Re: Specs on microwave power divider/combiners

On Aug 4, 9:42 pm, Jeff Liebermann <je...@xxxxxxxxxx> wrote:
On Mon, 4 Aug 2008 13:53:19 -0700, "Joel Koltner"



<zapwireDASHgro...@xxxxxxxxx> wrote:
I'm thinking that something like this:
http://www.mini-circuits.com/pdfs/ZN2PD-9G+.pdf...is a multi-stage Wilkinson
divider internally.  I'm curious how they decide to specify its frequency
range... anyone know?  Based on this data *** and others I've looked at, I'm
thinking that the low-end (1.7GHz) is spec'd based on the isolation between
the split/combined ports becoming "unreasonably" low (20dB, or thereabout),
whereas the high-end (9GHz) is based on the insertoin loss becoming
"unreasonably" high (~1dB over theoretical minimuum).

If you're just using a Wilkinson like this as a splitter and have a very good
match on the output ports, presumably you can use it to much lower
frequencies?  At DC, for a 50-ohm 2-way Wilkinson, the input is looking at 25
ohms (=VSWR of 3:1=return loss of -9.5dB, still acceptable in some cases) and
(since at this point it's just a resistive splitter, really) the overall
insertion loss is also 9.5dB or 6.5dB over the minimum.

---Joel

Your logic seems reasonable to me.  

I'll add that it's possible that with a specified maximum input power
of 10 watts, it won't take much insertion loss to hit the specified
125 mw internal dissipation limit.  Methinks something is fishy here
as that's -0.06 dB maximum internal loss.  Maybe they mean 125 mw over
the specified 0.5dB insertion loss.  However, 0.5dB loss with a 10
watt input, is 1.09 watts dissipated, which is quite a bit more than
the specified 125 mw internal dissipation.  Nice muddle here.

Some more clues can be found in the Power Splitter FAQ at:
<http://www.mini-circuits.com/pages/pdfs/pwr2-4.pdf>  (Sept 8, 1999)
  Q. Can performance parameters be traded off when specifying
  power splitter/combiners?

  A. Once frequency range has been designated, there is not
  too much leeway in tradeoffs between the various spec
  parameters. Usually all the performance parameters of
  a well designed power splitter simultaneously fall off
  as a function of frequency. However, near the band edge,
  isolation may be improved at the expense of VSWR by
  adjusting the value of the internal load resistor Zo.

There are other performance tradeoffs hinted further down the FAQ.

Not very specific, but enough to tell that MCL apparently specifies
the frequency range first, and then tweaked the performance parameters
to fit.  Whether you consider 3:1 input VSWR and/or 5dB loss to be
acceptable performance, is your problem.

I was hoping to find a test circuit and procedure, that might give a
better answer, but I couldn't find anything.

--
Jeff Liebermann     je...@xxxxxxxxxx
150 Felker St #D    http://www.LearnByDestroying.com
Santa Cruz CA 95060http://802.11junk.com
Skype: JeffLiebermann     AE6KS    831-336-2558

http://www.lazygranch.com/images/temp/mea802_4_2p5.jpg
FWIW, this is the inside of a MECA 802-4.2.5 power divider.
.