[time-nuts] time-nuts Digest, Vol 162, Issue 11 Phase Noise of Oscillators
kmec at aol.com
Sun Jan 7 11:54:14 EST 2018
To the learned group:
The recent comment (below) stirred my memory.
While by no means an expert, I remember form my early career at Westinghouse Defense building low noise S-Band VCOs for the AWACS radar system that oscillator noise was key to the performance of the pulse doppler radar (MOPA type).
In order to achieve the low sideband noise needed to see close-in low velocity targets in background clutter, the VC oscillator was designed with very low noise techniques.
A high Q coaxial resonator was used (large aluminum cavity with post), a PC board was made out of tellite (a type of cross-linked polystrene board like rexolite, only copper clad for etching) which had very low loss, sapphire trimmers were used, an LNA type transistor run at VERY low current (bipolar for best noise at the time) was used as the oscillator, and everything was filtered and bypassed with the best available parts to eliminate noise from outside the can. The output of this oscillator device was very low, like -5 dBm, so a buffer amplifier was used to get the can output up to the desired level. When I asked the oscillator designer why not simply run the oscillator transistor at higher current to get more output, he said that they had investigated the noise performance using all possible approaches, and found the lower the current in the oscillator transistor, the better the noise performance. This was also why an LNA type of transistor was used, as the 1/f noise of the device is directly translated to the output carrier (up-converted to the microwave frequency, if you will). That was why they bought a lot of transistors and then selected the ones with the lowest 1/f noise for use in the oscillator position with the slightly higher noise ones used as the output buffer amplifier.
I have in my collection of gear a unit made by Quantron which is a 1/f meter for transistors (302?). HP also made one (4xxx? my memory is bad, even though I had one and sold it on eBay). Very important to the oscillator folks which is why I kept one around (never have used it). These meters would help evaluate the best oscillator candidates at baseband frequencies.
In a message dated 1/7/2018 9:54:30 AM Eastern Standard Time, time-nuts-request at febo.com writes:
Does any limiter, soft or hard, [and perhaps any nonlinearity of power
term 3 or greater in the amplifier of an oscillator] cause the "baseband
1/f noise to translate up to the resonator frequency [a form of
More information about the time-nuts