[time-nuts] Close-in phase noise (long reply)

[wa1zms at att.net]

John-

The BPFs in the 5 to 20MHz chain are just 7-pole LC
filters with the goal of trying to keep any other
harmonics other than the desired at least -50dBc.
Xtal filters would be the better choice, no doubt.

The -50dBc level is clearly not the best
that one could get, but was enough for an earlier
240GHz project. I just used the same OCXOs and early
stages of multipliers to get the latest system running
on 630GHz.

In the 241GHz system, I ended up building a direct
frequency synthesizer to get 110MHz from a 10MHz
drive signal.  At the time, the Freq West PLL
blocks I used wanted a VHF signal to drive
the sampling detector to phase lock the L-band
cavity VCO. The original Freq West units used 5th OT
xtals for the commercial applications.

By later experimentation, I found that
the same sampling detector would also work with a
much lower frequency reference and still lock the
loop. The risk however is that the PLL might lock
on the wrong harmonic of the reference (i.e.: value of N)
or can have higher reference spur levels since the
PLL was designed assuming a VHF reference and not an
HF reference frequency. But this is not a commercial
design project, and I can live with a difficult alignment
procedure or initial power-up PLL lock troubles.

But all this aside, my efforts are currently aimed at
best close-in noise within the first 1KHz of BW around
the carrier.

The PLL bricks all seem to have several kHz of loop BW,
so my close-in noise going from 20MHz to 1320MHz should
be only slightly worse than 20Log(n), with n=66 in my case.
But I'm not ruling out the chance of 1/f noise (or similar)
showing up from the sampling detector or some other yet-to-be
determined source.

However my focus is currently on the 5MHz to 20MHz portion
of the LO chain and to be sure the gain stages are not
running near compression. I do still agree with your
earlier comment about getting the most from that portion
of the chain.

-Brian, WA1ZMS

-----Original Message-----
From: time-nuts-bounces at febo.com [mailto:time-nuts-bounces at febo.com]On
Behalf Of John Miles
Sent: Tuesday, December 16, 2008 7:53 PM
To: Discussion of precise time and frequency measurement
Subject: Re: [time-nuts] Close-in phase noise question...more info...


> More info on the LO chain:
>
> 1) 5MHz Wenzel OCXO <--Custom Osc for me.
> 2) MSA-1105 buffer MMIC and lumped LPF
> 3) 5MHz to 10MHz 1N5711 diode based doubler
> 4) MSA-1105 buffer and lumped LPF
> 5) 10MHz to 20MHz 1N5711 diode based doubler
> 6) 20MHz BPF

What kind of BPF?  A really narrow crystal filter would be nice here.  (You
have basically reproduced the 8568A/B's 20 MHz reference section.)

> 7) 20MHz drives sampling detector inside surplus Frequency
> West PLL block to lock 1320MHz cavity oscillator.

Sounds OK as long as the sampler loop's noise floor doesn't limit you.  I
haven't measured the in-band residual floor of any bricks but I'd be
surprised if an SRD multiplier wouldn't be quieter.

> 8) 1320MHz drives Frequency West SRD multiplier to 6.6GHz.

If I wanted to get to several GHz with what's in my junk box right now, I
would do what you did to get to 20 MHz, BPF it with a multipole crystal
filter, and then use a few more multiplier stages to get somewhere between
100 MHz and 1 GHz, a la the 8662A reference section, depending on the choice
of the next stage.

That VHF drive signal would go into either an HP 33002A or 33004A SRD
multiplier, or one of the Picosecond NLTL multipliers (e.g.,
http://www.picosecond.com/product/product.asp?prod_id=109 ) I picked up in
their fire sale when they shut down their fab.

> 9) 6.6GHz to 39.6GHz Milliwave diode multiplier/amp/filter
> 10) 39.6GHz to 79.2GHz in varactor doubler
> 11) 79.2GHz to 158.4GHz in varactor doubler
> 12) 158.4GHz into x4 sub-harmonic mixer

AFAIK the rest of the chain is fine.  I'd focus on getting rid of the brick
PLL, or at least taking pains to make sure that it's not the problem, before
worrying about the MMICs in your early stages.

Remember that there's no point in optimizing the PN of any one stage much
below the input-referred residual noise of the following stage.  MMICs, in
saturation or not, are pretty quiet.  Quieter than sampler loops anyway.

-- john, KE5FX


_______________________________________________
time-nuts mailing list -- time-nuts at febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.