Find all your DIY electronics in the MakerShed. 3D Printing, Kits, Arduino, Raspberry Pi, Books & more!





Ask MAKE is a weekly column where we answer reader questions, like yours. Write them in to mattm@makezine.comor drop us a line on Twitter. We can’t wait to tackle your conundrums!

image_sensor.jpg

Ian writes in:

I was looking at buying a digital camera, and read that there are two kinds of sensors that they can use to take a picture- CMOS and CCD. Can you tell me what the difference is, and if one is better to get?

Sure! It’s actually a pretty topical question, as the inventors of the CCD just won this year’s Nobel Prize! As you mentioned, there are two basic kinds of image sensor that are used in today’s digital cameras, CCD (charge-coupled device) and CMOS (complementary metal oxide semiconductor). They both work by converting light energy (photons) into electric charge (electrons), and the difference is in how this charge is read out.

To start, both kinds of sensor are made of a grid of ‘buckets’ placed evenly across a flat semiconductor surface. Each bucket acts as an individual sensor, which only sees a tiny portion of the image. By displaying a bunch of these tiny points in a grid (using a computer monitor or printer), we see the image.

Making color images is a bit more complicated. Because the buckets are sensitive to any wavelength of visible light, if we just looked at the results we would see a monochrome image. To get color information, we arrange the sensors into groups of four, and place tiny red, green, and blue color filters over them. Each group of four sensors is what we call a pixel, and it is interesting to note that modern cameras have millions of them.

Ok, so both CCD and CMOS sensors are basically just big arrays of individual sensors, so how are they different? The difference is in how the charges are collected and read out. In a CCD, the ‘bucket’ that collects charge is just a capacitor. To read the image data out of the CCD, the charge in each bucket is pumped individually over to an ADC (analog to digital converter), which actually measures charge. In a CMOS sensor, each bucket contains a photodiode and some amplifier circuitry. To read the image data out, the output of each amplifier is connected to an ADC through a multiplexer, which measures the voltage at each cell.

I don’t think that either technology is necessarily better, but each has its own quirks. There is an interesting site at dvxuser which talks about the different kinds of sensor artifacts associated with each kind of sensor. For most cases, though, I think that other specifications, such as ease of use and sensitivity to light, are probably more important to think about when choosing a digital camera. Good luck!

[photo by SarahCartwright]


Related

Comments

  1. Anonymous says:

    complementary, not complimentary.

    1. Matt Mets says:

      Thanks, fixed!

  2. Michael Black says:

    As an aside, some of the early “play” with digital cameras
    was using dynamic ram with the top pried off. You’d cycle
    through the row and columns, and the “picture” would appear
    the data pin(s).

    That would be the CCD scheme.

    The “Cyclops” was an early “digital camera” featured in Popular
    Electronics a few issues after the Altair 8800 hit the magazine’s
    cover. It was put together by the two guys who later formed
    :Cromemco”. That used such a RAM, but I can’t remember if they
    used an off-the-shelf dynamic RAM or something more specifically
    designed for the task. Memory says it was some oddball part that
    the authors were the only source for.

    There were later discussions in the hobby electronic magazines about prying the top off dynamic RAM to get a sensor. It was
    common knowledge, so I can’t point to any specific reference.
    In the mid-eighties, Steve Ciarcia had a digital camera project
    Byte, and that used dynamic ram that had been sorted or repackaged
    for camera use.

    The movie “Westworld” had a robot gone bad, played by Yul Brenner,
    and every time the audience saw what he saw, it was a terribly
    pixilated view. Either I read or I assumed at the time, circa
    1974, that it was a very low resolution digital camera used for
    that.

    Michael

    1. Matt Mets says:

      I’ve heard rumors about using DRAM to make a sensor, and haven’t been able to find much (except a patent from 1984: http://www.patentstorm.us/patents/4441125/description.html) on the subject. One thing that I am not clear on is how to form a grayscale image using the digital sense circuitry- it seems like it would see each pixel as either on or off, and nothing in between. The best I can think of is to perform multiple exposures, and either very the image exposure time or the sense comparator voltage to measure different levels. Either way should work, but will require many exposures to build up any sort of bit depth.

  3. KS Nayar says:

    I know several people with expensive SLR cameras that now collect dust because they are film cameras that people do not use anymore! I was wondering whether one could design a sensor that could be slapped onto the film-plate and use the film cassette space to wire a circuit that can capture the image in colour like any modern digital camera. This way these expensive cameras can be put to good use. Is this a feasible project for you to take up for the readers who can try building a project to convert film cameras into digital cameras?