motion detection / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

motion detection

previous 1 2 3 next

Move Thread

LAN_403

Marko

Thu Aug 16 2007, 12:54AM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

Pyroelectric crystals have an inherently high-pass response, hence the use for motion detectors.

Yeah, but here 'high pass response' is due to second law of thermodynamics itself, and I'm not sure if this would be what Matt needs. Their output manifests in really odd short pulses wich need quite an amplification before being usable.

I also fear that those sensors are quite expensive to be used like so.

ragnar

Thu Aug 16 2007, 02:52AM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

Thanks for your responses guys, last night I had a fiddle, probably most related to Andy / Avalanche's suggestion.

I used a variable resistor and a photodiode as a voltage divider to bias the gate of a small MOSFET, whose drain went to the LED. If I gently start to shadow the photodiode, the LED gently starts to warm up.

This wasn't really motion-sensing, but it looked cool anyway. I considered adding a diode + capacitor in there (essentially across the gate of the MOSFET) to slow down the 'recovery time' to give a delayed cooldown of the LED. I'll play with Andy's scribble to see if I can achieve what I was originally looking for.

Electroholic

Thu Aug 16 2007, 03:56AM

Registered Member #191 Joined: Fri Feb 17 2006, 02:01AM
Location: Esbjerg Denmark
Posts: 720

how about ultra sonic?
ive built this circuit quite some time back and it worked perfectly, i believe it was just some 741opamps and a 40khz crystal.

oops light intensity, dumb me...

distance seems a little bit easier tho

Steve Conner

Thu Aug 16 2007, 08:59AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Hi Matt,

I remember something very cool they had at the "Dome Of Discovery" when I was a kid. It was a whole wall sprinkled with little half-business-card-sized pieces of stripboard, each one having a CdS photocell, a transistor and a flashlight bulb on it.

Each photocell was shaded by a small cardboard tube so that it couldn't see its own torch bulb, and the whole wall was illuminated by a floodlight. The circuit was the classic cookbook one that turns a light on when it gets dark. When you stood in front of the floodlight, casting a shadow on the wall, the bulbs all lit inside your shadow. It was rather cool.

In more recent times, you can do some very cool Web 2.0 type things with IR floodlights, security cameras, a Mac with video capture, and a video projector. I wandered into this installation at a local art gallery last week:

The real Multi Touchscreen isn't just a cool piece of interactive art: it's a whole new metaphor for human-computer interaction. There's some way to make your own version using a piece of plexiglass, a webcam, a data projector, some LEDs and some clever software:

However, it's not much use, because there is practically no application software existing that understands multi-finger gestures anyway. JazzMutant's Lemur

and the iPhone are about two things I can think of.

Simon

Fri Aug 17 2007, 06:07AM

Registered Member #32 Joined: Sat Feb 04 2006, 08:58AM
Location: Australia
Posts: 549

Marko wrote ...

Pyroelectric crystals have an inherently high-pass response, hence the use for motion detectors.

I should clarify that I agree: pyroelectric crystals are probably not appropriate. It's just not the size that's the issue. For an array you want something cheap and simple, like phototransistors.

Playing with CdS sensors is still fun. That sounds like a cool exhibit, Steve.

ragnar

Tue Oct 02 2007, 03:16PM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

Here's a pic of what I've been able to build that remotely represents what I wanted to do.

On the circuit board, there is a phototransistor in the middle, with four red LEDs surrounding its periphery. As the light level decreases above the photosensor, the current to the LEDs increases.

Unfortunately, the effective "range" between dark/light is a bit narrow, because it's based on an absolute value of light, not a rate-of-change-of-light.

By choosing some different resistors and capacitors, I could tweak the circuit a little more, but for now I'm very, very pleased with what I could whip up in my apartment this evening. ^_^

sehrgut

Tue Oct 02 2007, 04:39PM

Registered Member #1017 Joined: Fri Sept 21 2007, 09:08PM
Location:
Posts: 24

I'm echoing Steve here . . . as much as I dislike op-amps, they're by far the easiest way to detect rate-of-change. Google "op-amp differentiator" and you'll come up with an array of circuits from the simple to the mathematically-precise. But I'd experiment with various size caps in series with the input, depending on what you wanted the resolution/sensitivity of the circuit to be.

Tom540

Tue Oct 02 2007, 05:40PM

Banned on 3/17/2009.
Registered Member #487 Joined: Sun Jul 09 2006, 01:22AM
Location:
Posts: 617

If you grab some lm567 tone decoders, they make really cool Ir detectors. They are an obsolete part but you can still get them from mouser or digikey.

Basically you setup the freq you want then through a cap feed the oscillator freq to a transistor with IR LED. Then use a photo diode and couple its output through another cap into the tone decoders input. If you're interested I can dig up a schematic. This works very well.

I've gotten them to work at about 2 feet on a robotics project. You can tweek the output filter and get really good results. I also added some crude Schmitt trigger inverter filters to their outputs.

-Tom

Tesladownunder

Thu Oct 04 2007, 09:22AM

Registered Member #10 Joined: Thu Feb 02 2006, 09:45AM
Location: Bunbury, Australia
Posts: 1424

This device responded to change in light intensity and would turn on as you approach and start flashing LED sequences and burbling away with random electronic tones then softly fade if you remained still. It also has a 16 tune doorbell chip. Lots of variation depending on which corner was upper most and how bright the ambient light was.

Probably you could work out the schematic from the photo as all wiring is point to point. (never learnt how to do PCB's and never made a schematic of it) Mostly cobbled together from odd circuits

It still goes after 20 years plus.

TDU

ragnar

Mon Oct 08 2007, 03:01AM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

Working for an evening with one or two op-amps and some negative feedback, I've got my desired effect. Steve Conner very kindly recommended a concept circuit that utilized bipolar LEDs and could differentiate between "getting brighter" and "getting darker".

I have two Rs and a phototransistor forming a voltage divider network between V- and V+. I capacitively couple one side of the phototransistor to the (-) input. The (+) input is connected to ground. The output feeds back into the (-) input with a 1M resistor.

Individually, each circuit works quite ideally, until I realised that it could sense the 50Hz flicker in fluorescent and even incandescent bulbs. A large capacitor across the phototransistor fixed this to some extent.

Confidently, I've made a 3 x 3 array, now to discover/confirm (as I had worried) that the phototransistors/LEDs can "see each other".

In a dark room, the whole board will oscillate in sync, because as any one sensor "turns off", the other sensors detect that and "turn on".

I've tried turning the phototransistors on their sides, so that they can't see anything until you move real close to the board, and this helped to some extent, but the array still interacts with itself in an unaesthetic way.

I've tried putting little black shrouds around the photosensors (to no avail), and little black shrouds around the LEDs (similarly to no avail).

The inherent problem seems to be this: the LEDs are sufficiently bright to illuminate just about any surface above them, which will no doubt trigger the surrounding parts of the array.

If the LEDs were cheap and crappy enough that they hardly illuminated anything other than their own clouded lens, the problem would disappear, but the effect would look poor, and perhaps even be invisible, in daylight.

Is this an unsolvable problem? Can anyone offer some insight?

(On a sidenote, I'm very, very pleased with this home etch and getting perfect 16 mil tracks (the power rails are 32 mil)

previous 1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob