The following essay I wrote is based on conversations I had with Dr. Paul Vella and Dr. Stephen Yang HOW A YIG OSCILLATOR WORKS - By David Alexander Straight A Ten Second Explanation. A YIG oscillator is a microwave oscillator who’s frequency is determined and controlled by a uniform magnetic field intensity. Typically, an electromagnet is used to provide this field and subsequently, a DC current can be used to control the oscillator’s frequency. The relationship between the current and frequency is very linear, making an excellent current-controlled oscillator. The Basic Principle of Operation Imagine that you had a compass pointing toward magnetic north. What would happen if you removed the glass enclosure, gaining access to the pointer, and then pinged the pointer just a little? You would be applying an impulse force on the end of the needle causing a slight angular rotation. If no other forces were acting on this needle then it would continue rotating forever. However, because this needle is actually a dipole in a magnetic field, a torque would be instantly be acting on the needle as soon as the impulse force becomes zero (think of the impulse as a Dirac function). Energy is stored in the needle, (ignoring non-conservative forces) and the needle would only diverge away from north just a little and then begin to return. Due to the inertia of the needle, it will continue swinging past north diverging the same amount, but in the opposite direction. This would go on forever if it weren’t for non-conservative forces being applied upon the needle. Now, note that this needle is actually resonating at some frequency (proof omitted) and also note that if the earth’s magnetic field intensity were stronger, then the needle would oscillate faster (the resultant forces on the needle would be greater). With this in mind we might be tempted to build an oscillator whose resonance is dependent upon a similar mechanism. Instead using a compass needle, unpaired electrons could be used. Where could we find these? Ordinary metal such as iron is a good source of free electrons, however there is a problem with using metal. It is lossy. Electrons moving within a conductive material will induce eddy currents, and therefore causing a low Q resonating object. What would help is some substance that is not conductive (or perhaps just a poor conductor). Ferrites happen to be a good source of electrons held in a poor-conducting substance. However, one thing remains quite important, and that is to have the electrons spaced about evenly within the substance. A material that has these properties is yttrium iron garnet (YIG), hence the name of the oscillator. Why do we want all these properties? The goal is to have all the electrons oscillating (wiggling) at the same frequency (rate) and only a uniform magnetic field within the substance will create this condition. If the field were not uniform then the electrons would be oscillating at different frequencies. Inducing a uniform magnetic field within the YIG substance, by external means makes the geometry of the YIG very important. Not all geometries of a substance yield a uniform magnetic field upon externally applied magnetic fields, take for example a transformer using a square shaped core. A good surface that maintains a uniform magnetic field is a very thin flat planar surface, however this shape is somewhat impractical considering the orientation that it must have to operate properly. A sphere shape is perfect for applying a uniform magnetic field (for a proof start by using infinitely thin spherical surfaces concentric to the center of the sphere). YIG spheres are predominant, because they have a uniform magnetic field on the inside upon application if a uniform external field so that all unpaired electrons see the same intensity and therefore will resonate at the same frequency. Now that a suitable tank circuit is realized all that is needed is something to keep the tank resonating (again, due to non-conservative forces even the YIG will quit oscillating without energy being added). From an electronics point of view the YIG sphere could be modeled as a parallel lumped element RLC circuit. The resistance is representative of the losses. If you take some lossy tank circuit and place a negative resistance in parallel with it then you will effectively remove the net loss. This is realizable by using a transistor or FET. Using impedance models one can see how a transistor could be set up as a negative resistance. Looking into the emitter of a transistor the impedance can be modeled as 1/((re + Zb)*B) where re is the internal junction resistance, Zb is some impedance in the base circuit and B is the beta of the transistor. Note that B is function of frequency and as frequency increases beta decreases. As beta decreases the impedance looking into the emitter increases and thus this looks just like negative resistance. Energy is coupled to the YIG electromagnetically using a very small inductor (on the order of Nano-Henrys) from the transistor. So that the YIG oscillates at some desired frequency, an electromagnet is used to apply a constant and uniform magnetic field to the sphere. These are the basics of how a YIG oscillator works For more information about YIG Spheres see Microwave Filters, Impedance-Matching Networks, and Coupling Structures by G. Matthaei, L. Young, and E.M.T. Jones Microwave Solid State Circuit Design by Inder Bahl and Prakash Bhartia For information on solid state microwave oscillators see "A Unified Approach to the Design of Wide-Band Microwave Solid-State Oscillators" IEEE Transactions On Microwave Theory and Techniques, Vol. MTT-27, No. 5, May 1979
Click here for a quick look at one photo if a YIG sphere 33k
If you have any questions contactme at dstraigh@polymail.calpoly.edu