What exactly is this MI thing?
This post provides some of my thoughts pertaining to the monolithic inductor (MI) in 4 Parts:
· A brief overview of NSC's MI technology
· The advantages of MI technology over existing similar technologies
· Some market opportunities for the MI
· One significant market opportunity possibly being overlooked
A brief overview of NSC's MI technology
It is very encouraging to hear that preliminary testing of the MI prototypes have produced positive results. I can understand that testing technology of this nature is not straightforward and is not something, which can be completed in a short period of time. This is because of the size of this passive circuit element. This circuit element is microscopic in size and has a value of only a few nH (nano-Henrys). One nano is 10 to the power -9, or 1 billionth of a Henry (the unit of measure for inductance). (A one Henry inductor is rare. It would have to be about the size of an apple crate.)
I venture to guess these MI prototypes are much more complex than simple inductors or transformers on silicon because it would be extremely difficult if not impossible to measure and test these elements because of the inductance contributed by the bonding wires, leads, and even the test equipment itself. So it is highly likely these prototypes incorporate active RF circuits employing these MI elements.
So anyway, these monolithic inductors are really small inductors. This means the circuits containing these inductors or transformers are only effective for signals exceeding 1 GHz. The size or value of the inductor or transformer is inversely proportional to their effectiveness with circuit frequency. In other words the smaller the inductor or transformer (measured in this case in nano-Henrys) the higher the signal frequency must be for this inductor to be effective.
The advantages of MI technology over existing similar technologies
Basically the invention provides a means for implementing an inductor or transformer on a SI (silicon) substrate and not having to rely on GaAs substrate as in the past. The thing that is particularly unique is that the magnetic flux lines are directed away from the substrate, which increases the inductance. This is because air provides a better medium for the magnetic flux than does silicon. In other words, the magnetic flux through air is denser than through silicon, thus resulting in higher inductance. An additional advantage for this invention is fewer layers of magnetic material are required.
If two more of these MIs are laid out in silicon in very close proximity then a transformer is created. In this case the magnetic flux lines from these inductors cross couple.
Some market opportunities for the MI
I believe the potential for this invention may be somewhat underestimated. Yes, NSC's TunnelMOS and HBT probably represent greater revenue potential, but I believe the potential for the MI is not far behind these other technologies.
The reason for my stating this is the following:
All RF circuits incorporate inductors and transformers in order to do the job they do. These inductors and transformers are used to produce such circuits as filters, mixers, oscillators, and coupling being RF stages. The human population continues to seek means of communicating more information at faster rates. The maximum amount of digital or analog information, which can be transmitted within a given timeframe is referred to as bandwidth. Think of a long water pipe with water flowing through it. There is a practical limit to the maximum rate water can be pumped through it. As the diameter of this water pipe increases the practical limit for the rate of flow increases.. Similarly as the carrier frequency increase in RF communications the amount of digital and analog information, which can be transmitted increases. It is also important to note, the power requirements for "pumping" this information becomes less. This is why communications are becoming more common in the GHz (Giga Hertz or one billion cycles per second) range.
It is far less costly and much more efficient if these inductive elements can be placed in silicon rather than be external to the IC. The MI increases the possibility of this in these types of applications by at least three fold over existing technologies.
One significant market opportunity possibly being overlooked
Some may recall my idea of utilizing the MI to assist in reducing radiated emissions. In a previous post I came to the conclusion that this is not a possibility, partially because of some input I received from a PHD in this area, which had some influence on my thinking. In this PHD's opinion the silicon real estate requirements would be much too great to provide enough damping of the rise and fall times of the digital signal. This PHD sited he has been challenged with a similar problem relating to reducing transients in his audio amplifier designs.
I realized shortly after coming to my earlier conclusion that his problem of transient suppression in an audio amplifier design and the reduction of radiated emissions are two entirely different problems. Certainly in his application the MI could not be utilized to solve his design problem due to the signal power levels and lower frequencies.
Here again is the problem:
About 20 years ago the FCC became aware of the increasing problem related to all electronic appliances including computers, printers and all other computer peripherals, etc. As digital clock frequencies increase harmonic (multiples of the clock frequency) radiated emissions are produced. This is because of the fast rise and fall times of these digital frequencies. For you mathematicians, mathematically this can be expressed as a "Fourier Series." So, bottom line electronic appliances began to interfere with radio and TV communications. The FCC past a law strictly limiting the magnitude of the radiated emissions allowed for all electronic appliances.
This law adds significant cost to the design of any electronic system because it dictates:
· how the circuits are laid on the PCB
· the additional cost of a multi-layer PCB, providing for a ground plain
· the introduction of filters, such as ferrite beads, in PCB traces
· the introduction of metal shielding and expensive conductive coatings
· increased design costs due to additional engineering time
Much of these costs could be reduced if the rise and fall times of these digital signals were limited before they leave the IC. Let's take in case the PC today. The system clock frequencies are in excess of 500 MHz (the fundamental frequency). The first harmonic is 1 GHz and continues to multiply to a measurable frequency of greater than 10 GHz. It won't be long before system clock frequencies exceeding 1GHz are commonplace, so the problem will only get worse over time.
So, bottom line I believe my idea of incorporating the MI in the bonding pads of all ICs could serve as a viable solution towards taming these digital rise and fall times; hence reducing the costs associated with designing electronic systems
Rick
-----------------
*** Credit to RKMJ from ragingbull for this post
AI
