Thermal "Camera"

I am trying to make a crude camera using an array of pyroelectric detectors and an lcd shutter.
Array initially will nominally 8 by 6.

The shutter will probably have a "top speed" of about 100Hz.

Its basically the same as a pir but used to give an indication of where in the fov someone is.

This will be kept at a constant temperature to avoid the need for thermal calibration although differences between them will probably require individual calibration.

I m assuming anything in the fov will be moving slow enough for this to work.

What I would like to do is select each detector in turn to multiplex to the input of the ADC.

I am unsure of the best way to do this though.

Fet on each detector connected to common line ?
I can calibrate each device in software but I am wondering if noise mat be a problem.

The absolute top rate that you can get ADC on an arduino to work would be about 125 khz. But that's individual readings. Divide that by 48 ( one reading for each of your 8x6 array) and you'll get about 2.5 thousand "frames" per second. But I suspect you might have to halve this due to the various pins sharing the same ADC.

This is still exceedingly fast though. Why the need for such speed?

100 HZ not KHZ

Boardburner2:
100 HZ not KHZ

Ooops :grin:

To explain further.

Element is selected then shutter is fired, repeat.

Should give about 2 frames per sec.

pyroelectric detectors and an lcd shutter.

Have you links to those devises?
AFAIK, PIR is very inertial, a few seconds response time

It is a gifted lucky bag of a few hundred unmarked devices.

Testing one with a pair od 3 d telly glasses would indicate a response time much less than that .

Sparkfun do one which I know responds at 1 KHZ.

Tested by a spinning chopper plate.

AFAIK they actually work best at high frequency not low.

EDIT
This explains how pir work.

Walk testing an alarm pir gives a response of about .5 sec .
I would expect such a device to wait for two transitions to avoid false alarms.

1in 3 seems not to work but that may be me blowing them up.
I do not know if their output is fet or otherwise, im guessing atm.

They are not marked , that could be deliberate as that is what happens often to those that fail test.

I am hoping to get enough workers to experiment with though.

Too expensive to buy individually for me for what is just an interesting experiment.

A typical PIR requires a Fresnel lens to work as a IR (motion) detector, that is because internally, the 2 elements null each other!

You need something like the MLX90614

I have used it in a project with excellent results.

A few weeks back, Hackaday had an article on a slow-scan homemade IR camera.

Ray

My Projects

Hi why do you need a shutter, just sample the array when you need a frame.

Tom.... :slight_smile:

TomGeorge:
Hi why do you need a shutter, just sample the array when you need a frame.

Tom.... :slight_smile:

In layman's terms, a mechanical shutter is used to control how long the pixels on an image sensor collect light. A simple mechanical shutter can be used to turn the entire sensor array on/off during the exposure. This eliminates the need for added electronics at each pixel location that would be used to turn on/off the pixel and store the charge (accumulated light). By using a mechanical shutter, a simpler, less expensive, and more efficient sensor can be used: one that has a higher fill factor (uses more of each pixel to actually capture light).

Ref: here

IR single element sensors often use a shutter or "chopper" to assist with noise reduction and establishing a zero reference for signal processing.

Ray

mrburnette:
A typical PIR requires a Fresnel lens to work as a IR (motion) detector, that is because internally, the 2 elements null each other!

You need something like the MLX90614

I have used it in a project with excellent results.

A few weeks back, Hackaday had an article on a slow-scan homemade IR camera.

Ray

My Projects

I have taped over one half and get an output using a chopper, weather or not its going to be useful I don't know.

Buying 48 of those suggested is not an option.

I saw the article which is what prompted this idea , unfortunately its too slow and I need to avoid mechanics.

Are you sure the LCD is going to work as a shutter?
Substrates are typically glass, and ordinary glass blocks thermal wavelengths very effectively.

AWOL:
Are you sure the LCD is going to work as a shutter?
Substrates are typically glass, and ordinary glass blocks thermal wavelengths very effectively.

No im not.

I do get measurable output but trying sensor with binoculars did not work.
Using a 6 in reflector telescope does.

I am hoping to get away with a cheap plastic lens like pir thermometers.

If I cant find something cheap that works it will have to be a rotating one.

Just found this though, looks much more promising