Generating 20 MHz from 5 or 10 MHz

The Ashtech Z-12T GPS timing receiver requires a stable 20 MHz frequency reference. My lab has only a 5 MHz H-maser reference and a 5 or 10 MHz Cesium reference so the goal is to generate a 20 MHz reference while best preserving the quality of the signal. The following is what I found out about frequency doublers and quadruplers. If you have comments or questions please let me know.


Generator 1 second stability 10 000 s stability
3325B 6e-13 1e-14
FS040 1e-13 6e-16

HP 10515A doubler


HP 10515A passive (diode bridge) frequency doubler



HP 3325B

Measured performance

3325B -- 5 MHz in vs. 20 MHz out

Spectral Dynamics FS040 quadrupler


Spectral Dynamics FS040-5 active frequency quadrupler

Measured performance

FS040 -- 5 MHz in vs. 20 MHz out

How these measurements were made

These measurements were made with a TSC 5110A time interval analyzer and an Oscilloquartz BVA quartz oscillator reference. The noise floor of the J1 vs. J2 output of the BVA and the noise floor of the channel A vs. channel B input of the 5110A is assumed to be better than the doubler/quadrupler being tested. Here is an actual measurement:

Noise test of OSA 8607 (output J1, J2) and TSC 5110A (channel A, B)


A pair of TSC 5110A Time Interval Analyzers


The OSA 8607 (inside a stainless steel thermal wine bottle holder)

In trade for the gum my assistant for this experiment was Nick(8):