To the Editor I read with interest your article in the August issue of Microwaves & RF, "Crystal Oscillators Continue to Shrink," p. 33. As a designer of RF/microwave function modules for commercial and military applications, I think it is fair to say that I am not much different from your readers who are involved in higherlevel microwave designs (beyond the component level) and, as such, need to develop and produce the reference timing oscillator for our customers, or buy such oscillators, usually in the form of a crystal oscillator, from an outside source. I fall into the latter category, in which I must procure the crystal oscillators from outside vendors. Thus, the ultimate performance of the systems and subsystems that I design falls victim to the offerings of these oscillator vendors.

Of course, there are excellent suppliers of crystal oscillators, but experience feedback october 2008 n Microwaves & RF visit www.mwrf.com 13 is the best teacher in finding these suppliers. It is not just a matter of searching through ads or scouring through web sites to compare and contrast spec on crystal oscillators from different manufacturers. Phase noise is one of the most important specifications for a crystal oscillator (or what is called "jitter" when it is specified in the time domain). And most manufacturers try to show the phase-noise performance of their oscillators in the best light possible, which is to list the phase noise on the data sheet at the most favorable offset frequencies possible, which may not be meaningful to a system designer.

There is no standardization as far as listing phase noise at an agreed-upon number of offset frequencies, such as 1 kHz, 100 kHz, and 1 MHz. Instead, the measured offset frequencies vary from manufacturer to manufacturer and within the different product lines from a single manufacturer. In these cases, it is difficult to compare this one spec across different oscillators. The most meaningful way to understand the phase-noise performance of a crystal oscillator is by means of the plotted performance versus offset from the carrier, as measured by a phase-noise measurement system. And to hope that the phase-noise measurement system was properly calibrated and the tests were performed in a shielded room and under the same conditions as measurements made by other crystal oscillator manufacturers. As it is unlikely that the measurement conditions were the same across the industry, one can understand how difficult it is to compare crystal-oscillator products from different manufacturerscertainly not by comparing spec from different company web sites!

Alan Darcon,Design Engineer Configuration Analysis Management Yonkers, NY