[TowerTalk] VNA Antenna Analyzer

[Jim Lux]

On 10/29/15 5:00 PM, Wayne Kline wrote:
> Fellow TT'ers
>
>    I bought a  VNA type antenna analyzer .  Just scratching the surface to it' s capabilities.  One thing I know I need to  build a set OSL Open Short Load
> termination calibrator fittings.   Spent some time on Google  SMA  homebrew   are easy to find... ( f I had bought a SARK 100  which  has an SMA connector I would be golden )    But  I  need info for an   UHF  or  N connector .
>
>      Not wanting to reinvent the wheel   anybody  have info on  such animals ?
>
>

Are you looking to homebrew?
Or buy something already made?

Anritsu, for instance, has a OSL 3-in-one
http://configure.us.anritsu.com/catalog/itemView.jsp?siId=OSLN50-1
but at $525, it's probably as expensive as your new VNA.

I don't know that I've ever seen a UHF PL-259/SO-239 cal set.


If you're just going to be doing it at HF frequencies, then you can 
probably cobble up a short or open pretty easily with an appropriate 
gender connector.  A small metal disc works nicely. for the open, the 
usual practice is to make sure that the center conductor ends, and the 
opening for the shield is completely blocked off.

If you have some N terminations, you could probably modify them to 
remove the 50 ohm resistor and replace it with a short or nothing.

Ultimately, if you're interested in precise measurements, you'd need to 
know the distance from the connector mating plane to the 
open/short/load, but at HF, with a connector that is an inch long, I 
doubt the correction is significant.

Or, if you know someone who has a good VNA that's in cal, you could 
measure your homebrew units.

There are also schemes to use stacks of adapters.. Cal with your SMA cal 
set, then get two SMA:N adapters that can mate. Mate them, measure the 
distance through the adapters.  Now you know the sum of the two adapters.
If you have all the various genders, you can measure a lot of different 
pairs, and do the three cornered (or 10 cornered) hat thing.  It's 
basically a solving sets of linear equations kind of thing.

Measurement 1 = A+B
Measurement 2 = B+C
Measurement 3 = A+C

So M1+M2 = A+2B+C
M1+M2- M3 = 2B, so now you know B

If you do it in complex coordinates, it's multiplying

M1 = A*B
M2 = B*C
M3 = A*C

so M1*M3/M2 = A*A

I'd not worry about the 0.36 degree phase shift from a 1cm error at 30 MHz.

and so forth.


You could also rig up a sliding short or sliding load, and do that style 
calibration.

What's really nice is if you get a few relays, you can build a little 
switch box that can do all the shorts/opens/loads.  You find someone who 
has a VNA to measure your switch box, and now you've got an electronic 
cal set.