eHam Forums => QRP => Topic started by: K5GUS on October 05, 2003, 09:17:19 AM

Title: 1/4 wave, 1/2 wave and random wire
Post by: K5GUS on October 05, 2003, 09:17:19 AM
Just getting into QRP and have been reading this forum to get lightweight antenna ideas.  I spend a lot of time in hotels across the country (and Asia) and plan on bringing a K1 or the new KX1 along for the ride.  I've read in these posts that a 1/2 wave radiator with 1/4 ground/counterpoise will work well.  When I read the Elecraft tips on using their internal tuners, it recommends staying away from 1/2 wavelengths of wire for random wire.  Why is this?  Is this only if your using random wire without a counterpoise?  Would a 1/2 wave wire with the 1/4 counterpoise be just fine with their tuner?  I currently have a small MFJ tuner (901B like) that currently works fairly well with a long wire and no ground with my IC706, but want to get into something lighter than my 706 and can use a wall wart or batteries instead of my 25a switcher.  I think my "neighbors" in my hotel will appreciate QRP power more than 100w anyways..

Thx and 73
de K5GUS

Title: 1/4 wave, 1/2 wave and random wire
Post by: WB2WIK on October 06, 2003, 11:43:42 AM
A half-wave, end-fed antenna is a good dimension to avoid because its feedpoint impedance is extremely high -- usually several thousand Ohms -- making it difficult to match, and almost definitely outside the range of any sort of auto-tuner.

Almost any other length is better.

"Random" is fine, as long as it randomly doesn't fall into the "half wavelength" category!


Title: 1/4 wave, 1/2 wave and random wire
Post by: K5GUS on October 06, 2003, 11:43:31 PM
The conflicting information I was talking about is this post:

where 1/2 wave is recommended with a 1/4 wave counterpoise or even without?! Your on the thread as well referring to a dipole.  So 1/2 wave isn't bad with a L or LC tuner, but doesn't work so good with autotuners that have limited range like the K1 or Kx1?  Or was the others posts incorrect?

As always, thanks for your input Steve.

Title: 1/4 wave, 1/2 wave and random wire
Post by: WB2WIK on October 07, 2003, 01:32:27 PM
In that older post, one respondant referenced a "1/2 wave wire with a 1/4 wave counterpoise" or something like that.  He said he'd make the wire 1/2-wavelength for the lowest frequency used, etc.

1/2-wave end-fed, with or without any quantity of radials or counterpoise, is one length that usually doesn't work, period -- as the feedpoint impedance is so high that any type of matching device used with it will be pretty lossy.  That's not to say it won't radiate -- of course it will.  In fact, a 1/2-wavelength of wire is technically *resonant*, meaning it is non-reactive.  Problem is, although it's non-reactive, its resistive component is extremely high.  Using virtually any other length wire is better, because it can be more efficiently matched.

The problem repeats every 1/2-wavelength, e.g., 1/2WL, 1WL, 1-1/2WL, 2WL, etc.  But not all ham bands are harmonically related.  So, if you install a 30 meter 1/2-wave wire (46' long), while that will be very difficult to match on 30 meters, it will match pretty easily on the other HF bands except probably 15 meters, where it will be 1WL and have a similar problem.

If you have zero feedline, which is often possible in portable installations, and can achieve a match between your transmitter and antenna, 100% of the energy from the transmitter will be radiated by the antenna, minus any loss in the matching network.  This is why "random wires" truly work just fine -- if you can match them.


Title: 1/4 wave, 1/2 wave and random wire
Post by: WB6BYU on October 08, 2003, 04:54:49 PM
The 1/2 wave end-fed wire is often recommended because
of the lower ground losses (certainly compared to a
quarter wire, where the ground radials are critical
to efficient operation.  I've used them, and they
certainly work, though I don't know if the performance
is that much better than some other alternatives.

First, if you just want something without much ground
loss, it doesn't have to be exactly half a wavelength
long.  Even cutting the wire to an exact length won't
guarantee that it is resonant when it is installed as
an antenna.  In the neighborhood of a half wavelength
the resistance and reactance change quickly with
frequency - and even more when the wire is several
wavelengths long.

Let's look at the matching for a moment:  an impedance
around 5000 ohms is easily matched with an "L" network
using two components with reactances around 500 ohms.
A standard "T" tuner used in this configuration will
simply have the output capacitor left at maximum
capacitance, where its reactance will be less than
the impedance of the antenna.  The main limitation is
the high Q and therefore large circulating currents
due to the high step-up ratio - this will increase
the losses in the tuner.  A well-made tuner, with lots
of space around the coil and good heavy leads, should
still be capable of reasonable efficiency.  The smaller
ones with the parts packed tightly into a metal case
may tend to run warm.  Another limitation is the minimum
capacitance available:  on 20m this will use 20pf, and
only 10pf on 10m.  Many tuners use capacitors that don't
go this low, especially considering all the stray

Another matching option is a parallel-tuned circuit.
Typically the wire is attached to the top of the coil
and the feedline is tapped between 10 and 20% of the
turns up the coil.  This is less critical of the exact
values of the components, as long as they are resonant.
A good capacitor may cover two or more bands with a
single coil, and adding parallel coils or capacitors
can extend it to other bands.  The same considerations
of rugged construction still apply.

