Last update: 31 October 2009 (added 5 articles in German)
Last update: 31 October 2009 (added 5 articles in German)
A "balun" is bi-directional adaptor between a BALanced device and an UNbalanced device. Twin-lead feedlines and dipoles are "balanced" (two signal conductors, with equal-but-opposite current: symmetrical). Coax cables and the typical antenna port of a transceiver or antenna tuner, are "unbalanced" (one signal conductor that is referenced to ground; a.k.a. "single-ended"). Baluns other than those with a 1:1 transformation ratio, also act as impedance transformer.
Two basic types: voltage baluns and current baluns:
 |
Voltage baluns comprise a transformer with at least two sets of windings; they may be interconnected like an auto-transformer. If the input and output section of the transformer has the same number of windings (ratio = 1), than the balun has the same voltage and impedance at its input and output. The impedance transformation ratio is the square of the voltage transformation ratio, which is the ratio of the number of windings. |
|
|
Current baluns typically consist of ferrite beads on a section of 2-conductor transmission line such as a coax, or conversely, coax wound multiple times through a ferrite ring, or tight-wound coil of coax windings (air core). Its purpose is to provide high impedance to common-mode current on the transmission line: it "chokes" that current. There is no impedance transformation. When placed between a balanced device such as a symmetrical antenna (ignoring stray capacitances etc. from surrounding objects that may actually make the antenna a-symmetrical), and a coax (unbalanced transmission line), the current balun rejects the common-mode and passes the differential voltage. |
Note: these
articles are copyrighted material; all related restrictions regarding access and
usage apply.
|
|
Articles:
|
|
|
Recommended book:
|
||||||||||
|
|
Articles in German:
|
I currently use two baluns:
|
|
1:1 current-choke balun at the output side of my antenna tuner |
|
|
4:1 balun between the 1:1 balun and the twin-lead feed line of my antenna (Cobra dipole only) |
My 1:1 balun is a W2DU-type current choke. I am (only) using 24 ferrite beads (no. 77 material from Amidon; $11 + S&H mid-2008 pricing from Amidon) that snugly fit on about 30 cm (1 ft) of RG-400 coax.
W2DU
ferrite-bead 1:1 current-choke
W2DU
ferrite-bead 1:1 current-choke "balun"
Close-up of a connector end of the balun
My 4:1 balun is a BL1 kit from Elecraft, consisting of a two-hole "binocular" (a.k.a. "pig nose") ferrite core and two sets of thermostat wire windings. The balun is rated for 150 watt (when connected to a 200 Ω load). Elecraft has since replaced this model with the BL2 kit with two stacked cores, switchable between 4:1 and 1:1, and rated at 250 W, for $35 + S&H (mid-2008 pricing).
Elecraft BL1 4:1 ferrite-core balun
On of these days I'll try and replace the "4:1 balun plus W2DU current-choke" with two W2DU chokes in a series/parallel configuration per VE2CV. See figure below. This is supposed to provide a low-loss wide-band balun, see here.
NOTE: to obtain 50 Ω at the unbalanced (parallel) side and 200 Ω at the balanced side, the VE2CV configuration requires that the two W2DU chokes be made with sections of 100 Ω coax! Closest available Z0 is 93 Ω; 100 Ω is typically only available as twinax. Not having been able to locate either value, I used plain 50 Ω coax. Obviously this significantly influences the measurement results! As the two current-chokes are put in series at the balanced output, the VE2CV can also be considered as a current-BALUN; the output is floating, which is not the case with voltage-BALUNs.
Two W2DU 1:1 current baluns configured as a VE2CV 4:1 current
balun
 |
Update 12-July-08: made a second W2DU choke balun and a quick-and-dirty hook up with 2x2 BNC chassis jacks tp create a VE2CV configuration. It appears to work. That is: I can tune my antenna pretty much to the same SWR and output power as before with the 4:1 and W2DU choke in series. However, tuning has become very finicky: the "low SWR dip" has become extremely sharp, and varies with my proximity to the antenna knob of my tuner. Maybe a problem with the ground/shield interconnections. To be continued... |
|
|
Update 27-July-08: determined that the observed behavior is due to shoddy soldering job that I did on one of the BNC chassis jacks. Got the big-gun soldering iron out and corrected the problem. |
|
|
Update 1-Aug-08: tried to measure my baluns. Put the BL1 and VE2CV 4:1 baluns back-to-back (1:4 + 4:1 = 1:1), and also measured each of these baluns stand-alone with a 200 Ω termination resistor. The test configurations are illustrated below and the analyzer plot is provided for each configuration. I'm still interpreting the results. First observation is that the two 4:1 baluns behave quite differently. The miniVNA analyzer only outputs one mW or so, so the BALUNs will never be forced into saturation during the measurements. |
The back-to-back configuration shows a transmission loss from close to 0 dB at 1.8 MHz, to close to 1.8 dB at 30 MHz.
