Category: Antennas

G3LNP balun with symmetric ‘matched’ load

My article G3LNP balun explored the operation of the G3LNP 4:1 balun on a 200Ω asymmetric load and found it exhibited extreme Insertion VSWR on what should have been an ideal impedance transformation but for the asymmetric element.

The balun is in fact a Voltage Balun and cannot be expected to work properly on asymmetric loads.

A correspondent proposes that the balun probably works very well on a nearly symmetric load such as a half wave dipole.

There are two aspectes to this proposition:

  1. the assumption that a common half wave dipole implementation is nearly symmetric; and
  2. the balun works well on a nearly symmetric load.

Continue reading G3LNP balun with symmetric ‘matched’ load

Last update: 14th October, 2017, 6:47 AM

G3LNP balun

G3LNP described a 4:1 balun for HF antennas in Radcom Nov 2017.

Above is the schematic supplied by G3LNP. He describes the dashed link at the bottom as optional, but uses it in his prototype so this analysis is with that link installed. The prototype used equal lengths of coax (1m PF100, an RG-6 like coax), and the toroidal choke appears to be 8t on a T130-2 powdered iron core.

Exploration of behaviour of baluns on extreme asymmetric load often reveals whether they work properly for asymmetric loads.
Continue reading G3LNP balun

Last update: 14th October, 2017, 6:50 AM

“No short antenna is more efficient than a resonant-length antenna”

A recent online posting asserted that an antenna is optimal when itself resonant, and fed with a resonant feed line length so delivering a purely resistive load to a source, and further that implementors needed to be careful that a shorter dipole could be offset to some extent by a longer feed line but it would be inferior because:

no short antenna is more efficient than a resonant-length antenna

… but does that stand scrutiny?

An NEC experiment

Lets walk though an experiment using NEC-4.2 models of a dipole of 2mm copper wire at 10m height at 7.1MHz over average ground (σ=0.005, εr=13).

Key assumptions:

  1. source has a Thevenin equivalent source impedance of 50+j0Ω;
  2. feed line is lossless.

The results are sensitive to the model assumptions.

We will calculate the ratio of radiated power to the power delivered by the transmitter to a matched load, let us call it TransmitEfficiency for the purposes of this article. Continue reading “No short antenna is more efficient than a resonant-length antenna”

Last update: 22nd September, 2017, 6:21 AM

Small untuned loop with MMIC LNA on 80m

This article documents a small experiment with a quite small untuned loop, and LNA and receiver on 80m to assess its ability to copy signals on the band as well as the station transceiver on large antenna.

A significant factor at 80m is that ambient noise is quite high. Let’s consult ITU P.372-12 for guidance.

Curve E is the median city noise, at 3.6MHz Fa is about 62dB. At a more detailed level, P.372-12 gives the median noise figure for Rural precincts at 51.8dB and that figure is more appropriate to the test location (large block rural residential).

The test was carried in a few spots, at 50-100 m from the main station dipole. Continue reading Small untuned loop with MMIC LNA on 80m

Last update: 27th October, 2019, 3:58 PM

Demonstration of the GR1606B for antenna Z measurement and calcs

This article demonstrates use of a GR1606B RF impedance bridge for measurement of the feed point impedance of a MHz loaded mobile whip. The antenna is roof mounted on a vehicle and measurements are made looking into 4m of RG58C/U, then transformed to feed point impedance using three tools:

Key metrics are:

  • F=7.05MHz
  • R=35Ω
  • ReactanceDial offset=200
  • ReactanceDial=442
  • X=(442-299)/7.05=34.33Ω
  • Z=35+j34.33Ω;
  • Feed line is 4m of Belden 8262 (RG58C/U).

Continue reading Demonstration of the GR1606B for antenna Z measurement and calcs

Last update: 13th September, 2017, 10:57 AM

Inductance of a loop of CAT5 pair

An online expert recently reported:

I tried to make an antenna loop for longwave with cat 5 and after it did no good I realized the twisted wires canceled each other out.

Or did they really cancel?

