[time-nuts] Why using HP5370 ext-ref is (maybe) a bad idea

Poul-Henning Kamp phk at phk.freebsd.dk
Mon Mar 3 02:26:55 EST 2014

In message <E7A494F6-78F1-4568-8BD3-D94A32DEBB8B at rtty.us>, Bob Camp writes:

>Do you have any idea how 'clean' your 200 MHz signal is? The manual 
>suggests getting it to -65 dbc for subs using a spectrum analyzer. That's
>pretty far down. I seem to recall the adjustment process being a bit tedious
>(lots of back and forth). 

My estimate is that all harmonics are at least -50 and most probably
-60 down.  The 10MHz is probably the worst.

My Lab is not really set up for RF work, so this is probably an
area where one of the hams could do lot more competent job than me.

I've attached a plot of one of the runs yesterday, beause things
are more complicated than I initially thought.

The vertical bars are AVG +/- STDDEV of 1000 sample TI on a 10MHz
from my HP3336 in start-com mode.

The X-axis is the phase offset set on the HP3336 in degrees, and
represents the phase difference between the ext-ref and start+stop
signals on the HP5370 plus some arbitrary offset due to cables etc.

Obviously, the phase difference has no systematic meaning for the red
bars, since it is free running on the OCXO at some frequency offset
from the input signal.

Yet, it is quite evident that there still is a periodicity in the
data, which peaks around 0, 5, 10 and 15 degrees.

The green bars however...

The initial artifact I noticed when I just plotted the STDDEV is
still there, around 9 degrees where both the average and the stddev
take a hop.

But that blib is peanuts relative to the 3-degree periodicity
for which I have absolutely no explanation, and the equally
evident 18-degree periodicity.

The 3-degree periodicity cannot be a simple harmonic, it it were
it would be a 1.2 GHz signal.  (360/3 * 10 MHz = 1200 MHz)

But what then ?

As in a canonical scientific paper, I have to conclude that more
research is clearly needed, and I'd really love to see what results
other people might get.

In the meantime, run you 5370 on internal clock, and rely on the
law of big numbers.


Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: intext.png
Type: image/png
Size: 20720 bytes
Desc: intext.png
URL: <http://www.febo.com/pipermail/time-nuts/attachments/20140303/9c50ed17/attachment-0001.png>

More information about the time-nuts mailing list