ferrite types

bead balun

voltage balun


iron powder

Core Baluns

(published in Electron #4, 2007)








This chapter is the third in a series of articles on baluns for antenna applications. The first article is an introduction with some background on balun types. The second article is on wire baluns. It is advisable to read the articles in the above order especially since each next chapter is building on information and formula's already explained earlier and referencing to this.






When going from an a-symmetrical to a symmetrical system some form of adaptation / matching should be applied to uphold the integrity of both systems. This may be accomplished by ensuring no current to flow anywhere except for the desired signal path. A simple and direct way to prevent other currents to flow is blocking-of all other current path'. A sleeve choke, 1 : 1 current transformer or current balun will perform this system function. The better this choke, i.e. higher impedance, the better this current balancing function in the signal path.




With the current balun requirements as settled, we checked the performance of a wire-balun in the previous chapter. Here we discovered this component could well be constructed for a limited frequency range (1 : 3), but was increasingly more difficult to design for higher bandwidth. Main problems were leakage inductance at lower frequencies and high parasitic capacitance to limit the high frequency side. Both problems may be attacked when applying a ferrite core at the balun since high permeability will enhance inter-winding coupling (prevent leakage inductance) while at the same time allowing for a lower number of turns (diminishing parasitic capacitance).


As an example we will calculate a 1 : 1 current balun around a 36 mm. toroide of 4C65 (61) ferrite material. For this function to be effective, sleeve impedance should be at least four times system impedance which we will set at 50 Ohm, so sleeve impedance should be 4 x 50 Ohm = 200 Ohm. At the lowest operating frequency of 2 MHz. this translates to a reactance of 16 μH. as we noticed in the previous chapter.


At the manufacturers web-site we will find for this material and core shape the winding factor AL = 170 nH/n2, so we will need a number of turns at this core:

n = 16 / 0.17 = 9,6 (10) , which leads to the construction as in figure 1.




Figure 1. Current balun 1 : 1




Note: The balun in figure 1 has been constructed using RG58 coaxial transmission line. Since the electromagnetic field is 'locked-up' inside the transmission line, the integrity is being preserved independently of the 'handling' of the line, whether in a straight line, curled up or around some core material. This applies to coaxial transmission-line as well as symmetrical line types. Since the latter is not as 'closed' as coax, turns should not be wound too closely and preferably not to overlap.






The balun as above has been constructed and measured in a condition of maximum unbalance, as discussed before. Results for transmission (insertion loss) may be found in figure 2.




Figure 2. Insertion loss of a 4C65 current balun



The insertion-loss for the 1 : 1 current balun on 4C65 may be directly compared to the wire balun in the previous chapter. It is easily noticed the 4C65 construction to perform significantly better, with the 1 dB loss position at the lower frequency side to have shifted from 1,8 MHz. to 0,9 MHz. and the high frequency side from just over 50 MHz. to over 150 MHz. In the interesting frequency range 3 - 30 MHz., insertion loss has improved from 0,16 dB on average to 0,07 dB and this may be regarded as almost 'ideal'.


As with the wire balun we also measured reflection with the balun again terminated into 50 Ohm. The results of this measurement may be found in figure 3.




Figure 3. Reflection of a 4C65 current balun



Again the improvement as related to the wire balun is clearly visible. SWR is very low (1,03) from the first measuring frequency at 0,5 MHz. (compare to SWR = 1,5 @ 4 MHz. for the wire balun), up to SWR = 1,25 at 200 MHz. (compare to SWR = 1,5 @ 50 MHz.).



There is not such a thing as a free lunch, so we may be curious as to what price these excellent properties?

As it happens, price to pay is very low. In the chapter on HF ferrite applications we found that above 3 MHz. maximum core load is determined by power dissipation inside the core, with highest stress at the highest operational frequency. In the same chapter we derived a formula for the system voltage that will make core temperature rise by 30 K. at the maximum internal power dissipation of Pmax.


