Hello,
I have the need to detect a piece of stainless steel sleeving on a plastic rod moving past a sensor.
This is proving difficult as hall effect proximity does not work for this and IR can't "see" the difference between the metal and plastic.
I need an inductive proximity sensor. ready made ones only come in industrial formats and they cannot run of low voltage (3.3-5V). Also expensive.
I am now wondering if I can just do this with a PIC or other micro. Many of them can already to capacitive touch by means of CVD.
I think that one should be able to do something similar for inductive sensing.
Has anyone have any ideas if this can be done or maybe even tried it?
Possibilities:
1. A reluctance sensor. A permanent magnet with a coil around it. The field will be disturbed by the passing steel and a voltage induced into the coil.
2. Field conductance. Drive an AC signal to a coil on one side of the passing sleeve and use a coil on the other side to sense the signal, the signal should increase as the steel works like a transformer core. This method is fairly easy to implement with a small MCU and you can make it immune to other magnetic disturbances.
3. Capacitive sensing. You can do this using an MCU with 'touch switch' features or you can use any signal source at a reasonably high frequency and use the sleeve as a dielectric to increase coupling to a plate sensor.
I was thinking of somehow using the capacitive touch principles where we use current source to charge the input and stray capacitance present and measure the charge time with a timer. The charge time then determines the capacitance and from that, touch and proximity.
I am wondering if one can do something similar with a small coil and apply a current and measure the voltage rise or something like that. Personally I have never done.
Very true - sorry.
The metal sleeve will be thin wall (0.6-0.7mm) piece of tubing. Stainless steel 304
length around ~3/4" : 20mm
OD of ~1/2" : 12.7mm.
The sensor distance between 1 and ~1/2" :12mm.
Yes, can use Pic MCU.
Design circuit like Inductance measurement application.
An opam with LC resonance and self oscillator, when metal place close coil, frequency will change. Use MCU to measure frequency will detect proximity.
The sensivity follow how much frequency change. Take care of heat and drift after operating time.
Use this method you can know that metal maybe is steel(iron) or aluminum when frequency increase or reduce.
Or use a coil with pulse like metal detector, if has no metal, the length of pulse reflect with longer time. When metal exist, the pulse time will be shortter.
I did some experiments with OP amp and LC on LtSpice yesterday. I was surprised how difficult it is to get that kind of oscillator to reliably work - even on simulation. I found some university exercise where it was mentioned that students should build this to experience the real world problems and complications of this.
I do not have a lot of experience with this type of thing.
I have no idea what initial inductance a practical suitable coil would have. Whether we are talking nH or uH range.
And of course, I have no clue how much inductance would be expected to change with the metal present.
I will have to do a whole lot of experimentation to find all this out.
What type of plastic rod ?
While a reflective optical sensor may struggle, a transmissive sensor shining right through the rod might be possible with a laser diode source.
Reflective might still work at a different optical wavelength.
One further thought. A hole drilled through the rod if practical, might be of help.
The plastic is some sort of poly line which contains fluid. So no hole :-(.
I have done some reflective tests with IR proxy and that could still be an option but the problem is calibrating it on the fly as well as temperature variation, dirt, wear and distance. The IR reflection values vary heaps depending on exact distance of the object from sensor. This makes it very difficult to detect objects based on reflectiveness.
I have not fully ruled that out yet, depending further research and testing.
The sleeve is thin, and the distance problematic, so all you are left with is a pretty effective shorted turn.
That shorted turn moving in and out of a tuned tank circuit should shift the resonant frequency fairly effectively.
As you have already discovered, there will be serious issues of drift and long term stability to overcome.
Another lesser known technique to explore might be a linear variable differential transformer (LVDT).
These do pretty much exactly what you want, produce a voltage output with axial displacement of a band or slug.
I have no idea if these are commercially available as just a discrete sensor coil, but they may be. https://en.wikipedia.org/wiki/Linear_variable_differential_transformer
It may even be possible to wind your own on a round plastic bobbin with three partitions and hook that up to an Analog Devices AD598.
This is a whole project with complications all on its own.
In my reality, it is only a small building block of a much bigger project and the brief was:
"Just use a sensor to detect the sleeve"
Funny how that often works.
This project is a simple LC meter based on PIC16F682A mcu. Description Here is another piece of laboratory equipment – LC meter. This type of meter, especially L meter is hard to find in cheap commercial multimeters. Schematic of this one came from this web...
I'd think about an eddy current loaded sensor, a coil
should act like a xfmr primary w/ open or shorted
secondary.
Outfits make angular position and proximity sensors
using pairs of coils and a conductive target plate.
Microchip Technology appears to offer either chip or
module products, and eval board is a good way to
prototype such a thing whilst working on ultimate form.
Matha Electronics Pallimuku , Electronics shop in Cochin, Kerala , India , Buy Electronic project kits , Electronic components, Industrial components, electronic spares description content Matha Electronics is a Landmark In Pallimuku , Ravipuram in Ernakulam and known for its perfect pricing and...