[time-nuts] Note on early VNA's

[Kmec at aol.com]

To the learned audience:
 
I agree that the 8410 is an excellent place to start to learn  about VNA 
architecture and issues. I, as well, learned the HP 8410, first  as manual 
system, then we did the automation ourselves (Westinghouse was too  cheap to 
buy a "bundled" system) using a 9825 desktop calculator, from the HP  app note 
of "Semi-automated Network Analysis". My education in this regard was  
superb!  My industrious Vietnamese grad student is learning the issues from  my 
8410 books and he is building a more modern version. 
However, a slight clarification about the early VNA's: They were not YIG  
based, that came later. The classic HP8542A system was a BWO based system 
with  the 8690 Sweeper and a bunch of plug-in drawers that worked with a Signal 
 Multiplexer to yield a 1-18 GHz system. A 5100 based synthesizer was used 
to  lock the harmonic converter to eliminate harmonic "skip" cal errors 
(Brooke  alluded to this) and the whole mess was driven by an HP 1000 
mini-computer and  had a reel to reel tape drive for mass storage! It was huge ( a 
three bay rack)  and cost $250K in 1968 money (a lot today, $1.5 M ??).  I had 
one of these  systems and still have parts of it! I automated my own newer 
8410 system in 1986  when I started my consulting company and used the 12 
term error model software  pak (HP11863??) in RM Basic on a 9826 computer (big 
step up!). While I do not  recommend this approach for anyone today, the old 
literature provides great  insight into the issues, where the errors come 
from and so on, as HP had figured  all this stuff out. It is a shame that 
shipping to Europe is so high, as a lot  of these systems and components are 
still around (I have six 8410 systems  still !). A mainframe is under $50.00 
and I bought a working 8411A converter for  $20.00 at Dayton this year (dont 
ask why, I guess it was too cheap....).
  I had just about every variant of this stuff, VLF through 40 GHz.  Man, 
HP engineering was tops in those days!
Still, I think a very dedicated homebrewer could build his own design for a 
 < 3GHz VNA from adapted "wireless" parts, but I am too lazy for that. I 
much  prefer hacking some proven hardware into what I need.
 
73
Jeff Kruth
 
 
In a message dated 6/2/2014 5:47:17 P.M. Eastern Daylight Time,  
time-nuts-request at febo.com writes:

Message:  7
Date: Mon, 02 Jun 2014 14:36:09 -0700
From: "Richard (Rick) Karlquist"  <richard at karlquist.com>
To: Discussion of precise time and frequency  measurement
<time-nuts at febo.com>
Subject: Re:  [time-nuts] VNA design
Message-ID:  <538CEE49.6000901 at karlquist.com>
Content-Type: text/plain;  charset=ISO-8859-1; format=flowed

On 6/2/2014 12:41 PM, Brooke Clarke  wrote:
> Hi:
>
> I started with the HP 8410 and added an  external computer.
> Since it can be used manually I think it's an  excellent way to learn
> about VNAs.
>  http://www.prc68.com/I/MWTE.shtml#NA
>

For my last 8 years at  Agilent before retiring in March, I
was doing advanced R&D on network  analyzers.  The newer
guys coming up didn't have an intuitive  understanding of
network analyzer architectures like I did. I
started  using the 8410 back in 1973 before I even worked
for HP.  Because of  the modular design, it was like a
teaching tool that forced you to  understand what was
going on.  When I mentored the young guys, I  would
explain to them a lot of principles based on the 8410.
Modern  network analyzers are too "automatic".
The 8410 puts modern VNA's into  perspective.  BTW, I used
to sit next to Dick Lee, who was a member of  the 8410
design team in 1963 at the dawn of the golden age of
microwave  instruments based on YIG tuned oscillators
and step recovery diode  samplers.

As you noted, the architecture was built around the YIG tuned  oscillator 
and certain things were done that way they were because of  that.

Rick Karlquist  N6RK