====== Gigabit PoE (802.3af) ======

  * http://en.wikipedia.org/wiki/Power_over_Ethernet#Power_capacity_limits
  * http://forum.mikrotik.com/download/file.php?id=10417&sid=6fc2e8eb2c2daca423ee4b91e19a1405
  * http://www.dipolnet.cz/poe_-_iee_802_3af_and_iee_802_3at_bib746.htm

  * **C = 100nF** (ceramic, stacked tin foil,...)
  * **L = 300uH**
    * Ferrite material 43 or 61 (toroid FT-37-61)
    * https://www.youtube.com/watch?v=t8mB6xz2NPI

{{http://www.valal.sk/upload/savefile_php/uploads/eb025411ea.png}}
{{http://www.infocellar.com/networks/new-tech/PoE/cable2.gif}}
{{http://electronicdesign.com/site-files/electronicdesign.com/files/uploads/2014/01/fig1.gif}}

===== Experiments =====

  * I've decoupled all of 8 gigabit ethernet conductors using 100nF ceramic capacitors and the link was still OK.
  * I've shorted + and - of 3 meters 100M ethernet pair through FT-50-61 toroid choke. For 4 and more windings it did not affected the link. When removing the winds the link quality was getting gradually worse.
  * I've put FT-50-61 toroid choke on + of the same pair and from 9-10 winds and up it prevented the ethernet from linking. When trying less winds the link quality was gradually better.
  * It's not good idea to use wire with 0.2mm diameter for more than 3A (too much voltage drop, too much heat). So probably 2A should be maximum.