UL(dissipation) = √(Pmax . XL (Q/6 + 1/Q))


with Pmax = 4 Watt for this 36 mm. ferrite core. At a highest operational frequency of 30 MHz. we find

UL(dissipation) = 115 V, allowing for a system power of 265 W in a 50 Ohm system.


We should be aware (again) that above calculation is for a system impedance of 50 Ohm, e.g. about the impedance of a resonating dipole antenna. Outside resonance, impedance will be much higher, even though the antenna tuner in the shack will have transformed all odd impedances back to 50 Ohm for the transceiver. The balun may therefore be stressed beyond allowable limits without the operator being aware of this.



Different cores


In the first chapter on baluns we discovered the minimum choke impedance of four times the system impedance. It is not important how this impedance is created, as long as this is high which is the same is we found in Ferrites for EMC applications. One therefore may be wondering whether only HF ferrites will qualify and what would happen when applying LF ferrite materials?

To test this, I made the same balun (10 turns on 36 mm. toroide) on 3E25 type of ferrite, a LF type of material with a high initial permeability of 6000, as compared to 4C65 (125).



Measuring set-up is identical to the other tests (under maximum unbalance conditions) and results may be found in figure 4.




Figure 4: Insertion loss of a 3E25 current balun



As may be expected the low frequency behavior is even better than with the 4C56 core, and will probably extnd to far below the lowest test frequency of 0,5 MHz., although we do not need this for HF applications. At the high frequency side we find a 1 dB loss at 100 MHz. because this LF material really was designed for low frequencies. At 100 MHz. even the material's loss-factor is at the end of its operational life.



Reflection properties of this 3E25 balun may be found in figure 5.




Figure 5. Reflection of a 3E25 current balun



Measurements extend to 0,5 MHz. but are so close to the X-axis that this is not showing in this graph anymore.

Again we find a very useful component for HF frequencies that may be applied to over 100 MHz.



Since parameters are different again for this core material, we calculate maximum voltage across this choke with the formula above, at highest operational frequency, to find UL(dissipation) = 100 V. This allows for a system power of over 200 Watt in a 50 Ohm system.


Although exhibiting somewhat less bandwidth and power than the balun on 4C65, also this 3E25 balun will perform an excellent job over the entire HF frequency range. Since this is a (very) low frequency material it is save to say that all ferrite material from the junk box will qualify as a core for a current balun.



Parasitic capacity

As we have discovered, these current baluns will perform well as long as the sleeve impedance is high. In practice we find a parasitic capacitance in parallel to the choke with a diminishing impedance with frequency. Since the value of a capacitor is related to the voltage across, we like to keep first and last turn of the choke as far apart as possible. When measuring the effect of distance it appears that for these current baluns a distance of about one centimeter (two RG58 diameters) is sufficient for this capacitance effect to be low.

An alternative winding technique has been tested to keep this first to last turn far apart. This technique first puts half the number of turns on the core and then crosses over to complete the other half of the turns in the opposite direction. First and last turn will end up half a core diameter away. As it happens, cut-off frequency was lower than with the straight winding method, so this allegedly 'better technique' is no improvement.


Bead balun


In the last chapter and also before the role of the parasitic capacitance at the high frequency side was clear and was a limiting band-width factor. To diminish this capacitance we may consider to not coil-up the transmission-line around the core, but 'coil the core' around the transmission-line. The latter situation arises when stringing ferrite beads on the transmission-line since we also know that every time the line passes through the center hole of the toroide, this accounts for a full turn. The string of beads therefore acts like many reactances in series, as also applies to parasitic capacitances.


When constructing this 'bead-balun' at RG58 coaxial cable (diameter is 5 mm.), a well fitting toroide shape is TN10/6/4. Using ferrite type 4C65, winding factor AL = 52 nH/n2, so for a total inductance of 16 μH we would need over 300 beads. This balun would be 1,35 meter long, which is not really practical anymore. When selecting 4A11 type of material (AL = 286 nH/n2) we would need 56 beads with a total length of 25 cm., or when selecting 3E25 (AL = 2250 nH/n2) only 3 beads at a length of 3 cm.


