uch has been made in thepast five years of the poten-tial for CMOS imagers andof the impending demise of the in-cumbent image-sensing technology,CCDs.Strong claims by the proponentsof a resurgent CMOS technology have been countered by equally force-ful claims by CCD defenders. In a pattern typical of battling technolo-gies (both with significant merits but also lacking maturity in some re-gards), users have become leery of performance representations made by both camps. Overly aggressivepromotion of both technologies hasled to considerable fear, uncertainty and doubt.

Imager basics

For the foreseeable future, there will be a significant role for both typesof sensor in imaging. The most suc-cessful users of advanced image cap-ture technology will be those whoconsider not only the base technol-ogy, but also the sustainability,adaptability and support. They willperform the best long term in a dy-namic technology environment that the battle between CCDs and CMOSpromises to deliver.Both image sensors are pixelatedmetal oxide semiconductors. They accumulate signal charge in eachpixel proportional to the local illu-mination intensity, serving a spatialsampling function. When exposure is complete, a CCD(Figure 1) transfers each pixel’scharge packet sequentially to a com-mon output structure, which converts the charge to a voltage, buffersit and sends it off-chip. In a CMOSimager (Figure 2), the charge-to-volt-age conversion takes place in eachpixel. This difference in readout tech-niques has significant implicationsfor sensor architecture, capabilitiesand limitations.Eight attributes characterize image-sensor performance:•

Responsivity

, the amount of sig-nal the sensor delivers per unit of input optical energy. CMOS imagersare marginally superior to CCDs, ingeneral, because gain elements areeasier to place on a CMOS image sen-sor. Their complementary transis-tors allow low-power high-gain am-plifiers, whereas CCD amplificationusually comes at a significant power penalty. Some CCD manufacturersare challenging this conception withnew readout amplifier techniques.•

Dynamic range

, the ratio of a pixel’s saturation level to its signalthreshold. It gives CCDs an advan-tage by about a factor of two in com-parable circumstances. CCDs stillenjoy significant noise advantagesover CMOS imagers because of qui-eter sensor substrates (less on-chipcircuitry), inherent tolerance to buscapacitance variations and commonoutput amplifiers with transistor geometries that can be easily adaptedfor minimal noise. Externally cod-dling the image sensor through cool-ing, better optics, more resolution or adapted off-chip electronics cannot make CMOS sensors equivalent toCCDs in this regard.

Reprinted from the January 2001 issue of

PHOTONICS SPECTRA

CCD vs. CMOS:

by Dave Litwiller

Choosing an imager means considering not only the chip,butalso its manufacturer and how your application will evolve.

CCD vs.CMOS

CCD vs. CMOS:

BiasGenerationClock &TimingGenerationOscillatorClockDriversLineDriverAnalog-to-DigitalConversionGainTo FrameGrabberElectron-to-VoltageConversionPhoton-to-ElectronConversionCharge-Coupled DeviceImage SensorCamera(Printed Circuit Board)

Figure 1.

On a CCD, most functions take place on the camera’s printed circuit board. If the application’s demands change, a designer can change the electronics without redesigning the imager.

Facts and Fiction

•

Uniformity

, the consistency of response for different pixels under identical illumination conditions.Ideally, behavior would be uniform, but spatial wafer processing varia-tions, particulate defects and ampli-fier variations create nonuniformi-ties. It is important to make a dis-tinction between uniformity under illumination and uniformity at or near dark. CMOS imagers were tra-ditionally much worse under bothregimes. Each pixel had an open-loop output amplifier, and the offset and gain of each amplifier varied con-siderably because of wafer process-ing variations, making both dark andilluminated nonuniformities worsethan those in CCDs. Some peoplepredicted that this would defeat CMOS imagers as device geometriesshrank and variances increased.However, feedback-based ampli-fier structures can trade off gain for greater uniformity under illumina-tion. The amplifiers have made the il-luminated uniformity of some CMOSimagers closer to that of CCDs, sus-tainable as geometries shrink.Still lacking, though, is offset vari-ation of CMOS amplifiers, whichmanifests itself as nonuniformity indarkness. While CMOS imager man-ufacturers have invested consider-able effort in suppressing dark nonuniformity, it is still generally worse than that of CCDs. This is a significant issue in high-speed ap-plications, where limited signal lev-els mean that dark nonuniformitiescontribute significantly to overallimage degradation.•

Shuttering

, the ability to start and stop exposure arbitrarily. It is a standard feature of virtually all con-sumer and most industrial CCDs,especially interline transfer devices,and is particularly important in ma-chine vision applications. CCDs candeliver superior electronic shutter-ing, with little fill-factor compromise,even in small-pixel image sensors.Implementing uniform electronicshuttering in CMOS imagers requiresa number of transistors in each pixel.In line-scan CMOS imagers, elec-tronic shuttering does not compro-mise fill factor because shutter tran-sistors can be placed adjacent to theactive area of each pixel. In area-scan (matrix) imagers, uniform elec-tronic shuttering comes at the ex-pense of fill factor because theopaque shutter transistors must beplaced in what would otherwise bean optically sensitive area of eachpixel. CMOS matrix sensor design-ers have dealt with this challenge intwo ways: A nonuniform shutter, called a rolling shutter, exposes different linesof an array at different times. It re-duces the number of in-pixel tran-sistors, improving fill factor. This issometimes acceptable for consumer imaging, but in higher-performanceapplications, object motion manifestsas a distorted image. A uniform synchronous shutter,sometimes called a nonrolling shut-ter, exposes all pixels of the array at the same time. Object motion stops with no distortion, but this approachconsumes pixel area because it re-quires extra transistors in each pixel.Users must choose between low fillfactor and small pixels on a small,less-expensive image sensor, or largepixels with much higher fill factor ona larger, more costly image sensor.•

Speed

, an area in which CMOSarguably has the advantage over CCDs because all camera functionscan be placed on the image sensor.

