Can a dipole be fed at its end (or near end)?
No one argues a dipole cannot be fed at its center or even quite a bit off center. Some suggest it is impossible to feed a dipole at its end without some sort of “other” conductor to “push against.” My simulations suggest one can “push against” the opposite terminal of whatever power source exists at the end leaving it with very high voltage wrt the rest of the antenna system. Presented below is a test measuring the magnetic field strength near the dipole wire vs. position along the wire. I perform the test twice differing only in the position of the energy source.
The “ideal” dipole test setup
Figure 1 portrays the ideal test configuration. Non-conductive supports (blue) hold a dipole (red) trimmed for operation in the 30m amateur radio band. In actuality, I made two dipoles from yellow insulated lab wire, one with feedpoint midway (point C), the other with an LNR Precision transformer at the end (point E).
The “real” dipole test setup
It’s springtime here in Virginia and nightly thunderstorms with the occasional tornado watch are the norm. Sometimes you have to sneak in a test wherever and whenever possible. Such was the case when I finally had an evening available to perform a quick test.
Yeah that’s my smoker just beyond the dipole’s end. The far support is a non-conductive fiberglass mast. The rear support, behind the photographer, is my house. Each yellow wire dipole has black markings at one foot intervals to provide quick guidance for sensor placement during tests.
The antennas under test
In a perfect world I would have one dipole wire for the 30m test with two feed positions. I instead made two antennas: one center-fed and one end-fed. I did perform S11 on both to trim for the 30m band. Unsurprisingly, both are the same length. The VSWR for the center-fed is under 1.5. The VSWR through the step-up transformer for the end-fed is also under 1.5.
Dipole testing electronics
If we are to test a dipole with no extra conductors, such as feedline, then we must place the same restrictions on the testing gear. Fortunately we live in the 21st century where battery operable tiny computers with WiFi built-in provide slick solutions.
Let’s introduce the active components used to energize the antenna and measure the magnetic field.
30m Transmitter Raspberry Pi assembly
A way to energize the dipole without any feedline was remarkably simple with the stack seen here.
From bottom to top…
- Raspberry Pi-Zero W with WiFi that works surprisingly well for its size (Attached to my interior WiFi at the other end of my house no problem).
- Spiffy 3 volt lithium power Hat to power the R-Pi.
- Header extender,
- TAPR 30m LPF and amplifier Hat.
In the above photo, the transmitter assembly connects directly to the center of the 30m dipole using a BNC-Banana adapter. With the TAPR 30m hat, the transmitter generates about 150 mW of power measured into 50 ohm load.
A bit of tape on the opposite side ensures the stack stays so.
More details available in my other article…
Some earlier photos of the transmitter
Here you see the transmitter held together with rubber bands. It is attached to the broadband LNR Precision transformer from their 10/20/40 product reviewed earlier in this end-fed configuration.
The recent purchase of a PL-259 to male BNC adapter (not easy to find) shortens the connection between the transmitter assembly and the transformer. It’s tight.
No wires. No transmission lines. Self powered. An antenna energized by this transmitter is as isolated as one can get.
Magnetic field Raspberry-Pi sensor assembly
Allow me to introduce my little friend… Frankensensor.
This is what happens when you take the electronics described in an earlier article into an Ace hardware store for packaging ideas. Key electrical components top to bottom include:
- Honest to goodness passive magnetic sensor from Beehive Electronics (the yellow loop with wand).
- Logarithmic RF power to voltage converter (aka RSSI) and a 9V battery to power it.
- Raspberry Pi stack (details below) to measure the voltage and convey the readings via WiFi (no additional wires).
For structure we have:
- One 6″ disposable drywall knife modified with a couple of holes at the blade edges.
- Wire ties arranged upwards to a couple of curtain clips, or some such things, that hang on the antenna wire.
- More wire ties to ensure conductive wiring isn’t holding weight.
- A bunch of hot glue in key spots to keep the antenna clips pointing up.
- Some double-sided tape to keep the sensor wand always coplanar with the antenna wire (not that a few degrees would make much difference).
The goal is pretty simple… maintain the sensor loop the same distance from and orientation with the antenna wire and hold it there during any specific measurement. Indirectly sensing antenna current this way perturbs the antenna a bit less than a clamp-on ammeter… or so I suppose. It was easy enough to try.
This contraption isn’t my proudest moment in fabrication, but it works quite nicely.
Sensor Raspberry Pi stack
Zooming in a bit on the previous photo provides a better view of the active electronics.