Actually I was out last month with some other QRP folks
and we tried several antennas, including end-fed half
wave wires on the K1 and K2.  A brief summary:

1)  both the K1 and K2 built-in tuners would match a
40m half-wave wire with quarter wave counterpoise.
Usually.  Sometimes there were quirks.  (Supposedly one
of the tuners had been optimized for matching such

2)  another ham brought a 40m half-wave wire with a
home-made tuner.  The tuner had been adjusted at home,
and in this installation needed no further adjustments
to give a low SWR.  This seems to be a reasonable option
for quick setup (requiring just one wire), although the
relative efficiency of backpacking-sized tuners may not
be that high.

3) a full-wave end-fed wire for 40m outperformed either
of the half wave wires, probably because it was up
higher in the air.  (Both of the other wires were not
more than about 8' above the ground.)  I had no
problems matching it with an "L" net tuner.

4) A half-wave dipole up in the clear was at least as
good at the full-wave wire overall.

My suggestion would be to use a wire that is a bit
shorter or longer than an exact half wavelength
(depending on what works best with your tuner.)  It can
be matched directly from the K1 or K2, though there may
be times when setting a tuner outside on the wire and
running coax into the operating position may be more

Also, if you are using an external tuner, design it
for good efficiency on the wire you are going to use
it with and the bands that are most important to you.
This may mean a parallel-tuned circuit with plug-in
coils or capacitors to extend the tuning range.  A
plastic case is probably a good idea to reduce the
stray capacitance and eddy currents from coupling to
the coil.

Title: 1/4 wave, 1/2 wave and random wire
Post by: DK3QN on October 15, 2003, 06:21:02 PM
From practise I've found that lamda/2 wires provide much better results than lamda/4 wires if you are restricted on implementing a sound radial system.

Reason is that for the lamda/4 wires you definitely need a pretty good RF ground whereas for the lamda/2 a couple of wires (or even none) are just good enough.
As already pointed out by others earth losses are a significant issue with lamda/4 wires.

The issue of high feedpoint impedance with end-fed lamda/2 radiators can be handled by using a 'magnetic balun'. 'Magnetic balun' is the name of a broadband matching black box sold e.g. by WIMO, Germany.

In fact, there is very little 'magnetics' in the box.

The basic layout can be found here:

Comment: the number of actual turns is 8 (only 5 are shown). The 3 wires need not to be drilled but may be run in parallel.

This is all in German language but should be understandable from the drawing anyway.

If you DIY the cost of the whole device may be in the range of less than 15 US $. It's worth a try!

I used such a 'balun' during my stay on the Lofoten Islands this summer (IOTA EU-076) together with a 10m piece of wire and it worked out quite well on 14Mc and 18 Mc (abt. 5/8 lamda). I also had 2x 10m each radials laid out on the ground.

Using a 15m to 23m wire the device is said to provide good results on all HF bands starting from 80m (not tried yet). A matchbox may be needed here.

The WIMO version I have in use - although spec'd at 100 watts - can take a max of about 50 watts CW over longer periods of time. When running it at 100 watts CW the SWR deteriorates all of the sudden after some time which means the ferrite core saturates heavily.
So, watch out to choose a core big enough for your desired power level.

There is another variant of such a device to be found here:

The FT-50 core used here can definitely take only QRP power levels probably up to 10 watts or so.
I do not have any experience with this implementation.

It reminds me a lot of the so-called 'Fuchs-Kreis' ('fox circuit') which is reported in old German HAM radio magazines and has become more famous again recently in the German QRP scene. The 'Fuchs-Kreis' basically is a RF transformer (one band only or switched for multiband) with a coax fed low impedance coupling coil wound across a high impedance antenna coil. You can also see parallel C's making the transformer resonant at the operating frequency.

Because of the high voltages involved all of this is focussed around QRP power levels or up to a hundred watts or so.

Give it a try.

73, Klaus, DK3QN

Title: 1/4 wave, 1/2 wave and random wire
Post by: BIRDDOG on October 19, 2003, 05:29:41 AM
I'm sure Wayne was talking about a random wire only. There are several suggesting staying away from any multiple of 1/4 wave with a single random wire.

I've been reading around and seeing different opinions.

When using a 1/4 wave counterpoise, you are forcing the feed point to be at the low voltage high current point of the wave. So, under that thinking, a 1/2 wave wire should work good with a tuner.

Also, when you use a 1/4 wave counterpoise with a 1/2 wave wire. Count the total length of your antenna. Remember that the counterpoise is part of the antenna. If some don't like that thinking....simply turn the counterpoise directly 180 degrees extending out from the main wire...then they might like that count. What do you have now? Some will call it an off center fed dipole. Think about what it looks like now....doesn't it look like a Windom? Measures like a Windom. Windom's work good. They were using Windom's, fed with a single wire, for multiple bands, with no tuner, 70 years ago.

Also, if you're looking for something simple and easy for traveling. I read about a "W3EDP" antenna. It is simply two pieces of wire. The 85 foot piece goes to the hot socket on your QRP rig. The 17 foot piece goes to the ground on your QRP rig. Seen a few write ups on this. Seems to work pretty good. And it's easy.

Here's a page that lists a bunch of antenna discussions, including a few for the W3EDP.