Designing Directional Couplers for Broadband HF-UHF

[xez]

Hi all,

Has anyone got any good tips on broadband (HF->UHF)
somewhat physically compact directional coupler
SWR / directional watt-meter designs?

I'm also looking for equations or quantitative
design parameters relating to their design and operation.

Two use cases are of interest:

1) Specifically for weakly coupled power / SWR
probe operation, I'd like somewhere very generally
in the region of 1% to 0.001% coupled power.
The probe coupler should not to impose much load or
power limitation on the primary line which may be
operating with 0.001W to 1000W kinds of powers, though
0.1 to 10W would be most commonly of use / interest.

I'd like to cover all of HF, VHF, and UHF with one
simple design *if* that's practical, though having one for
each band wouldn't be a bad compromise if it greatly
simplifies or improves the design.

2) Low insertion loss, highly directive circulator / isolator
functions at modestly high power levels (e.g. 0 to 1kW),
and good operational characteristics over a single
specific narrow band of operation. The cases of HF,
VHF, and UHF are most relevant since other techniques
are applicable above or below those ranges.


For (1), the weakly coupled power probe, I've seen
designs using just a ~1/100th wavelength probe
alongside a transmission line. I haven't references
for the coupling ratios achieved with such a small probe
line length, but I'd guess it's in the 0.1% range in
order for it to be useful for its purpose. I question the
directivity these are achieving, and would welcome some
quantitative information on the subject.
I wonder how much smaller than 1/100th of a wave long
one could make the coupling probe line and still get
good directivity and achieve around 0.1% coupling ratio?
Has anyone any references relating to design types or
design performance parameters?
I understand that 1/4 wavelength is the optimum length
for a secondary directionally coupled line, though in
that case I'd expect there'd be a very high percentage
of directionally coupled power, well in excess of what
I'd need for a measurement.

For (1), the weakly coupled power probe, I've seen
designs using ferrite coupled transformers for the HF
frqeuencies that are surely much smaller than
1/100th wavelength in physical size,
though I am not aware of specific details of their performances or maximum achievable bandwidths.
I'm inclined to be skeptical as to whether such
a ferrite transformer design could work well for
HF + VHF + UHF, or which could easily work with
higher power levels at any frequency.

 

[xez]

Looks like I found a few related references to the
questions I had:

http://members.tripod.com/michaelgellis/direct.html
http://n2pk.com/RLPmtr/RLPmeter.html
http://n2pk.com/VNA/VNAarch.html
http://www.telepostinc.com/Files/phipps-1.pdf
http://www.arrl.org/tis/info/pdf/5904024.pdf
http://www.noding.com/la8ak/m12.htm
http://www.eham.net/articles/1024
et.al.

It seems like most broadband designs fall off
either on the lower end of HF and/or the upper end
of UHF, but they seem to be able to function decently
well in simple / compact forms given how broadband
and tolerant of power levels they actually can be.

 

[berkeman]

I did a search at the ARRL.org website, and got some good hits, but the better ones are in the "members only" area -- sorry about that. Have you looked in the ARRL Handbook? It's got some reasonable circuits for directional couplers and SWR measurements. Your local technical library should have copies of the ARRL Handbook.

http://www.arrl.org/search/?exp=1&q=swr

 

[xez]

Thanks for the help.

I've got an old copy of the book which is
(surprisingly) mostly useless on the topic,
but I'm glad to hear that the new ones are better.

I'll look through a newer one and see what they have to
say.

The links I turned up show that most of the topologies
in common use since ~ 1959 are either the
coupled line probe ones or the ferrite transformer
ones I was somewhat familiar with, though they've
done some impressive jobs at making the ferrite
transformer ones more broadband in a couple of cases.

Given the broadband 'black box' module units I've seen
I thought perhaps I was missing some key topology that
might be totally different and better than the others.