To FrameGrabberLineDriverClock &TimingGenerationOscillatorAnalog-to-DigitalConversionGain

E l e c t r o n - t o - V o l t a g e C o n v e r s i o n P h o t o n - t o - E l e c t r o n C o n v e r s i o n

Complementary Metal Oxide SemiconductorImage SensorCamera(Printed Circuit Board)

C o n n e c t o r B i a s D e c o u p l i n g B i a s G e n e r a t i o n

Column AmpsColumn Mux

R o w D r i v e r s R o w A c c e s s

Figure 2.

ACMOS imager converts charge to voltage at the pixel, and most functions are integrated into the chip. This makes imager functions less flexible but, for applications in rugged environments, a CMOS camera can be more reliable.

CCD vs.CMOS

With one die, signal and power trace distances can be shorter, with less inductance, capacitanceand propagation delays. To date,though, CMOS imagers have es-tablished only modest advantagesin this regard, largely because of early focus on consumer appli-cations that do not demand no-tably high speeds compared withthe CCD’s industrial, scientificand medical applications.•

Windowing

. One unique ca-pability of CMOS technology isthe ability to read out a portion of the image sensor. This allows el-evated frame or line rates for small regions of interest. This isan enabling capability for CMOSimagers in some applications,such as high-temporal-precisionobject tracking in a subregion of an image. CCDs generally havelimited abilities in windowing.•

Antiblooming

, the ability togracefully drain localized overex-posure without compromising therest of the image in the sensor.CMOS generally has natural blooming immunity. CCDs, on theother hand, require specific en-gineering to achieve this capabil-ity. Many CCDs that have beendeveloped for consumer applica-tions do, but those developed for scientific applications generally donot.•

Biasing and clocking

. CMOS im-agers have a clear edge in this re-gard. They generally operate with a single bias voltage and clock level.Nonstandard biases are generatedon-chip with charge pump circuitry isolated from the user unless there issome noise leakage. CCDs typically require a few higher-voltage biases, but clocking has been sim-plified in modern devices that op-erate with low-voltage clocks.

Reliability

Both image chip types are equally reliable in most consumer and in-dustrial applications. In ultraruggedenvironments, CMOS imagers havean advantage because all circuit functions can be placed on a sin-gle integrated circuit chip, mini-mizing leads and solder joints, which are leading causes of cir-cuit failures in extremely harshenvironments.CMOS image sensors alsocan be much more highly inte-grated than CCD devices. Timing generation, signal pro-cessing, analog-to-digital conversion, interface and other functions can all be put on theimager chip. This means that a CMOS-based camera can besignificantly smaller than a comparable CCD camera. The user needs to consider,however, the cost of this inte-gration. CMOS imagers aremanufactured in a wafer fab-rication process that must betailored for imaging perfor-mance. These process adapta-tions, compared with a non-imaging mixed-signal process,come with some penalties indevice scaling and power dis-sipation. Although the pixelportion of the CMOS imager al-most invariably has lower power dissipation than a CCD,the power dissipation of other circuits on the device can behigher than that of a CCDusing companion chips fromoptimized analog, digital andmixed signal processes. At a systemlevel, this calls into question the no-tion that CMOS-based cameras havelower power dissipation than CCD- based cameras. Often, CMOS is bet-ter, but it is not unequivocally thecase, especially at high speeds (aboveabout 25-MHz readout). The other significant considera-tions in system integration are adapt-ability, flexibility and speed of change. Most CMOS image sensorsare designed for a large, consumer or near-consumer application. They are highly integrated and tailored for one or a few applications. A systemdesigner should be careful not to in- vest fruitlessly in attempting to adapt a highly application-specific devicefor a use to which it is not suited.CCD image sensors, on the other hand, are more general purpose. Thepixel size and resolution are fixed inthe device, but the user can easily tailor other aspects such as readout

Figure 3.

Are they really stars? For an ideal detector, each pixel’s response to a photon would be identical, and the “starlight”would be confined to the area of the star.

CCD vs.CMOS

Choose Your Imager

CMOS imagers offer superior integration,power dissipation and system size at theexpense of image quality (particularly inlow light) and flexibility. They are the tech-nology of choice for high-volume, space-constrained applications where imagequality requirements are low. This makesthem a natural fit for security cameras, PC videoconferencing, wireless handheld device videoconferencing, bar-code scan-ners, fax machines, consumer scanners,toys, biometrics and some automotive in- vehicle uses.CCDs offer superior image quality andflexibility at the expense of system size. They remain the most suitable technol-ogy for high-end imaging applications,such as digital photography, broadcast television, high-performance industrialimaging, and most scientific and medicalapplications. Furthermore, flexibility means users can achieve greater systemdifferentiation with CCDs than with CMOSimagers.Sustainable cost between the two tech-nologies is approximately equal. This isa major contradiction to the traditionalmarketing pitch of virtually all of the solely CMOS imager companies.