|
|
For the BL1 terminated with 200 Ω, the analyzer sees a flat 50 Ω from, and an SWR better than 1.1 (1.8-30 MHz). |
|
|
For my VE2CV terminated with 200 Ω, the analyzer sees an Rs of 50 Ω around 1.8 MHz and gradually going down to 37 Ω at 30 MHz; SWR increases from better than 1.1 at 1.8 MHz to almost 1.8 at 30 MHz. |
Test configurations with VE2CV and BL1 baluns back-to-back (1:4 + 4:1)
(only top configuration tested so far)
Transmission
mode plot of the above test configuration (adjusted by -1.2 dB for coax loss)
(click on image to enlarge)
Test configurations
for BL1 balun, VE2CV balun, and terminator resistors
Analyzer plots (left-to-right): BL1 with 200 Ω termination, VE2CV with 200 Ω,
and straight 50 Ω
(click on image to enlarge)
As suggested by Gerd, DO1MGK (SK), I conducted some more measurements of the W2DU chokes, to try and get a good feel for the signal attenuation (0 dB loss desired), and common-mode attenuation (high-dB loss desired). Test-configuration set-ups are illustrated below. The analyzer plots below it show something close to 0 dB insertion loss (after adjusting for ≈1.2 dB loss of the "black" coax) and about 20 dB common-mode attenuation on the shield (1-30 MHz). The latter is not bad, but not great either. A minimum of 25-30 dB would be good. I have put two of these chokes in series (2 x 24 beads): transmission loss 1.1-1.3 dB (several connectors more in series than with the depicted single choke configuration), and 25-34 dB common-mode attenuation on the shield (1.8-30 MHz). That's better! Note that the resulting choke is not nearly as compact as the coax-on-toroid choke described further below (which uses different ferrite material).
Test configurations of the W2DU current choke
Analyzer plots
(left-to-right): 50 Ω termination, insertion loss, common-mode attenuation
(click on image to enlarge)
Update 11-Aug-08: after consulting with Felix, HB9ABX, I evaluated a different form of current-choke: 6 windings of coax on a triple-stack of ferrite rings made of #43 ferrite material. I used Amidon FT-140-43 cores ($2.35 each + S&H, mid-2008 pricing). Using RG-58A/U coax, it is hard to fit more than 6 windings on these toroids. The windings are spread out evenly around the ferrite cores. The same toroids can be used to suppress TVI/RFI problems (e.g., by winding several turns of the wire-pair from your stereo set to your speakers). That's why I had eight of them in stock, just in case. I stacked three of them with a couple of tywrap cable ties.
Ferrite toroids (#43 material; Amidon FT-140-43)
6-winding current-choke on 3 toroids
(tape measure in cm)
The test-configuration set-ups are illustrated below. The corresponding analyzer plots show SWR better than 1.1, about 1.3 dB insertion loss (I actually used a configuration with more connectors than shown in the diagrams; at least several tenths of a dB right there) and between 28 and 38 dB common-mode attenuation on the shield. This is a decent broad-band current choke for 1.8-30 MHz!
Analyzer plots
(left-to-right): with 50 Ω termination, insertion loss, common-mode attenuation
(click on image to enlarge)
Another way to loop coax through a ferrite ring is the "split winding" method illustrated below. When I have some time on my hands again, I will compare its performance against the "contiguous winding" method used above. Apparently the chocking performance for HF (up to 30 MHz) is better than that of "split wound" version (ref. W8JI).
And maybe I'll evaluate an air-wound 1:1 coax choke next! No costly ferrite-cores are needed, just a short length (30 cm, 1 ft) of plastic pipe with a diameter of 7-12 cm (3 to 5 inch) from the DYI store, on which about about 6-8 meters (20-25 feet) of 50 ohm coax is close-wound (about a dozen turns). Some long nylon cable ties keep it all together. Clearly not as compact as a stack of ferrite rings...
My twinlead-to-BNC
adapter
My twinlead-to-banana-plug
adapter
©2007-2008 Frank Dörenberg F/N4SPP
