Tests of big depth metal detectors ( TR and PI systems )

[green]

The cast lead targets were intended to be representative of 'large targets' , not copper/silver coins. Such targets might be lead /bronze vessels, bronze axe-heads, cannon balls, weapons, and doubtless plenty of other stuff. They're not necessarily going to have low freqs / high TC values, but they should give a medium-strong signal.
The coin-cache modelling is a particularly difficult target to mimic, though we have made a good start, I think.

Some technical stuff:
I noticed you stating that a 1 ounce copper coin has a TC of 500 usecs (f_corner = 318 Hz ). This is quite a bit different to what I would expect. Unfortunately, I don't have a good non-dugup example of our British 1 oz copper coin, only many corroded ones. But I do have a US Peace dollar. This is just under 1 oz, and 0.900 silver, so not exactly the same, but... I measured it having F_c = 1080 Hz (TC = 147 usec) [measured at 13kHz] and the 'George Payne' quoted f_c = 800Hz ( TC = 200 usec) [measured at 6.6kHz]. So I'm surprised you're seeing 500 usec for a 1 oz coin. Do you have a US silver Dollar (Morgan/Peace) in your collection? It would be good to see how it measures up on your rig.

I'm going to rummage and see what copper pipes I can find, just to see how they compare to your sample.

If a coin cache and a poured lead target had the same decay, shape and amplitude using a PI detector. Would a VLF detector think they were the same or at least similar?

[Skippy]

That's a tough question, I'm expecting the answer to be 'NO', as I am under the impression that the cache doesn't behave like one lump.

My earlier post:"I've read (somewhere...) that with coin caches, the eddy currents of one coin can pass to neighboring coins, such as ones below, and this is repeated with less intensity each transfer. This is alleged to cause a cache to have a larger eddy-current lag, which would presumably make the cache appear a higher-conductor/lower corner-frequency target to a VLF. I've no idea about PI response"

Caches seem to be harder to find than big lumps. Some of this is no doubt because they don't give as strong a signal as the total mass of coins would suggest. But it may also be the unusual response that fools VLF's. But... I've not found a cache in real life, so I can't speak with any authority at all !

Of course I was planning on using the VLF in non-motion all-metal mode, so the detection is going to be more PI-like, and 'odd responses' that fool a VLF in 'discrimination mode' might not be important.

The kind of odd responses I'm thinking of are "iron wraparound" , where a target indicates so 'high up' the disc scale, it's off the range. I've seen tests on the XP Deus where this off-the-scale reading caused complete silence. Other machines see it as 'Iron'. And it's more common in stronger soil/dirt, I understand. An expert on Dankowski's forum said the Fisher Coinstrike was unusual, in that it had a wide disc scaling such that targets never wrapped around, and it was good in tough dirt, particularly when the targets were 'high up' silver coins.
But the business of iron reading as "silver" is another example of VLF's being fooled.

[green]

[Some technical stuff:
I noticed you stating that a 1 ounce copper coin has a TC of 500 usecs (f_corner = 318 Hz ). This is quite a bit different to what I would expect. Unfortunately, I don't have a good non-dugup example of our British 1 oz copper coin, only many corroded ones. But I do have a US Peace dollar. This is just under 1 oz, and 0.900 silver, so not exactly the same, but... I measured it having F_c = 1080 Hz (TC = 147 usec) [measured at 13kHz] and the 'George Payne' quoted f_c = 800Hz ( TC = 200 usec) [measured at 6.6kHz]. So I'm surprised you're seeing 500 usec for a 1 oz coin. Do you have a US silver Dollar (Morgan/Peace) in your collection? It would be good to see how it measures up on your rig.]

Recorded data for the first graph with coin at two locations. Centered on coil, amplifier out not saturated at 6usec delay. The second, target was moved closer to edge adjusting for maximum signal. Coin position had a small effect on decay curve. The data for the second graph was taken for a thread discussing constant current vs constant rate Tx. Longer Tx(16msec) moved the constant slope decay(TC)closer to Tx off. Maybe used a different coil and amplifier from the first graph. TC a little higher than first graph. Any thought why the TC, PI vs VLF is so much different? VLF Tx frequency looks like it made a difference, 147 vs 200usec. Would you expect it to be closer to 400usec if VLF Tx frequency was lower? With the PI, 160usec Tx pulse I need to look at the decay curve past 500usec before the decay becomes straight. The TC of the borrowed silver dollar and copper coin were close.

