A. Isolation

Study for isolation between RF blocks can be divided into two stages: In the 1950’s to the 1980’s before the RFIC development, studies were focused on the shielding. As long as an electronic equipment contained RF blocks, a metallic box made by material with high conductivity usually covers each RF block for the good shielding. It is very heavily involved the mechanical works and the cost of the shielding boxes is sometimes more expensive than the device and all of the parts in the circuits. However, nobody dare to think about the implementing of the RF blocks with integrated circuit technology. On the other hand, in the implementing of the digital integrated circuit, the dramatic reduction of cost was very attractive and encouraging. In the 1980’s, the scientists and engineers started to integrate RF circuit using the integrated technology. It was, of course, a very brave action. Up to 1995, the experiment for isolation had been achieved with a great step forward. The isolation between blocks approached 40-50dB. RFIC technology was becoming realistic and mass production began.

There are three kinds of isolation. The first is the isolation between RF blocks. The second is the isolation between digital blocks. The third one is the isolation between RF and digital blocks. The first two have similar difficult points and have the same issue of the solution with the exception of the case when the PA block, which must be specially treated, is involved. The third one is somewhat more complicated. Normally the power of the RF block is much higher than that of the digital block since the current in the RF block is in the order of mA while the current in the digital block is uA. From the power viewpoint it looks like that a digital block is easier to be disturbed than the RF block. From the frequency viewpoint, an RF block is easier to be disturbed than a digital block because the signal in a digital block is basically a pulse. A narrow pulse is a wide band signal. It contains high frequencies, including RF components even its digital frequency is lower than the RF. Therefore, the third isolation is somehow difficult than others.

A question might be raised up to a sensitive reader: how many dB of isolation is enough for the practical purpose? In the wireless communication system, the ideal isolation level should be around 130dB, which is 10 dB higher than the total gain of the useful signal from antenna to the data output terminal. When the un-isolated signal goes through the entire communication system and gets a maximum of gain, 120dB, it is still at a level under the sensitivity of 10dB. Then, the communication system never find its existence. As mentioned above, when the isolation between blocks is achieved up to 50-60dB, it makes RFIC becomes realistic and available to applied into a system, say, communication system. However, it is still far from the ideal goal and the research of isolation must be continued.





B. High Q Inductor for IC

The project of high Q IC inductor is the key growing point in the future RFIC as well as SOC development.

At present, the Q value of an IC inductor at RF range is about 5 to 10 in al l of IC processes. It is much poorer than that of a chip of discrete inductor, in which the Q value of inductor is about 80 to 150. The low Q value of spiral inductors brings about the additional noise to the RF blocks, and it is hard to be applied in the implementation of a filter or an oscillator.

The second problem is that a large die area must be provided to the spiral inductor. In order to reduce the cost and to simplify the hardware works, most IC inductors are constructed on the chip. The die area for each spiral inductor is about 10 to 100 times larger than the area of the device or other parts. Looking at any of existing RFIC layouts, about 60% to 80% of the area is occupied by the IC spiral inductors. The IC spiral inductor becomes the main factor of the total IC cost.