Since maximum dissipation of these beads is scaling with the root of volume (see Ferrites in HF applications) , maximum power dissipation of a 4A11 bead of this size is 30 mW. Maximum voltage per bead is: UL(dissipation) = 1,6 V., allowing for maximum voltage of 90 V. for the 56 bead balun, which translates to maximum system power of 160 Watt in a 50 Ohm system. Although all ferrite batches are created equal, some beads may be more equal than others making up for a different voltage distribution, so a maximum 100 Watt would be a safer limit for total system power.

When preferring somewhat larger RG213 cable with outside diameter of 10 mm., the bigger size ferrite bead TN23/14/7 would fit and would have a winding factor AL = 485 nH/n2 when 4A11 type of ferrite will be applied, requiring only 33 beads. The balun will now have a total length of 24,8 mm. and would allow for a maximum voltage of 2,9 V. / bead, or in total 95,7 V. allowing for 183 Watt in a 50 Ohm system.

Except for using a lower number of beads, everything else did not really change that much.


Also the route along the 3E25 road is not really better since the number of beads, the μ' and μ" will be lower making maximum allowed system power even lower still. It is clear bead baluns will be superior for very-wide bandwidth applications (EMC measurements) but will be less practical in HF (only) applications since the simple 10 turn, 36 mm., 4C65 toroide is already quite near 'perfect' for these applications.


Voltage balun


Up to now we have been discussing current baluns since we found in the first chapter this to be the device we would need to match a symmetrical dipole to an a-symmetric transmission-line. Still one may regularly find voltage baluns recommended for this type of applications in ham magazines. The usual construction of this voltage balun is a trifilarly wound transformer, wired-up such that the balanced connection is taken around a 'ground' connection and the a-symmetrical side with reference to this ground terminal as in figure 6.



Figure 6. Example of a 1 : 1 voltage balun



In figure 6 we find the generator connected across two windings of the transformer with the balanced side also taken from two windings. Since all windings consist of the same number of turns (trifilar) and are closely coupled, input and output voltages are equal.


In the diagram the balanced output is situated around a 'grounded' terminal, which may not be so very much 'grounded' anymore at the end of the feed-line. According to 'Reference data for Radio Engineers', standard hook-up wire is showing an inductance of 1,6 mH/mile (1 μH/m) for frequencies between 3 - 30 MHz. For a feed-line length of 10 m., inductance is 10 μH or around 600 Ohm at 10 MHz. It is clear this 'ground' terminal is at low impedance any more.


The model of figure 6 may be constructed as in figure 7, when applying the same principles as in our earlier designs (10 turns on 36 mm. toroide of 4C65 material to obtain 200 Ohm at 2 MHz.). The generator is connected across two windings in series, so effectively across 10 turn.





Figure 7. Trifilar 1 : 1 voltage transformer




Measurement at ideal operating conditions


To measure this transformer at ideal operating conditions, I constructed two transformers and connected these 'back-to-back' to keep the balanced side in-between the transformers at a perfectly balanced level. Results of the measurements have been 'halved' to represent each of the transformers.



Results of the transmission measurement may be found in figure 8.




Figure 8. Insertion loss of a trifilar voltage balun



The insertion loss as in figure 8 is showing a fairly good transformer that starts dropping-off below 1 MHz. and above 60 MHz. In between the graph is flat and is showing low loss.



Reflection characteristic of the trifilar voltage balun may be found in figure 9.




Figure 9. Reflection of a trifilar voltage balun



The graph of figure 9 is showing SWR which is somewhat better than the first wire balun in the previous chapter at the low frequency side but definitely worse at the high frequency end. Also in-between cut-off frequencies the SWR is not really low at 1,25. All in all this is a marginally performer.



