Induction sensor

ot used one, but i think they detect metal moving past the sensor end at speed

What I really meant to ask... Is there a typical circuit design they use and what is it? There are so many of nearly identical size and specs that I'm taking a guess there is a common circuit they all use.

All the units I have seen are based on an internal oscillator (a few hundred kilohertz for he small ones) and a means to measure the eddy current load on that oscillator, with a decision point to switch the output logic signal, indicating metal in the field at some specific magnitude of eddy current load on the oscillator coil. The two main types are ones that spray field out all around the end of the coil, as well as in front of it, and those that have field shaping cores that keep almost all of the field in front of the coil (concentric poles), so that the unit can be flush mounted in a hole in metal without seeing this metal, or close bedside other units, without interacting with each other. The down side of the shielded form is that it cuts the sensing range for a given diameter unit, about in half.

You may want to get a large surface diameter type. Stainless is not as sensitive as other types of metals. Depending on your application, 2 wire types require a 120/240 AC in the loop of lets say a relay coil.., you can get normally closed/normally open types and the average handling current is around

200ma, that is normally enough for a 120v 10amp type relay coil with a medium mechanical mass.

The 3 wire types require DC voltage 12..24 volts DC on most of them going to the Brown Wire (+), Blue wire(-) and the black wire will be a transistor output.

You can get those in PNP, NPN, N-on/N-off etc.. the handling current on those are around 200 Ma DC>

for the NPN types, it is used as a sink (pulls to common when on). We this type, you can switch a variety of devices that use a varying range of DC voltages, you simply pull it to ground/common.. With the PNP type, it becomes a source device. What this does is generates output from the same (+) as the sensor is connected too. The PNP types are not so popular..

So you have your pick now.

--
"I\'m never wrong, once i thought i was, but was mistaken"
Real Programmers Do things like this.
http://webpages.charter.net/jamie_5

I don't think you can possibly build one of these sensors for what you can but several.

Here is a low priced source:

Despite the small costs of the available units I'd still like to build one myself.

I have some pot cores and magnet wire for the coil, and uPs and relays etc for the logic and output. But...the oscillator and detection circuit is new to me and where I need some direction. I've googled and found some good theory but very little on circuit specifics (except a 'hobby' metal detector). Any help is appreciated.

I don't have a construction article for you, or even a schematic, but I think you might get started by building a Hartley Oscillator and sense the change in its DC supply consumption as the sensed metal quenches its oscillation.

one

etc

new

How about an oscillator of which an induction coil is a component. The frequency changes when metal approaches the coil. An appropriate filter lets through the changed frequency. The level of the signal is detected and if high enough a buzzer sounds. Would that work?

R

Sorry, I meant to say that, but must have assumed it was obvious. The oscillator inductor is the coil that produces a field in front of the sensor.

The ones I have seen have the oscillator actually lose enough gain to stop oscillating, and the supply current takes a step (down, I assume). The logic output just detects the changes in the supply current to the oscillator.

Is this attribute unique to Hartley oscillators? Does it not apply to other types as well (Pierce, Colpitts, etc)?

I think it applies to any LC oscillator. I picked the Hartley, because if you are going to wind your own inductor, it puts more of the complexity into the coil, and less intro the rest of the circuit. It is not hard to bring the tap out on a home made coil.

Man, it has been a long time since I last built a 1 transistor oscillator. I plinked around a little and quickly discovered that I have the supply current situation backwards.

The oscillator is biased to draw a significant and continuous supply current, but once it gets going, the rectification effect of the base junction kills the forward bias for most of the cycle, greatly lowering the average collector current, concentrating it into narrow pulses at the positive peak of the cycle. Putting a piece of metal in the inductor field not only lowers the inductance, but also the Q, so the oscillation tends to start and die out , rapidly, so that the transistor spends most of the time in an intermediate state between no oscillation (high current) and active positive feedback oscillation (class C low current operation). The hardest part for me to model is the eddy current losses in the target as the distance is varied.

This may or may not help,

you can download a free .pdf which has "6 components and less" electronic projects. One of the said projects was a metal detector. Might be something to give you a few ideas.

Regards, JW