[Skippy]

I had pondered this discrepancy between my figures and George's when I made the measurements.
There's a few possibilities:
1) to accurately find a targets corner frequency, you should measure it at a frequency close to that corner freq (+100%/-50% should be OK), so neither George nor myself are anywhere near on that count.
2) there's nothing technically precise about George's figure, he just stated '800Hz' in one of his online articles. There could be a +/-10% error on that, no one knows.
3) while I've gone to some effort to get an accurate reading (using more than one method), it's possible my own measured f_c of 1090 Hz could have a 10% error
4) I've just measured the one circulated Peace Dollar. I've no idea how it compares to other circulated examples, uncirculated examples, or to Morgan dollars.
5) This one is possibly the important one: Skin effect. Silver dollars are quite thick, and it's quite possible that at 13kHz, the currents don't circulate so deep, meaning there's effectively less metal being measured, and a resulting higher f_c. It's still likely that at 6.6kHz there's a small amount of this effect.
I have tried doing direct measurements of targets by driving the coil from a sig-gen, then looking at the results on a 'scope. But it was damn difficult to calibrate, and to correct for all the phase shifts, so the end result was poor. I'm not sure it's worth another try.

So I don't know how low the Peace dollar's f_c is. It may be 700 Hz (227 usec), it may be 1150 Hz (138 usec)

And I don't have any other huge high-conductor coins that are similar. I have a couple of silver 'quarters' and many clad 'quarters', they have f_c approx 1600 Hz ( 100 usecs)(when measured at 13 kHz) though I haven't got precise measurements. The 100 usec figure is not too far from your graph's 130 usec ( = 1225 Hz). I'n not enthusiastic about using clad coins for these tests, I prefer something uniform and homogenous.

[green]

More tests. Copper squares cut from 1/2 inch tubing, about .8mm thick. Was going to test 10, 20 and 40mm squares. TC should be linear with length 10mm(40usec) 40mm(165uec), didn't test 20mm yet. Thinking about 5, 10 and 20mm thick poured lead targets. The form I wrapped the foil around was 64mm instead of 60mm as suggested. Needed a little more lead to get 20mm thick. A 1oz copper coin has a TC close to 500usec, maybe need a 40mm thick lead disk for large copper or silver coins. I still think the poured lead would make a more repeatable target than a number of coins. How the coins are stacked can effect detection distance. Just something to think about.

Charted two 64mm diameter lead disks. One about 5mm thick, another about 18mm thick. Both targets were spaced 3inches from coil when recording. Didn't expect the two targets to be that close. Charted with time scale 200 and 800usec full scale. Need 800usec full scale to get TC of some larger targets. Don't know if time constant is important when target and GB sample are taken at less than 200usec.

[garyq]

Large Copper Test Pieces

My posting may or may not be of some interest for this discussion.

Many years ago I obtained 2 large solid pure copper pieces that were created during a pouring of pure copper at a mining company’s copper smelter.
I have pictured them beside a 1919 penny (97% copper) to give an idea of the dimension and size of the copper pieces.



Some time ago I performed air tests using the 2 large copper pieces to get some idea of their depth potential and also compare different settings between older and later models of Pulse Induction detectors that I’ve had in my possession.

In regards to the actual size of the copper pieces the thicker shorter piece to the left of the coin weights 1165 grams and its dimensions are L 9cm (3.5”) x W 9cm (3.5”) x D 5cm (2”)

The copper piece to the right of the coin weighs 1260 grams and its dimensions are L 21cm (8”) (at its longest point) x W 15cm (6”) (at its widest point) x D 2.5cm (1”) (at its thickest point).

Air Depths using the strongest setting on the PI with an 18” mono coil a clear response on the 1165g piece was achieved at 81cm (32”) and the 1260g piece was 92cm (36”)
With an 18"DD coil it received a response on the 1165g piece at 68cm (27") and the 1260g piece at 86cm (34")

As a comparison when a standard aluminium drink can lying side on was squashed flat and using the same PI setting with the 18” mono a clear response was also achieved at 92cm (36”)

The picture below was during some in the ground tests I performed years ago and had previously placed the 1165g copper piece down inside the deepest hole at the centre of the picture at a depth of 76cm (30") and was able to receive a response with an 18" mono coil.

Actually the coil in the picture is a 12" mono when I was testing it over a gold nugget on the end of the wooden rod below the coil inside the hole.