Measuring under practical conditions


In the above situation the voltage balun has been tested under 'ideal' conditions, i.e. a condition of perfect balance. In a practical situation there usually is some form of unbalance, up to total unbalance in an adverse situation. Therefore all earlier baluns have been tested in a fully unbalanced situation to prevent surprises when in an operational situation. To compare balancing qualities of various balun types, also this voltage transformer has been tested under the same conditions with one output terminal grounded, to simulate a condition of maximum unbalance. The voltage balun however is not a symmetrical device as may be appreciated when regarding figure 6. One output terminal is connected to a winding to the generator, while the other output terminal is left open. For a complete impression we therefore have to measure this voltage balun two times, with different output terminals grounded.


Note, this unequal situation also exists for the current baluns. In testing with the 'center-conductor' grounded we always have been measuring under most unfavorable conditions. With the other terminal grounded in a current balun, we would have measured just an other transmission-line with only transmission-line damping as a non-ideal condition. Also when measuring in reflection only a 'perfect' transmission-line would have been found.



Two measurements, each with a different output terminal grounded may be found in figure 10.




Figure 10. Insertion loss of a trifilar voltage balun, one side grounded



In figure 10 we indeed see different behavior depending on the grounded terminal. Even the best of the two graphs only is showing insertion loss below one dB over a limited HF frequency, and only above 10 MHz. This is not the type of behavior we would like to see.



To complete this series, we also have measured the reflection characteristics, as shown in figure 11.




Figure 11. Reflection of a trifilar voltage balun, one side grounded



Again in figure 11 we find a component that will not qualify in an antenna system and for which reflection is un-acceptably high at all frequencies. Note the vertical axes of the diagram has been changed with respect to earlier and comparable diagrams to show anything at all on this voltage transformer.


This 'weird' behavior may be better understood when looking again at figure 6; by grounding one of the output terminals the voltage transformer is partly shortened out. The transfer characteristics may be affected less because of capacitive coupling 'across' the transformer.


When adding all up it is obvious this voltage transformer is not the ideal component (to say the least) to connect a balanced antenna to an unbalanced transmission-line and it is surprising (this type of) voltage baluns are still to be found in some antenna system and / or be recommended in magazines for this application.



Some conclusion about 1 : 1 baluns on ferrite cores


From the above it may be concluded that a current balun in general is the component of choice to connect a symmetrical antenna to an a-symmetrical feed-line. When constructing this current balun, any ferrite material from the junk-box will do. When applying a 36 mm. ferrite toroide, ten turns will be sufficient when applying 4C65 material and 6 - 8 turns are optimal for other (less HF) type of material. For this type of current baluns, more turns usually will make a lesser quality component since increasing parasitic parallel capacity will decrease maximum usable frequency.

Since the 1 : 1 current balun is based on isolating the input from the output terminals, the same quality may be applied to good use to stop undesired HF currents to arrive at the transceiver. It's therefore good practice to also apply this current balun / sleeve choke at the other side of the feed-line at the receiver or at any other electrical appliance that is picking up HF currents (audio / video equipment).


Voltage baluns should be avoided around balanced antenna systems because of sensitivity of these components to any type of unbalancing (trifilar balun) or because of diminishing internal coupling (other flux-transformer types).



Current baluns on powder-iron (carbonyl) cores


So far we discussed sleeve chokes / current baluns in the form of larger coils 'on air' or on ferrite cores.

Sometimes also baluns on powder-iron / carbonyl cores are being offered, which should be regarded with some care. To compare with ferrite materials, let's look at the 36 mm. toroide shape. Using 4C65 ferrite with an initial permeability of 125, we needed 10 turns to make an inductance of 16 μH, that will provide a parallel inductance of 200 Ohm at 2 MHz. When we were to apply e.g. Amidon grade 2 carbonyl core (permeability = 10), of the same shape and dimensions, we would require 3,5 x as many turns (√(125 / 10)) to obtain the same impedance. Since the center core hole is 22 mm. 35 turns of RG58 will not pass and even using miniature (teflon) coaxial cable will be challenging, next to boosting parasitic capacitance to an un-acceptable level.

For current baluns / sleeve chokes, powder iron / carbonyl type of materials are less suitable materials.


Please contact me for your remarks and suggestions at:


Bob J. van Donselaar