somewhere I have seen that it is possible to build a Balun by using a transformer in the following way:

enter image description here

The explanation I have found is that the fact that the ground is only at the primary coil will make that the signal at the primary coil is unbalanced with respect to ground (and I understand this), while the signals at the terminals in the secondary circuit are balanced with respect to ground. I do not understand the reason of this.

  • 1
    \$\begingroup\$ there will be winding-winding capacitances, those may or may NOT be balanced. \$\endgroup\$ – analogsystemsrf May 4 '19 at 3:58

the signals at the terminals in the secondary circuit are balanced with respect to ground.

They are balanced, but not with respect to ground. They won't be referenced to ground at all, except through whatever parasitics couple them to ground.

If you want them balanced with a common mode voltage at ground, you'd want to either use a center-tapped secondary, or terminate the balun outputs with a center-tapped structure.

  • \$\begingroup\$ Precisely what do you mean with "They are balanced, but not with respect to ground"? Which are their values? \$\endgroup\$ – Kinka-Byo May 3 '19 at 17:24
  • 2
    \$\begingroup\$ The output voltage relative to ground could be anything. There's nothing in the idealized model that determines a particular voltage relative to ground. It could be -1000 V, +1000 V, or anywhwere in between. In the real world there will be, like I mentioned, parasitics that will determine the actual voltage from the output wires to ground, but you can't predict from what you know now what it will end up being. \$\endgroup\$ – The Photon May 3 '19 at 17:39
  • \$\begingroup\$ Ok I have understood that there is not any link between their value and the ground. But, what does balance mean in this case? \$\endgroup\$ – Kinka-Byo May 3 '19 at 18:50
  • \$\begingroup\$ It means that relative to whatever their common mode voltage is, both outputs are varying with equal and opposite waveforms. \$\endgroup\$ – The Photon May 3 '19 at 19:44
  • 2
    \$\begingroup\$ By virtue of symmetry, the secondary terminal voltages are balanced with respect to the midpoint of the transformer secondary winding. Remember that "voltage" is shorthand for "potential difference;" voltage has to be measured between two points. In this case, the terminal voltages are equal and opposite when measured from the middle of the secondary winding. If you grounded the middle of the secondary winding - i.e., the "center tap" - the voltages on the secondary terminals would be balanced with respect to ground. \$\endgroup\$ – Brian K1LI May 4 '19 at 0:30

enter image description here

The secondary impedance is only as balanced as the load with respect to some other impedance such as ground or freespace. It does not depend on the turns ratio. But stray capacitance coupling of windings must be also balanced to the primary ground at higher frequencies.

Here is another method for RF which works as a tri-filar winding on a core. The inductive impedance determines the usable range in frequency 1~3 decades, which is true for all isolated transformers.


enter image description here

For RF, Current Baluns are spec'd by common-mode attenuation at some impedance. e.g. 25dB min 10MHz to 100 MHz, or 1MHz to 30MHz or...

"Voltage Baluns" used in RF are called "3dB splitters" and are measured by BW and isolation between split ports in both 50 and 75 Ohm varieties.

Split power loss is -3.5dB, as no one yet has ever made one for RF that is lossless. They use a tiny ferrite 'hybrid transformer" with 2R termination to ground.

enter image description here

This is what I call a Voltage Balun.

  • \$\begingroup\$ The only way current can be opposite polarity to voltage is in parallel LC resonance FYI @joribama -1 \$\endgroup\$ – Tony Stewart EE75 May 3 '19 at 23:00
  • \$\begingroup\$ You are right I am wrong. I was missing that source current is opposing \$\endgroup\$ – Tony Stewart EE75 May 3 '19 at 23:56
  • \$\begingroup\$ OK, I thought I was going mad. :^) \$\endgroup\$ – joribama May 3 '19 at 23:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.