By James C. Rautio
Spiral inductors and silicon … from an RF point of view it seems silly. After all, we have all that lossy silicon just microns away from our loss-sensitive inductor. We need low loss or our VCO won’t start, or our LNA will have a high noise figure. Put that inductor anywhere but on the silicon!
But on the silicon it must go. If we are to enjoy our tiny little smartphones with half a dozen or more radios in one tiny little hand held package (which is half battery to start with), and to do so at a reasonable price, the inductors have to go on the silicon. This makes loss in the silicon our enemy. Before we can attack our enemy, we must first understand our enemy.
It seems to be common knowledge these days. Si RFIC inductor loss is largely, or at least partly, due to inductive coupling with the substrate. Imagine a planar spiral inductor positioned a short distance above a perfectly conducting sheet. (Note: You can call the sheet “ground” if it makes you feel better, but on an RFIC, that is very nearly a meaningless term.)