Image censoring with radio frequency (RF) in CMOS is a combination of light sensing chips and wireless communication. Typically, we were first engaged in the article, “RF Design Issues and Challenges in a CMOS Image Sensor Process”, because of the circuit design process required to make a functioning Radio Frequency transceiver. Radio Frequency transceivers are electronic devices that receive and demodulate radio frequency signals, and then modulate and transmit new signals. By using these radio frequency signals, it is a way to limit local interference and noise. We thought it would be an interesting concept to grasp for our future benefit if we decide, design and implement these communication devices into CMOS Image Sensors or CIS.
The first type of CMOS pixel circuits created were called passive pixels. They had good fill factors but suffered from very poor signal to noise performance. [2] Modern CMOS designs mostly use active pixels, which put an amplifier in each pixel, typically constructed with three transistors. [2] The more transistors added to CMOS Image Sensor designs the less noise interference, which is a good thing especially for the CIS with RF since the RF will produce noise interferences.
CIS is a relatively simple circuit that involves Photodetectors, these are very light sensitive, when activated they take in how much light each one is being exposed to and stores it by collecting these readings into pixels. Typically, each circuit would form four pixels; two green, one blue, one red. These sensors individual circuits are expanded until a resolution pixel size is met. CIS is very popular among many devices these days for many reasons, the biggest is perhaps how easy they are to produce which makes them cheap. What makes CIS so cheap is there are many manufacturers making CMOS chips at very affordable prices, the other option for digital image capturing is CCD, Charge Coupled Device.
CCD is cheaper to make individually when compared to CMOS but at large scale the CMOS CIS prove to be much cheaper thanks to the large number of foundries that produce various CMOS circuits. Furthermore, other than being very economically efficient they are capable to be very smart. What gives them the capability to be smart is that since the circuit is almost entirely CMOS, calculations can be made without an intense data path and processor carrying out corrections on each individual CIS circuit, the image can be corrected on each individual pixel and then passed on which is beneficial for many reasons such as; low processing consumption, more accurate pixel correction, and quicker image processing.
The article selected makes some notable points relatable to what we have already talked about in class. The circuit being described is 0.18um, in class the transistors are sized at 0.22um. The 0.22um size used in class, for at least a few examples, were never defined as the standard that needed to be used. It is impressive that the circuits designed in the article are that small, furthermore, the circuits in class are strictly transistors but here inductors were involved in the LVS. This is interesting because lecture gave the impression the CMOS would be a circuit and the components, such as the inductors in this article, would be two separate portions. This is not how the circuit is designed, however, the foundry needed to put inductors in the CMOS portion.
This demonstrates that if this same circuit from the article had an issue of overheating and it needed to be monitored constantly then a simple thermistor could be included at the foundry and a CMOS RF Image sensor processing with self-monitoring temperature would be made. Another point this article related to class was that, in class everything is about achieving maximum efficiency from the circuits. An interesting thing here is they have two goals; Wirelessly Communicate and Detect light for the Photodetectors. If they focus on the Photodetectors running as efficiently as possible then the inductors will either not perform or perform poorly, so instead they added a fifth metal at a thickness of 2um at the foundry so that the inductor performs at its absolute best, assuring the wireless communication is solid. In the process of gaining maximum performance for wireless communication, the transistors will have a smaller dopant level thus limiting their performance to lower than what it could have been.
The article extensively mentions one of topics from the CMOS class and that is the idea of power distribution and noise in these circuits. It is always an innovative idea to be able to control the amount of power that is consumed and not carelessly drain it. Hence, you must be able to put into prospective the Energy and Power equations. To find energy, the first thing that needs to happen is define the Instantaneous Power. Instantaneous Power is found by P(t) =I(t)V(t).
Energy is the rate of power consumption, and is defined by equation (1) listed below. Average power is the energy consumed by a circuit over a given period time. It can be found using the equation (2) listed below. The article mentions the chips must be low power due to the small die size. If the chip already has minimal power and is also receiving radio frequencies, that are at extraordinarily high frequencies, it seems that the level of noise in the chip would become a major issue, perhaps severe enough to hamper its communication capability or the behavior of the transistors. Capacitors are always a good source to reduce the amount of noise interference when dealing with AC circuits. Granted the diagram shows a capacitor which filters noise when it leads to ground, this capacitor doesn’t directly lead to ground.
The article mentions the chips must be low power due to the small die size. If the chip already has minimal power and is also receiving radio frequencies that are at extraordinarily high frequencies, it seems that the level of noise in the chip would become a major issue, perhaps severe enough to hamper its communication capability or the behavior of the transistors. Capacitors are always a good source to reduce the amount of noise interference when dealing with AC circuits.
Granted the diagram shows a capacitor which filters noise when it leads to ground, this capacitor doesn’t directly lead to ground. These topics bring up the famous topic of CCD vs. CMOS. Both sensors were created around the same time, but CCD was preferred at the beginning. CMOS required smaller features that could not be attained at the time and CCD operated at a faster speeds. Later, better technology was created to implement CMOS sensors. It was also realized that CCD sensors have higher bandwidth and more noise.
“Students in The University of Mississippi Electrical Engineering’s Digital CMOS/VLSI Design course researched a contemporary issue and wrote a blog article about their findings for presentation on SemiWiki. Your feedback is greatly appreciated.”
References:
[1] L. Truong, D. Zhang, T. Leitner, and B. Mansoorian, “RF Design Issues and Challenges in a CMOS Image Sensor Process,” Image Sensors. [Online]. Available at: http://www.imagesensors.org/past workshops/2013 workshop/2013 papers/07-09_069-truong_paper.pdf. [Accessed: 11-Mar-2016].
[2]“CMOS Fundamentals,” CMOS Fundamentals. [Online]. Available at: http://www.siliconimaging.com/cmos_fundamentals.htm. [Accessed: 11-Mar-2016].
[3] Matthew Morrison. (2016, March 1). CMOS/VLSI-Lecture 11 [Online]. Available at:
https://www.youtube.com/watch?v=V4pvT8FjXEA [Accessed: 11-Mar-2016].
[4]“Teledyne DALSA – A Teledyne Technologies Company,” CCD vs. CMOS. [Online]. Available at: https://www.teledynedalsa.com/imaging/knowledge-center/appnotes/ccd-vs-cmos/. [Accessed: 11-Mar-2016].
[5]“Theory & Definitions,” Voltage, Current, Power & Energy: Definitions. [Online]. Available at: http://meettechniek.info/measurement/theory-definitions.html. [Accessed: 11-Mar-2016].
[6]Mattias Myrman, “De-aggregating and dispersing dry medicament powder into air,” [Online]. U.S. Patent WO 2003086517 A1, October 23, 2003. [Accessed: 11-Mar-2016].
Share this post via:
Build a 100% Python-based Design environment for Large SoC Designs