches)*%IACS*width(inches) and TC=.07usec/mm^2*thick(mm)*%IACS*width(mm), hope to be with in 10 or 20% of actual TC. Been looking for some possible targets to purchase to check formula with longer TC targets. https://www.jmbullion.com/copper/copper-bars/ maybe 5oz cube, 1 lb cube and or the 1/2 lb bar. The 5oz cube calculates TC=4500usec, 70 lb silver bar should be a lot higher. Any thoughts on other possible targets? I read Tx time should be 3 to 5 times the target TC for a good Rx signal. Just trying to understand the variables and what it would take to sense the targets.

Time constant for the 1x1x.25inch aluminum bar reply #67(45*.25*60*1=675usec), higher than the 450usec I listed on the chart. The curve looks like it hadn't straightened out so maybe 675usec isn't far off. Used 160usec Tx 1000pps, probably need a longer Tx time and lower Tx rate to chart decay.

[Skippy]

Just for curiosity, I tried out the experimental formula:
TC = 0.073 x L x T x %IACS
.. on your earlier copper pipe samples of 40x40x0.8 & 10x10x0.8, which you measured as TC = 165 and 40 usec respectively.

Using 100% for the copper IACS, the calculated values of TC = 237 and 59 usec resp. are obtained, 44% longer than measured. This is a bit disappointing, I was hoping the copper would be repeatable, it would be an obvious 'reference' material to make samples from, as it's universal.

Regarding the aluminium bar: I see you've calculated used an IACS figure of 60%. This is the pure Al figure, is your sample pure? I would guess it's some structural grade alloy, which typically have IACS figures of 30% to 50%. These lower % values would give a better match to your measurements. Did it seem 'soft' when you cut it?

Regarding the large targets, those copper ingots are quite cool, most other large metal samples I can think of are round bar stock for lathe turning. But estimating/simulating a 70 lb silver bar ... I'm unsure how.

When I stated 37% IACS conductivity for the drinks can metal, I was going from memory. I found the actual data and the forum post it's buried in, in Eric's Vallon thread. 38.2% was the figure I came up with:

"I've done my can-metal resistivity test. I carefully spiral cut a 3mm wide strip from the middle of a can. The total length of my sample was 1.85 metres, and area is 0.105mm x 3mm. I did a good 4-wire resistance measurement, and got R = 0.265 Ohms.
Putting all that into R = p.L/A gives
resistivity = 45.1 x 10e-9.
Comparing this to a 100% copper value of 1.724 x 10e-8 gives a conductivity figure of:
Conductivity =38.2% IACS.
This compares very well with the figure for 3004-series Al alloy in the table I linked to earler of 39 - 43%. 3004-series is a common can alloy, apparently."

[green]

Just for curiosity, I tried out the experimental formula:
TC = 0.073 x L x T x %IACS
.. on your earlier copper pipe samples of 40x40x0.8 & 10x10x0.8, which you measured as TC = 165 and 40 usec respectively.

Using 100% for the copper IACS, the calculated values of TC = 237 and 59 usec resp. are obtained, 44% longer than measured. This is a bit disappointing, I was hoping the copper would be repeatable, it would be an obvious 'reference' material to make samples from, as it's universal.

Regarding the aluminium bar: I see you've calculated used an IACS figure of 60%. This is the pure Al figure, is your sample pure? I would guess it's some structural grade alloy, which typically have IACS figures of 30% to 50%. These lower % values would give a better match to your measurements. Did it seem 'soft' when you cut it?

Regarding the large targets, those copper ingots are quite cool, most other large metal samples I can think of are round bar stock for lathe turning. But estimating/simulating a 70 lb silver bar ... I'm unsure how.

When I stated 37% IACS conductivity for the drinks can metal, I was going from memory. I found the actual data and the forum post it's buried in, in Eric's Vallon thread. 38.2% was the figure I came up with:

"I've done my can-metal resistivity test. I carefully spiral cut a 3mm wide strip from the middle of a can. The total length of my sample was 1.85 metres, and area is 0.105mm x 3mm. I did a good 4-wire resistance measurement, and got R = 0.265 Ohms.
Putting all that into R = p.L/A gives
resistivity = 45.1 x 10e-9.
Comparing this to a 100% copper value of 1.724 x 10e-8 gives a conductivity figure of:
Conductivity =38.2% IACS.
This compares very well with the figure for 3004-series Al alloy in the table I linked to earler of 39 - 43%. 3004-series is a common can alloy, apparently."

