From bec@haystack.mit.edu Mon Oct 25 21:23:23 MET 1999
Received: from dopey.haystack.edu (dopey.haystack.edu [192.52.61.54]) by picasso.geod.uni-bonn.de with ESMTP (8.8.6 (PHNE_14041)/8.8.8) id VAA13015; Mon, 25 Oct 1999 21:23:22 +0200 (METDST)
Received: (from bec@localhost)
	by dopey.haystack.edu (8.8.6 (sendmail_886_v2)/8.8.6) id PAA21049;
	Mon, 25 Oct 1999 15:20:40 -0400 (EDT)
From: Brian Corey <bec@haystack.mit.edu>
Message-Id: <199910251920.PAA21049@dopey.haystack.edu>
Subject: Re: YEBES and question marks
To: petrov@picasso.geod.uni-bonn.de (Leonid Petrov)
Date: Mon, 25 Oct 1999 15:20:39 EDT
Cc: aen@haystack.mit.edu, baa@casa.usno.navy.mil, dgg@aquila.gsfc.nasa.gov,
        hase@wettzell.ifag.de, mueskens@mpifr-bonn.mpg.de,
        nothnage@picasso.geod.uni-bonn.de, sorgente@hp138.mpifr-bonn.mpg.de,
        tme@cygx3.usno.navy.mil, vicente@cay.oan.es, vlbi@oan.es,
        weh@ivscc.gsfc.nasa.gov
In-Reply-To: <199910251901.VAA12884@picasso.geod.uni-bonn.de>; from "Leonid Petrov" at Oct 25, 99 9:01 pm
X-Mailer: Elm [revision: 109.18]
Status: R

Dear Leonid,

>   Now I understand your argumentation. It sounds convincing. Well. Let take
> that a bad distributor spoils only phase cal. What should YEBES folk undertake
> in this situation? What should they do in a short-period perspective, what
> should they do in a long time perspective? What can they do before the test
> experiment? Do we need a test experiment then? If yes, what should we measure?

Wow -- that's a lot of questions!  I'm going to refrain from a detailed reply
because I'd first like to see some confirmation of the theory, namely, an
improvement in the delay and rate scatter by fringing with manual phases.
If manual phases don't help, then everything I've been saying is off the mark,
and you should ignore me.

*If* manual phases do help, then...
 short term: avoid using the 5 MHz distributor for phase cal as well (are
               additional outputs available right at the maser?), or
               thermally insulate the distributor from its surroundings, or
               move the distributor to a more thermally stable location.
 long term:  same as short term, plus, if the distributor must be used,
               investigate its temperature sensitivity.  From the data
               available, it appears the sensitivity is of order 100 ps/deg C,
               which is extremely high.  Paul Burgess warned about buffers
               that use tuned amplifiers.  It could be that the circuit is
               intrinsically sensitive and should be replaced with a
               broadband, more stable design. 

> Do we need a test experiment then?

What does the manual fringing say?

 --Brian