Designing and marketing new metal detectors.

[ivconic]

It's been a while...
Listing through numerous articles i founded something that correlates with what we talked about here:


Non-magnetic stainless steel:
High quality 300 series stainless steels (e.g. Type 304, 316, etc.) are the most difficult
metals to detect due to their poor electrical conductive qualities and, by definition, have
low magnetic permeability. These are commonly used metals in the food processing
and pharmaceutical industries.
When inspecting non-conductive products, a stainless steel test sphere typically needs
to be 50% larger than a ferrous sphere to produce the same size signal. When
inspecting conductive products, a stainless steel test sphere needs to be 200% – 300%
larger than a ferrous sphere to produce the same size signal.
When detection specifications include non-ferrous and/or stainless steel, the particular
metals, and sizes should be identified. Correct identification of what particles should
be detected is critical because these metals have many varieties and they all look slightly
different to the metal detector.

[Davor]

I actually gave this a thorough think-over. Most of the nowadays detectors are fixated on ground balancing techniques, deep penetration into the ground and whatnot. But the whole point is that small low conductive stuff is very ground-like, hence the only property that makes something needle-like pop up against the ground is it's geometry. I'd say the differential concepts in all-metal mode could work well.

[WM6]

I'd say the differential concepts in all-metal mode could work well.

35 years old, now forgotten technology, can detect not only stainless needle, but fresh grass too.

[Tinkerer]

It's been a while...
Listing through numerous articles i founded something that correlates with what we talked about here:


Non-magnetic stainless steel:
High quality 300 series stainless steels (e.g. Type 304, 316, etc.) are the most difficult
metals to detect due to their poor electrical conductive qualities and, by definition, have
low magnetic permeability. These are commonly used metals in the food processing
and pharmaceutical industries.
When inspecting non-conductive products, a stainless steel test sphere typically needs
to be 50% larger than a ferrous sphere to produce the same size signal. When
inspecting conductive products, a stainless steel test sphere needs to be 200% – 300%
larger than a ferrous sphere to produce the same size signal.
When detection specifications include non-ferrous and/or stainless steel, the particular
metals, and sizes should be identified. Correct identification of what particles should
be detected is critical because these metals have many varieties and they all look slightly
different to the metal detector.

Note that they use spheres as test targets. This eliminates the variables of target orientation or field vector orientation.

Using spheres, each metal alloy shows it's TC due to conductivity. However, complex targets, the ones with conductivity and permeability make differentiation a bit more difficult.

Tinkerer

[Davor]

I think it is not that much of importance because your targets are not neatly aligned for you to detect, but are scattered randomly at random depths. You could speak of some order in chaos in case your targets ARE neatly placed on the ground surface, and have similar properties, e.g. needles or coins. In such cases you just removed some variables. Because the real targets are randomly scattered and at various depths, in real life metal detecting you get nowhere if your detector is not capable of dealing with several orders of magnitude responses, and such big dynamic range will take care of much smaller variations due to shape and orientation for most of the situations.
I'd say the most significant effect in detection that may be attributed to shape or orientation is discrimination precision, but not so much detection of presence.

[ivconic]

Note that they use spheres as test targets. This eliminates the variables of target orientation or field vector orientation.

Using spheres, each metal alloy shows it's TC due to conductivity. However, complex targets, the ones with conductivity and permeability make differentiation a bit more difficult.

Tinkerer

Yes, sphere is perfect reference target for series of testing.
And it is illustrative what they founded out, relating to differences between steel and iron signals!
I guess that does explain why we had troubles with needle detection, before.

[Tinkerer]

Yes, sphere is perfect reference target for series of testing.
And it is illustrative what they founded out, relating to differences between steel and iron signals!
I guess that does explain why we had troubles with needle detection, before.

When talking about target response, we need to add the information of the detector technology used.

For example, with a traditional PI, we get the reactive response due to permeability and the resistive response due to conductivity added since both types of response have the same phase.

With VLF it is different.

With the TEM method, the responses are of different phase, but I use the time domain for the extraction of the target signal so I must look at different times to find the desired signal response.

With the TINKERERS TEM method, purely permeable target like ferrite, that has no conductivity, is 180 degrees apart from a pure conductive target, like gold.

Iron has conductivity and permeability.

If I could get hold of spheres of similar size, but different alloys, there would be notable differences in the response.

For example, there is an alloy, called Mangalloy. It is about 85% iron and 14% manganese. It is extremely hard and used for rock drilling and such. In spite of the high iron content, it is not magnetic. As a tramp metal, it can not be extracted with even a very powerful electro-magnet.

As you would expect, it's response is nearer the response of gold than the response of iron.

If we know which target we want to detect, we can design the detector for this kind of target. If we want a detector that is good for any and every kind of target, it will never be the best detector for a single type of target.

But, now that we can do some considerable number crunching with very tiny MCU's, we can design a detector that can take into account many, many variables and extract signals that we did not even know they existed 20 years ago.

Tinkerer