Good observation on the copper tubing. I'm planning on some more tests with copper. Not knowing the %IACS for the other materials has confused me. Try different Tx on times and shape with two different coils, (1000pps, 6250A/sec, 160usec on) (10pps, 62.5A/sec, 16000usec on) (10pps, 1A constant current, 16000usec on). Figure 8 Rx coils, 1.5inch and 8inch diameter to try.

The 1x1x.25inch square was cut from a 4ft bar purchased at Lowes, can't find any specs. Don't think it's high grade.

Probably should repeat the above tests with the 1x1inch aluminum foil, aluminum can side, lead square and aluminum square, trying to do better measurements.

Anything else I should be trying or paying attention too when doing the tests?


If anyone knows of a formula to calculate the time constant for different shape targets to compare with the test results I would appreciate it. I would guess the time constant for the 70 lb silver bar could be around .1second.

[Ferric Toes]

Here is a photo of a full size replica of a Spanish silver bar, except that it is made of aluminium. It has a conductivity of 25.8% IACS which is well down compared to pure aluminium at 61%. It weighs 12 lbs compared to the 70 lbs if it were silver. We were not allowed to borrow a genuine bar as they were worth $10,000 each in 1982 when the original tests were done.

I did measure the TC of this bar but cannot remember what it was. I will try and find my old notebook. A real silver bar I reckoned would be several milliseconds. I will also try and find details of the coil which was probably 500uH and we could pulse 50A peak with the 60V supply. We ran the first coil rather too long on the dockside and the plastic housing started to melt. In the sea on the sled it was fine as the sea acted as a infinite heatsink.

Eric.

[Skippy]

I would model the 70 lb (32kg) silver bar as being 17cm x 17cm x 10cm in size. This gives Vol = 0.003 m³ and mass = 31.5kg. Estimate %IACS conductivity at 90% (it's going to be corroded/pitted, may not be that pure etc).
Then estimate TC = 0.073 x 170 x 100 x 90 = 110000 usec = 110 msec., matching your estimate nicely.
The real figure could be 50% longer, due to bulk effects, and the shape being 0.1m x 0.3m rather than square.

"Anything else I should be trying, or paying attention to when doing the tests?"

Make good accurate measurements of your test targets specs, such as thickness. Measure the thickness of 20 layers of foil (or 16 etc) to get a good result. Keep all your targets so you can go back and re-test them.

I'm sorting out a load of 'kitchen foil alloy' aluminium sheet for some testing at the moment, it's common in thickness from 0.03 to 0.10 mm.

[green]

Finally purchased a digital micrometer, been using a 6in vernier caliper so the thickness measurements should be better. Modified the control circuit for 160, 1600 and 16000usec Tx on time and started taking data. Repeated some testing with the 8inch figure 8 Rx coil to compare with a 1.5inch figure 8 Rx coil(160usec Tx). Both charted similar curves with short TC targets(very close to the same as the 8inch figure 8 Rx coil tested awhile back). Then tested the 10x10, 20x20and 40x40mm pieces cut from copper tubing with the 1.5inch figure 8 Rx coil. Got 50usec TC vs 40usec TC(tested awhile back)with the 8inch figure 8 Rx. The temperature has been a lot colder in my workshop,garage(around freezing)lately. Tested the 20x20mm piece cut from copper tubing at 32degF and around 100degF(used a heat gun to increase target temperature and wrapped it to keep it warm until ready to test, used an infrared thermometer to read temperature when recording data). Four tests(32, 100, 32 and 100degF)to see if things repeated. The same temperature runs repeated but the different temperature runs charted a different slope. Copper changes resistance about 20%/100degF. Maybe that has something to do with the difference. The decay slope for the 20x20mm copper squares didn't straighten on a linear log chart like I expected so I need some more testing to determine why. Repeat the copper pieces with the 8in and 1.5in figure 8 Rx coil. Any other thoughts?

[green]

I'm sorting out a load of 'kitchen foil alloy' aluminium sheet for some testing at the moment, it's common in thickness from 0.03 to 0.10 mm.

Measured some kitchen foil I have. Two brands regular strength(.0178mm), one brand heavy duty(.0241mm). Wondering what you are getting.

Measured mils(.7 and .95) and converted to mm.