Why does an inductive proximity sensor only detect certain steel alloys?

[Infraviolet]

I've built an inductive proximity sensor circuit based on a colpitts oscillator with an inductor*. When the inductor is brought close to some metals it has a significant change in the oscillation frequency due to the way the presence of metal alters the inductor's inductance. And yet some metals it just doesn't detect.

*commercial inductor, radial coiled wire design, small and compact, not sure of the internal details or core type as the datasheet doesn't describe them

The main purpose of the design is to detect screws to use them as non-contact "end"-stops and other position and millimetre range distance markers in robotic systems. It detects most screws just fine, including down to small M2 screws at many millimetres range. And yet others it completely fails to detect, it seems some metals have no ability to change the inductor's inductance, however close and however large a sample. Aluminium and brass seem to have NO effect at all, not less effect than steel as is usually suggested in literature, no effect. Even chunks of either sized at several cm in all dimensions brought near the inductor don't change the frequency, yet tiny pieces of some steels do. Doesn't detect a reel of Tin-Lead solder either. Yet more strange is that some steels aren't detected either.

I don't know the proper alloy description, but some screws I have are made of a steel with what might be considered an almost "red" hint of colour to it, consider the ideas of "cool" blue-ish grey versus "warm" red/yellow-ish grey, these screws are a "warm" metal colour. Also I find the inductor detects the steel found in a bunch of screw drivers and spanners I had to hand very well, particularly tools with a black coating, but it doesn't detect some steel "dowel pins" I presented to it, these dowel pins are very shiny and polished indeed. As far as thin (2mm, 3mm) steel bar stock goes, it seems to more favourably detect "rough looking" types with a somewhat flecked surface finish, and only detects some samples amongst shinier more polished steel alloy types.

None of the metal items I presented it with are from sources which stated exactly which alloy wa used, except perhaps some of the screws, and not the ever-so-slightly "red" ones either. So I can only describe them by surface appearance.

I'm not so much trying to debug it, as the screws I most crucially need to detect are all made of steel alloys which it does detect, but I'd be very interested to understand what difference in the metal compositions causes some to not affect the inductor at all.
Thanks

[TimFox]

Note that the "stainless steels", especially the 300-series "austenitic" alloys are nominally non-magnetic.
To get a reliable indication with an inductive proximity sensor, you need a metal alloy with a high-ish permeability, such as cold-rolled steel, to shift the inductance of your coil far enough to get a frequency shift above the threshold.

[Kleinstein]

Not all steel is ferromagnetic.
Usually a ferromagnetic metal will shift the resonance frequency down, often quite a bit. A non magnetic material with usually shift the resonant frequency up, as the eddy currents replell a magentic field.
So it depends on how the sensor is build. Some metal detectors use this to tell apart iron and othe rmetal.

[Infraviolet]

Thanks, I hadn't realised that steels' magnetic properties could be so varied. I'll have a go at cross comparing whether the examples my inductive proximity sensor can't detect are also those which aren't so strongly attracted to magnets.

The eddy currents note makes sense, an earlier version of the sensor did detect Aluminium with a faster frequency output, may have been as that version was sprawling on breadboard and this one is compact on a PCB so the eddy currents had more extended lengths of wire to be induced by and reinduced in to. And indeed, though I forgot to mention it, all the alloys the sensor can detect result in a lower frequency of oscillation.

[ConKbot]

The good news, it's not just steel types, even how it is been worked can matter. I have a stainless vacuum mug, and it's mostly non magnetic. But where it necks down, and near the rim of it, a rare earth magnet sticks. These areas are work hardened more than others in the forming, and the work hardening makes it become more ferromagnetic than the fully soft base metal.

[ejeffrey]

it seems some metals have no ability to change the inductor's inductance, however close and however large a sample. Aluminium and brass seem to have NO effect at all, not less effect than steel as is usually suggested in literature, no effect.

There is an effect but it might be _much_ smaller.
It depends on your nominal operating frequency and how sensitive your measurement is. 

Also a simple oscillator circuit like yours mostly only detects the change in inductance.  You can get more information if you use a circuit that can measure loss (resistance) as well.

Another way to measure is to measure mutual inductance.  If you have two coils next to each other and excite one, the pickup in the other will be screened by a nearby metal.

[Infraviolet]

Regarding mutual inductance, would that only work if you have the transmitting and recieving inductors on opposite sides of the metal sample, or does it work in the kinds of geometrical configurations which would be considered "reflection". Might give me an idea for a future project.
Thanks

[Nassau Cable]

Hello. Your inductive proximity sensor circuit, which uses a Colpitts oscillator with an inductor, can detect some metals but not others. This is because different metals have different properties that affect how they interact with the inductor in your circuit. Ferromagnetic metals like iron, cobalt, and nickel can significantly change the inductance of the inductor, resulting in a noticeable change in the oscillator frequency, while non-ferromagnetic metals like aluminum, copper, and brass have a much lower effect on the inductor. However, even non-ferromagnetic metals can still cause some change in inductance depending on their shape, size, and proximity to the inductor.

It is important to note that not all steel alloys are ferromagnetic, and their magnetic properties can depend on their composition and processing history. Some types of steel can be heat-treated to change their magnetic properties. Additionally, the surface finish and texture of the metal can also affect the sensitivity of the inductor. Rougher surfaces can provide better contact and coupling with the inductor, while smooth and polished surfaces can reduce the area of interaction and coupling.
To improve the sensitivity and reliability of your circuit for detecting a wider range of metals, you may need to further characterize and optimize your circuit.

[armandine2]

Principle of Operation

snip from IFM catalogue