The RF to voltage (RSSI) board hangs on the end of the sensor wand and has an LED to indicate power. It has a pair of wires to convey the measured voltage to the Raspberry Pi assembly that includes from bottom to top…
- 16 bit, four channel, A/D hat (only one channel used for this test),
- Battery power hat using 3 volt lithium cell. The five LEDs indicate full charge,
- Raspberry Pi-Zero-W with built-in WiFi,
- Some tape to ensure it stays together.
The battery hat connects directly to the header of the Pi-Zero. The height of the battery hat precludes direct connection of the A/D requiring the use of the intermediate header.
All these components are readily available to anyone to replicate the test.
Pro-tip: The Pi-Zero will NOT boot if the power to the RF-to-voltage board is on and a voltage appears on the A/D. I have no idea why and it took a long time to puzzle through this issue. If you build one of these, just connect the 9V battery after booting the Pi-Zero.
More details in my other article…
The control point
With both Raspberry Pi computers booted and connected to my home WiFi, I shelled into each of them with my computer and monitored the transmit signal with a portable shortwave receiver.
Click to open the view of the control windows below.
- Shell into the sensor R-Pi upper left showing various measurements in millivolts.
- Shell into the transmit R-Pi lower left using PiCW to generate the test signal on DIO4. The TAPR hat filters and amplifies this signal.
- Red shell with a text file editing session to record the millivolt measurements vs. sensor position.
Finally the result
Converting the millivolt readings to A/m requires a bit of math. We have the inverse and logarithmic nature of the RSSI module, the published formula from Beehive Electronics and then conversion to A/m. Here is a quick view of the math in the GNUplot command file.
# power-voltage Y=mX+b parameters of the AD8318 IC in the # RSSI module as measured in lab by KX4O m = -.024433333 #volts/dB (differs from -.025 V/dB data sheet value) b = 26.7535 #dBm (differs from 20 dBm data sheet value) # Inline conversion to Tesla per Beehive Electronics datasheet (for 10.1 MHz), # then to A/m (x 797700) from the RSSI voltage. plot [0:44] \ '2019-06-01_EFHW_Data_001.txt' using 1:((10**(($2/1000/m+b)/20-5.259))*797700) '2019-06-01_CFHW_Data_001.txt' using 1:((10**(($2/1000/m+b)/20-5.259))*797700)
GNUplot rocks! That I happened to be within the scale and range for each component in this signal chain was dumb luck. Behold…
If we assume the strength of the magnetic field immediately around an antenna wire is proportional to the current at that point…
…you can clearly see the classic half-wave dipole current distribution.
- Both the end-fed and center-fed antennas operate just fine without any additional conductors or feedlines.
- The end-fed energy is less than the center-fed. If you do the math, it emits about 1.6 dB less power than the center-fed.
- Given the small toroid size, the LNR Precision transformer is probably the reason for the small loss, but that’s just a guess.
- I have no explanation for the lower readings of the center-fed at 37-40 feet. Yeah they bother me too, but that’s life in Realville. This doesn’t detract from the overall message conveyed by the graph so I’ll file a copy of this perturbation under “dammit” and move on.
- Center-fed, end-fed (or “pretty darn close to end-fed” if you like), you can clearly see the dipole operates well from either feed point position.
I can find no reason to suggest a dipole cannot be fed anywhere along its length without aid of other conductors to “push against.” Simulations agree. Yeah the above graph reminds us there is a loss penalty to pay for a proper transformer to feed a dipole at its highest impedance point, but there’s nothing magical going on here. It just works.
As discussed here…
…potential problems lurk when you actually do attach a feedline although most feedline lengths are quite forgiving. The story continues to evolve on this web site…
…and may well forever. They key takeaway from this test is for owners and users of EFHW antennas to fret not what some folks have to say about their aerials and just use them with some confidence there is no bogeyman lurking to spoil their day.
If my transformer efficiency guess above is true, later models with higher power ratings may shrink the difference. I have a later model higher power LNR transformer from their EF-QUAD product and another transformer from MyAntennas. Both are much much larger and heavier than the mostly QRP transformer used for this data. I will repeat the tests above with these two newer units to explore this topic a bit more.
UPDATE June 3, 2019
I did test the transformer from the LNR-QUAD. It has a more robust toroid within.
Let’s see what happened.
Well, the small step-up transformer from the LNR-10/20/40 EFHW and the larger transformer from the LNR-QUAD perform identically. Both are about 1.6 dB lower radiated power than the straight up center-fed dipole. I suppose the good news is the size of the toroid governs max power handling, but not efficiency.