Ultrasonic Transducer Alarm

[hugoender]

Why not build one similar to an ultrasonic range finder?
The Tx unit sends a pulse which hits the opposite wall and the Rx unit receives the echo. You would sound the alarm if you didn't receive a return pulse, or the pulse came back too soon or the pulse came back too late.
If there is no wall to reflect off, then you would simply sound the alarm whenever you received a return echo from any distance. This would be easier to implement than the first example.

Well there is a wall nearby (about 2-3ft from the circuit) but if a person crossed, the signal would not be absorbed but instead reflected. Therefore you cannot make a circuit that detects when the circuit does not receive a return pulse.

To detect if a pulse came back too soon or too late is essentially what we have been discussing (the doppler effect).

The best would be if there was no wall to reflect the signal. Then I could make the circuit just detect when a signal is reflected back to the circuit by a person passing by. However, this is not the case and there is a wall. Maybe I could fine tune the power so that the signal is not powerful enough to travel to the wall and back... however, the difference in distance from the wall to the circuit and a person to the circuit is so small that it is probably not possible to fine tune it to such a degree (especially when running off a battery that is not a stable voltage source since it will run low).

 

[hugoender]

You might be right mneary. I was just looking at my concept and people may be sonic reflectors instead of absorbers in which case the circuit might have to detect an increase in RX power. Anyways, I think I am out of suggestions

Yeah mike... the people will not absorb the signal, they will reflect it back to the circuit.

There HAS to be a way (simple way) to do this ...

 

[audioguru]

A fat person wearing a thick winter coat won't reflect the ultrasonic beam. A naked person with "abs of steel" will.

 

[hugoender]

A fat person wearing a thick winter coat won't reflect the ultrasonic beam. A naked person with "abs of steel" will.

lol well there is no telling who will be passing by this alarm

 

[Mikebits]

To detect if a pulse came back too soon or too late is essentially what we have been discussing (the doppler effect).

Not exactly. Doppler measures frequency change, The method Kchriste is talking about measures time changes. Sorta like sonar. So a sonic device emits a pulse and expects a reflected return pulse in x amount of seconds. Any deviation from expected is considered a intrusion. One major difference is a doppler system can emit from TX continuously, where a system as Kchriste describes must be pulsed. In the latter case, you will use counters as your measurement device.
Simple in concept, not so simple to implement.

 

[hugoender]

Mikebits, the whole time I was reading your explanation of what kchriste meant, this was exactly what I was thinking:

Simple in concept, not so simple to implement.

 

[hugoender]

Today I will brainstorm some more and see what rough ideas I can come up with. However, I have a feeling that I am trying to simplify something that is inherently complex (or more complex than I would like). If that is the case then hey what the hell, maybe I will try the complex and see if I can learn from it

 

[hugoender]

This actually works pretty well:

https://www.electro-tech-online.com/custompdfs/2008/07/k49-2.pdf

and here's an "interesting" design:

**broken link removed**

So I went back to this post and looked at the kitsrus link. I then proceeded to look for their first version of that circuit and found: https://www.electro-tech-online.com/custompdfs/2008/07/k20.pdf

I read the circuit description they gave and looked at the schematic but I do not understand how that circuit works. They say in the description that the phase changes when person moves... but isn't it the frequency that changes? Even when there is no person in the way, the signal that is reflected of the wall will be out of phase from the original signal. Wouldn't this cause a false trigger? Or is this circuit assuming that there is no wall for the signal to get reflected back to the receiver when there is no person in the way?

If you guys could take a look at this circuit and give me your thoughts, I would greatly appreciate it.

 

[mneary]

On kit20 I would replace the 741's with something a little less lousy like the TL072. I would also consider replacing T3 with a JFET, and change the bias accordingly (much higher impedance input, maybe 10Meg/3Meg.) I would also put a 100n capacitor across the power input as well as another across the op amp supply pins. And I would move the op amp bias point to the center of their common mode range instead of Vcc/2 (the two 10k resistors.)

I also haven't found how kit20 distinguishes phase. It looks like an amplitude detector to me (senses changes in amplitude). The 4.7u capacitor in parallel with 10k is the low pass filter similar to the other circuits, and the 100n in series with 100k is the high pass filter. They've included an envelope detector in place of a Schmitt trigger. U3 is a crude but effective latch.

Note that the kit20 schematic doesn't include a transmitter, I guess you just borrow the one from kit49.

 

[pc88]

If you guys could take a look at this circuit and give me your thoughts, I would greatly appreciate it.

Here's a Poptronix article which explains a lot of what's going on:

https://www.electro-tech-online.com/custompdfs/2008/07/UltrasonicMotionDetector.pdf

Basically it tracks the derivative of the envelope of the incoming signal. If it's too large, it sets off the alarm.

This means that initially you'll get a false reading when you power up the unit, but it's also able to adapt itself to whatever environment (i.e. reflections) it's in.

 

[hugoender]

Thank you for that link pc88.

Yeah, it seems like this project is a bit much for me. I mean, I can probably build it and have it work but all I will be doing is reading someone else's schematic and just making the circuit. Hardly learning if at all. I guess I am going to have to go with the PIR or IR solution. I will create another thread to start developing a new circuit.

Thank you all for your help. I have learned a tremendous amount of information that I otherwise probably would not have come across for quite some time.

 

[mneary]

You can do it. Take it step by step.

 

[raifushika]

ei guys please help in searching any simple alarm circuit.....

i need it for our project in our shop in school.....

im a electronics engg. tech. student...

will some can help me???....

 

[mneary]

Have you read through this thread from the beginning? What questions do you have?

 

[hugoender]

I still havent gotten around to making the PIR sensor alarm (or the thread for that matter). Been way too busy but now I think I will have so more time and might start looking into it again. So standby for more questions