18.04.2012 - DIY camera: Start here |
A pinhole camera is an old concept. With a small enough hole an image can be projected on a plane opposite to it.
Incidentally, this mechanism is quite independent of the wavelength observed.
The optimal size of the pinhole is calculated by this formula:
Where f is the distance from the hole to the plane, lambda is the wavelength, and d is the pinhole diameter.
The image, projected from a pinhole will be quite faint. To read it off, we need a good sensor.
Normally, that would be a job of a CCD matrix or somesuch, but that's not our way.
Photodiodes are made in a variety of wavelengths, and are normally senitive enough for the task.
With only one photodiode it have to be moved across the image plane.
So, here is the basic design.
Make a box:
Make a pinhole mount in it:
Put a cartesian robot inside:
Put sensor on the robot:
And you got a camera. :)
Details follow on the next page
|I know I'm very late to this party, but have to ask...What your final transimpedance amp. consisted of, as far as components, and your final favorite photodiode, used?|
20.03.2019 01:43 - Roddy
|I'm not sure which resistor are you talking about.
If it's the one in the perfboard photos, then this is so out of date as to be irrelevant.
18.04.2015 01:15 - Artlav
I see that you have used a 20k Ohms resistor in your circuit, what capacitor did you used to go with it?
14.04.2015 16:16 - Alejandro
|Two columns in one direction align naturally, as long as the mechanics aren't too loose.
Two adjacent columns going in opposite directions, however, are still an open problem.
It's repeatable enough to align the central parts by a fixed offset in post-processing, but there are edges where things get non-linear. I'm thinking of making a sort of a correction table, by taking a picture of a grid.
If that's what you were asking.
09.02.2015 09:56 - Artlav
|You made a great system. I Working on a same concept by moving the sensor with servos.
I have a question, How did you manage to rectify the offset of a pixel column to another?
Have A nice day!
09.02.2015 03:26 - Chyneuze
|The noise is RF in origin - what helped is wrapping the box in grounded foil, eliminating ground loops and putting the sensor in a shielded box of it's own.
Light/reflection noise is not something i noticed or encountered, and painting it black will be tricky, since many "black" paints turn bright white in NIR. I'm not immediately aware of any paint that would be uniformly black across the entire spectral range of that camera.
The code is too simple to be worth posting - drive the steppers, sense the light, write to the card. All three are there in Arduino's set of demos.
08.02.2015 23:53 - Artlav
|I get what you are saying in regards to electrical noise
If you make the inside of your box matte black, like all cameras are, you will notice a big difference in noise level and contrast.
I'm working on one of these right now, the initial test camera obscuras I've built work WAY better when you all the internal surfaces black.
Will you be posting your code???
07.02.2015 14:35 - cyberteque
|@j: Not really. The sapphire window, peltier stabilised 3400nm one was the most expensive at about $57 equivalent.
@Andre: They would be waaay too large. Sensor size is pixel size - the larger the sensor, the lower the resolution is.
22.01.2015 22:34 - Artlav
|Photodiodes from IBSG...that could not had been cheap.|
22.01.2015 22:04 - j
|Hi, just a thought but have you tried using a solar cell as a sensor?
The ones used on solar lamps work reasonably well, or you could also harvest them from old photoelectric smoke alarms.
I found two nice 7mm2 diodes this way, good for several mA when put next to a TV remote.
22.01.2015 21:50 - Andre
|@vimtut0r: I did, back in 2013 (the pages are quite a bit out of date). Simplifies the code a lot, makes the image somewhat cleaner.
@The Lightning Stalker: Pieces of coloured glass, "hot mirror" window from a DSLR, UV filter glass, polished piece of silicon, and several others.
22.01.2015 13:04 - Artlav
|Very cool project!
Why don't you just place an ATTiny or something similiar right next to the OpAmp to reduce the signal length?
22.01.2015 12:09 - vimtut0r
|What filters are you using for wavelength selectivity?|
22.01.2015 04:44 - The Lightning Stalker
|Thanks a lot!|
Ill let you know if it worked for me
AD820 for the pre-rebuild stuff.
AD8605 for the latest version.
In my case i needed to measure sub-pA currents, thus the need for very low-offset op-amp.
|Which op amp did you used?|
Im trying to build a green index sensor!
Comments disabled for lack of purpose and excess of spam.
If you have a comment on some project, ping me on the corresponding forum.