Re: End Fed antennas w/ uBITX #ubitx

Jerry Gaffke

I did that back-to-back test of two matchboxes.
I'd say they are plenty efficient for my purposes from 3.5 to 30 mhz.

My 2 year old matchbox is referred to here as xfmrA.
Text on the box says EFHW-8010, should be exactly like the one reviewed in the March 2016 issue of QST.
I popped this one open long ago.
It's pretty much as described on that facebook page where they are trying to build clones.
Got two stacked toroids, each has an OD of 2.43", ID of 1.40", height of about 0.5" each.
Has 14 turns of enamel copper wire 0.072" diameter, doubled up on the first two turns
for the primary 50 ohm connection.  A couple caps in series across the 50 ohm port,
the facebook page seems to think the series combination is 100pf.
The SO239 port and the grounding post are adjacent at the bottom, have their 
grounding tabs soldered together for a short and hefty connection.  
That mechanical support eyebolt at the top has no internal electrical connection.
Facebook page thinks the toroids are Mouser PN 623-5943003801,
which is at least the right size, no idea if that's the right material.
But before trying to roll your own, look over that facebook page.
Apparently not as trivial as you might think to hit multiple bands like this.

Also borrowed a newer matchbox, about 2 months old, label says MEF-330-1K.
This was bought as a separate matchbox, not part of an antenna system.
I believe the only significant difference between the two is a vent to the atmosphere
on the side of this new one, not present on the old one.
Referred to below as xfmrB

Here's the raw data:

          -----2500-ohms-----     ----50ohm-load-----------------
           xfmrA       xfmrB       ohms-in       vout-for-50v-vin
 3.80mhz   45.6+j3.2   45.4+j5.9   38.3+j11.5    48v 
 7.20mhz   41.7-j5.5   45.3-j4.3   37.4+j1.6     52v 
14.25mhz   31.8-j3.9   33.5-j4.3   31.8+j4.7     60v 
21.35mhz   33.6+j6.9   35.4+j6.4   42.3+j12.7    48v 
29.00mhz   67.0+j12.9  65.3+j14.0  63.0+j1.5     38v 

First column shows input impedance of xfmrA when output has 2500 ohms across it.
Second column shows the same for xfmrB
Third column shows impedance into the xfmrA coax port, xfmrA is hooked back to back with xfmrB
using wires of 1 or 2 inches, and xfmrB has a 50 ohm dummy load on the coax port.
Fourth column shows peak voltage across the dummy load of the above back to back arrangement
when xfmrA is driven with 50v peak of RF.

Impedances were measured using the AQRP Vector Impedance Analyzer from Kees.

The voltage measurements were made with a Rigol DS1052E using a single 10x probe.
The probe was first used to scale the RF source for 50v at the input port,
then the probe was moved to measure the voltage across the dummy load.
I don't have a lot of faith in this Rigol, but should be fine for such relative measurements. 

I chose a 2500 ohm load for the matchboxes when measuring their impedance,
as this should give close to a 50 ohm input impedance given the 1:7 turns ratio of the toroids.
The measured impedance was somewhat less in all cases except 30mhz.
So for some reason the transformer seems to exhibit an impedance ratio in excess of 1:49
at most frequencies, and thus gives a good match for an end fed halfwave which typically
has an impedance somewhat greater than 2500 ohms. 

Given that the input impedance was mostly lower than the 50 ohm load impedance 
in that last column, it's not too surprising that the output voltage sometimes exceeded
the input voltage.   But there are some weird anomalies in those figures I haven't quite 
figured out, and I'm not totally convinced my measurements are accurate.

At all frequencies, the dummy load got plenty hot.  
The toroids of the box I had taken the lid off of did not.

A few more general notes, not part of the measurements presented here.

I've measured SWR into my EFHW antenna system, my figures agreed with 
the QST article and with the charts on the website.
I believe the charts on were created with 100' of RG8X coax
from matchbox to the SWR meter plus RF source.
My SWR measurements with a very short coax showed higher SWR figures out-of band,
but the SWR was perfectly fine within the bounds of most bands  (not all of 80m or 10m).

The SWR minimum is pretty well centered on a usable part of each band, 
due in part to the small airwound coil in the wire near where it enters the matchbox.
A possible exception is 80m, the SWR min is centered on the CW portion of the band,
a tuner of some sort would be needed to use it at 4.0 mhz.

Given that an RF source can pump power into this antenna system without
anything getting hot, I can only conclude that most of that power is properly radiated.

In what pattern it gets radiated is debatable.
Here's an FAQ for this antenna system:
Near the bottom are four links, clicking on the first we see plots for a straight half wavelength of wire 
at 40 foot above the ground.
Ideally, an 80m antenna would be a half wavelength above the ground for an optimal pattern,
so 40 meters, or 130 feet.  I'd bet most hams with an 80m antenna aren't climbing trees
and rooftops much higher than the 40 foot of the plots shown in that last link.
At 40 foot, the 80m plot shows it's going pretty much straight up, a good NVIS antenna perhaps.
The others all show lobes that look to me as if they could do some good for DX.
For any ham who plans to use 80m, this EFHW is nearly equivalent to a dipole at the same height.
The other bands are gravy, and that gravy doesn't look bad.

Yes, having several different antennas would be better.
A 30 mhz rotatable beam is easy enough, and would work much better than this wire.
But for some of us, it's just a hobby.

Jerry, KE7ER

Join to automatically receive all group messages.