Parallel connection

I constructed a loop of one Cat 5 pair and measured its inductance when both wires are bonded at the ends.

The conductors are 0.5mm diameter and spaced 0.9mm. To estimate the inductance we use the geometric mean radius (GMR) as the equivalent radius of the pair. GMR=(0.5*0.9)^0.5=0.67, diameter=1.34mm. So let’s calculate the inductance of a single turn circular loop of 0.8m perimeter and round conductor of 1.34mm diameter.

The estimate above is 850nH.

Above is the measurement, the screen is not readable, but it is 852nH, very close to the estimated 850nH. Continue reading Inductance of a loop of CAT5 pair

Last update: 2nd September, 2017, 8:52 PM

GPS antenna plot

I bought an inexpensive GPS antenna on eBay, the requirement was for one that operated from 3-5V to suit both of my GPSDOs. The antenna is mounted in an electrical junction box on conduit above the roof for reliable coverage.

The question was whether the active antenna with 5m of RG174 coax and SMA male connector at A$6 posted was any good.

To map its behaviour, it was attached to a Trimble Thunderbolt GPSDO and Lady Heather used to plot rx signal level over about 30h.

Above is the plot. The pattern is not quite symmetric as there are obstructions in play, in fact some of the dips in performance are explained by specific trees and the roof profile. There is a gap to the south at low elevations, GPS satellites don’t fly there (MEO inclined 55° from the equator) .

Overall, it reaches similar strength at the zenith as other antennas tried.

Overall evaluation, it seems to work ok though the coax is a bit rough.

Last update: 30th August, 2017, 6:25 PM

Messi & Paoloni Ultraflex 10 coax cable

Messi & Paoloni Ultraflex 7 coax cable compared M&P UF7 with RG-213. This article does a similar comparison between M&P Ultraflex 10 and LMR400UF.

Both cables are of similar size, ~10mm overall, stranded centre conductor and foil+braid outer conductor. The shield stranding is different and the foil is copper in the UF10, aluminium in the LMR400UF.

Let’s take the loss factors calculated for TLLC and de-construct the conductor and dielectric loss for each line type.

Above is a comparison of the cables. Continue reading Messi & Paoloni Ultraflex 10 coax cable

Last update: 18th August, 2017, 7:14 PM

True balanced tuner

A recent long running thread on QRZ entitled “True balanced auto-tuner” was sure to tease out some pretty woolly thinking… the word “true” was enough to signal the outcome.

There are only three words in the title, we can dismiss “true” as a harbinger of woolly thinking, and though people will argue the toss on the appropriateness of the term “auto-tuner’, most people share an understanding of the meaning. “Balanced” is another problem altogether.

After thirty odd posts, there has been no definition or discussion of the term balanced, or its advantages or disadvantages.

One of the recommendations by several posters is the old is new again solution, the once popular link coupled tuner and the work of W5ZQ featured in one of those recommendations.

W5ZQ and WW8J

Topology

W5ZQ describes a tuner inspired by WW8J. W5ZQ extended the design and provides a writeup on optimising balance.

Above is W5ZQ’s partial circuit. In the article he describes and shows:

  • adjustment of the grounding point of the output tank; and
  • current meters which presumably attach to J2 and J3.

Key to analysis of the topology is that the centre of the output inductor is grounded. This results in the circuit tending towards equal but opposite phase voltages on the output terminals. Continue reading True balanced tuner

Last update: 22nd May, 2020, 12:34 AM

WSPR – data mining 40m 01/08/2017

I mentioned elsewhere that I dowloaded the WSPR archive for 1/08/2017, and particularly analysed 40m spots.

There were close to 1,000,000 spots for the day, about 340,000 on 40m, and about 20,000 individual transmissions reported during the day (40m).

Tx / Rx mix

Above is a pie chart of the mix of Tx Only, Rx ONLY and Tx/Rx stations. The largest group is Tx Only, 44% of stations do not contribute reception records. Nextly, 30% are Rx Only, and 27% Tx and Rx. Continue reading WSPR – data mining 40m 01/08/2017

Last update: 7th August, 2017, 7:09 PM