[Elecraft] K2 Antenna Tuner Question

[Julian (G4ILO)]

Fine, George. I'm probably not going to lose too much sleep over the losses
in 2.5m of RG-213. It might be possible to reduce it a bit by using 300 ohm
ribbon. The trouble is that the feeder has to share some plastic ducting
with another length of 213 for the 6-meter antenna. I don't have any
experience of using twin lead but I seem to recall that for it to work
correctly and preserve balance, other conductors need to be kept away from
it.

At present, the KAT-2 will find a match with my antenna on all bands,
except 40m. It produces a 1.9:1 SWR on 40m, but I noticed that it clatters
away whilst reading 2.0 or 1.9 all the time, which suggests to me that
nothing it tries makes much difference to the matching. My loop is roughly
a half wave in circumference at 40m so its impedance at that frequency must
be almost infinite. Interestingly, the SG-239 at the feed point did seem to
match it well enough that I was able to make a few contacts with it. I
think the solution to this will be to insert a 40m trap into the opposite
side to open circuit the loop at 7MHz, turning it into a halo.

The other thought is whether, given that I'm trying to match a wide varying
range of impedances, I would get better results using a current balun. I
think the balun I'm using is a voltage balun although the constructional
details didn't say. It is two x 14 turns of wire, bifilar wound on an
FT-125-61 core, with one pair of opposite ends of the two windings connected
together and to the unbalanced shield, the other two ends to the balanced
terminals, and a link from one balanced terminal to the co-axial center.

I've been studying the N0SS/W1CG design and wondering if there would be
anything to be gained from rewiring it like that, but as the balun is now
in a rather inaccessible location I'm also experiencing a strong sense of 
"if it isn't broken, why fix it?"

73,
--
Julian, G4ILO. (RSGB, ARRL, G-QRP, K2 #392)
G4ILO's Shack: http://www.qsl.net/g4ilo



  ==== Original message ====

> > George,
> >
> > Perhaps what you don't realise is that the KAT2 can be turned off
> from the
> > menu so that it is effectively out of circuit.
> 
> Actually, Julian, I am quite familiar with the KAT2 circuit and its
> use. I have done a lot of circuit modelling around the KAT2 component
> values to investigate the suitability of the tuner for various
> feedpoint impedances.
> 
> The point of the suggested
> > exercise was to see how well the SG-239 matches the antenna compared
> with
> > the KAT2, just for curiosity, since the 239 appeared to be quite
> lossy when
> > installed at the feedpoint.
> 
> Thank you for making that clear - I failed to understand how the units
> were to be hooked up. You are entirely correct that by removing the
> KAT2, one could evaluate the SGC in its stead. I haven't looked into
> the SGC schematic so I am unfamiliar with its component values,
> relative to the KAT2.
> >
> > Personally, I suspect that any benefit in using an "optimum feedline
> > length" (presumably with the aim of presenting a matchable SWR at
> the TX
> > end)
> 
> Actually a "matchable line input impedance at the TX end . . ."
> 
> might be negated by the extra losses incurred by using a longer than
> > necessary length of co-ax. I have some charts on my web page that
> > demonstrate just how high co-axial feeder losses can be when the
> antenna is
> > a long way from 50 ohms. I claim no credit for these charts by the
> way:
> > they were produced using an Excel spreadsheet that was mentioned in
> an
> > article in RadCom about a year back.
> 
> The TLDetails program is outstanding in its capability for doing
> t-line analysis and modelling. I believe that AC6LA is the author - it
> is part of the XLZIZL spreadsheet system that he offers. It is also
> available as a free-standing executable program. Check his website for
> downloads.
> 
> Actually, our group has given up on coax for any field antenna
> operations where multiband performance is required. As you point out,
> regardless of how well the K2 can match the line, if 50% of the power
> or more goes to heat up the coax, it is academic. And that is not an
> unrealistic figure, even with RG-58 or 8X, with some antennas on some
> bands that we have modelled and later tested.
> 
> On the other hand, either 300 ohm TV twinlead - for the lightest
> weight applications - or regular 450-ohm ladderline show remarkably
> small loss even at SWR>10 so one can feel free to "fiddle" with the
> line length (as long as things remain reasonable!) to facilitate
> matching a difficult antenna setup.
> 
> Our general approach is to use EZNEC 3.0 to model the antenna and
> estimate its driving point impedance on the bands of interest. The we
> use the TLDetails program to transform the antenna load Z to the line
> input Z. It is here that we can evaluate the effect of different line
> types, lengths, etc.
> 
> The Reg Edwards' (G4FGQ) program "LTUNER" designs the optimum or exact
> L tuner for a given load impedance to provide a specified input
> impedance, typically 50+j0 ohms. Those component values can be
> compared against those afforded by the tuner involved to estimate the
> liklihood of the tuner performing the match.
> 
> The effects of not using the correct values can be evaluated by a
> program that I have written which determines the input impedance of an
> arbitrary L or PI network presented with a given load. It is designed
> for rapid component value changes so that one can quickly arrive at
> the best compromise between what the tuner can do and what it should
> do to most nearly match the load. If you would like, I will be happy
> to email you an executable of it. It is called NDEXPL3.EXE . . . Reg's
> program is available from his website with which I assume you are
> familiar.
> 
> The KAT1 is the hard nut to crack. With its limited number of L and C
> combinations and the values of the components, its impedance matching
> range is much smaller than that of the KAT2 both in terms of resistive
> and reactive components.
> 
> Interesting business!
> 
> 73/72, George
> Amateur Radio W5YR -  the Yellow Rose of Texas
> In the 57th year and it just keeps getting better!
> Fairview, TX 30 mi NE of Dallas in Collin county EM13qe
> K2 #489  IC-765 #2349 IC-756 PRO  #2121 IC-756 PRO2 #3235
> 
> 
>