CMOSIS continuously develops new image sensor technology to improve the performance of its image sensors. This technology supports our custom sensor developments, and is also deployed in our standard products.
The dynamic range is typically limited by the readout process of the pixel. Techniques have been developed in the past to cope with this, usually by non-linear compression of the signal.
CMOSIS developed a new pixel that allows readout of a photodiode with a wide dynamic range, which maintains a linear response to light. After exposure, the photodiode is read out via two transfer gates to two sense nodes. Two signals are then read from each pixel. The first signal only reads charge transferred to the first sense node, with maximal gain. This sample is used for small charge packets and is read with with low read noise. The second sample reads the total charge transferred to both sense nodes, with a lower gain. Pixels with a read noise of 3.3 electrons and a full well charge of 100,000 electrons have been demonstrated, resulting in a linear dynamic range of 90 dB.
Global shutter pixels are needed to capture the scene for all pixels at the same moment in time. CMOSIS developed a unique global shutter pixel that combines true correlated double sampling with an excellent shuter efficiency. Each pixel uses two memory elements, which sample the signal and its reference value. During readout, the difference between both samples is calculated, through a process called correlated double sampling. Noise levels below 10 electrons can be reached, and a shutter efficiency of 99.999% has been demonstrated. This means that if 100,000 photons hit the pixel, only one photon will disturb the image stored in the in-pixel memory elements. The other photons are used to capture the next image.
CMOSIS global shutter pixel is the only pixel that combines excellent shutter efficiency with low read noise through correlated double sampling. It is also the only global shutter pixel that is compatible with backside illumination technology.
These slides illustrate an implementation in CMOSIS' CMV sensors.
CMOSIS has developed patented, fast AD converters with cyclic and ramp ADC topologies.
The majority of CMOSIS products use counting ramp AD converters, which have been developed at 10, 12 and 14 bit ADC resolution. Counting ramp AD converters use a comparator and a counter in each column. The comparator compares the sampled data in the column amplifier with a reference sawtooth waveform. The output of the comparator enables or disables a counter in the column. The ADC performs correlated double sampling together with AD conversion through sequential conversions of the reference and signal sample, and by a smart inversion of the comparator output between both cycles. This differential process also cancels clock delays. That allows to operate the AD converter at much higher clock frequencies than usual in the industry. As a result, we convert a full row to 10 bit resolution in less than 2.7 µs in our CMV2000 & CMV4000 products
A second generation AD converters uses faster conversion cycles through the use of a local oscillator in the column, which allows to increase speed by another factor of 4.
Backside illuminated CMOS image sensor technology has been developed over the last years to improve the sensitivity of very small pixels for consumer applications in the visible band. CMOSIS uses this technology not for consumer applications, but for scientific and industrial imaging.
CMOSIS developed a backside illumination flow for extensions of the wavelength band beyond the visible range. Devices have been developed for detecting light outside of the visible band, in the near UV (200-400 nm) or extreme UV band (15-50 mm). Near UV imaging has been demonstrated on global shutter pixels. Developments in extreme UV were done with low noise high dynamic range pixels. Since every photon generates more than 15 electrons, and our read noise is only 3.3 electrons, we are capable of counting individual photons in the EUV band. Even spectroscopy is possible by measure the energy of the photon from the number of generated electrons in the EUV band.
Large area image sensors, with dimensions up to wafer scale have been reported for several years in the scientific and commercial literature. Economically, the use of such sensors has been very limited because of low production yields. Every wafer contains defects which result in most cases to a failing sensor. These killer defects are hard to avoid, the sensor has to tolerate them. We patented a new methodology to drastically reduce the sensitivity of CMOS image sensor pixels to defects occurring on a wafer. This allows to make these large devices also on an economic scale. Applications can be found in medical X-ray equipment, large and medium size photography, and scientific instrumentation.
Time Delayed Integration or TDI imaging is used to image moving objects. By synchronizing the pixels with the motion of the camera or the object, the effective exposure time can be increased. Implementations in CMOS have traditionally been difficult because of the lack of a charge addition circuit. The application requires the combination of a global shutter and a low-noise readout method. We found new ways to solve this problem. This will typically be applied in earth observation instruments, inspection systems and machine vision.