How to design an infra-red circuit to detect small obstacle (e.g. finger) ?

Hi.

It would help if you stated what optical setup you have in mind. Would it be enough to detect that a narrow beam has been interrupted?

A simple way to make a fairly sensitive transceiver and detector is to use a NE567 tone decoder. Depending on the gain you need, an additional op-amp stage might be needed or you can use a phototransistor as a gain element with suitable external components.

This virtues of this technique are low parts count and relative insensitivity to DC or line frequency light.

If this interests you, say so and I'll provide such details as are not obvious.

Larry, - I do need a narrom detection indeed. So if I have a 2 feet separation and there is a narrow beam where if someone place a finger or pass his finger thru the narrow beam quickly it should detect that.

- As per level of gain- I need to make a sound of a certain frequency when the obstacle is detected.

So, if you can please provide me details on how to make a circuit like that it will be wonderful.

Thanks, Anwar

You can buy the little receiver blocks like TV sets use. They detect modulated IR, very sensitive, 40 KHz or something. The transmitting LED could be modulated by - gasp - a 555 or something.

John

You might want to look at safety devices for punch presses etc. I think they use an array of emitters and detectors.

Best regards, Spehro Pefhany

I'm not going to do the detail design or any simulation on this. I have built similar circuits and used the NE567 in this manner, so I can assure you that this technique will work with appropriate design and adjustments for your specific application.

The NE567 provides these functional blocks:

1. An RC oscillator with a useful semi-digital output on pin 5. This can drive a low-power LED directly or, with an extra transistor, as much as your LED can take.
2. A synchronous demodulator, multiplying the signal on pin 3 to produce filtered baseband on pin 1. You can set the demodulator bandwidth according to how fast you want detection to happen versus how much out-of- band noise you wish to reject.
3. A reference and comparator to detect when the level on pin 3 exceeds a threshold, producing an open-collector output on pin 8.

You most likely need to provide some gain between the photocurrent signal and pin 3 of the NE567. The net response of that stage, from photo-current to pin 3, should have 90 degrees of phase shift to work well with the NE567 demodulator. (That might be 270 degrees, so the LED might have to be turned around or the polarity of its BJT driver inverted.) For modest gain, a phototransistor can be used in a circuit with the following topology to provide this function:

VCC + | .-. | | | | '.' .---------o | | || .-. o------||- PhotoRx | | | || | | | '-' | | |/ o-------| .----o |>

| | | --- .-. o---. --- | | | | | | | .-. | | '-' | | --- | | | | --- '----o '-' | | o---' | | === === GND GND (created by AACircuit v1.28.4 beta 13/12/04

This can be done with a very compact layout near the receiving phototransistor to minimize noise pickup. The LED can be driven by some twisted pair.

As for making a tone when pin 8 is driven low by detection of the beam, (or opened by interruption of the beam), I will leave that problem to you.

....

....

What I like about the NE567 for this purpose is that the detection bandwidth can be adjusted and carrier frequency adjusted up to a hundred kHz or so. The circuitry is quite generic and compact. And there is no issue trying to match transmit and receive frequency, even if low precision parts are used.

I'm just an EE student right now, so I am trying to learn from this question. Please tell me what is wrong. That being said: Why does the IR beam have to be modulated? What would be wrong with just putting a straight IR beam (IR LED with a lens?) hitting an IR photodiode or phototransitor. Invert the signal from that (just hook up the connection at the collector of the phototransistor), and hook it up to an LED or relay for the indicator light.

....

With careful attention to the optics and in a benign environment, that might work. Using modulated light and narrowband detection, problems with ambient light can be easily precluded and the effort to build that circuit is comparable to or easier than getting optics designed and aligned to respond only to the intended beam.

I think Holtek makes IR transmitter ic's that would work well too.

Detecting IR at one modulated frequency reduces the ambient noise greatly, higher signal to noise. That allows you to drop power levels, etc., and still get it to work under all conditions.