Help for Bias tees (20-3000 MHz)

Please, can somebodyhelp me in design of bias tees in the frequency range 20-3000 MHz?I want to use them for an active antenna to power the active element from the coax itself.They may be available in market but I want to know the design.

Making a bias T work over that frequency range is a bit tricky.
There may only be 3 components to worry about; the RF coupling capacitor, RF blocking inductor and grounding capicitor, but getting it all to work !
The coupling cpacitor has to be a low reactance at the lowest frequency of use but still free from parallel resonance at the highest. Not too much of a problem.
The inductor is probably the biggest headache. it must be a high impedance over the full range, probably some sort of multiple inductor or careful choice of ferrite loading and you may get something to work without a problematic resonance.
When it comes down to it it is about keeping all the resonances that will cause a **** out of the wanted signal well out of the frequency range of interest; and the range you are working over is quite problematic if you are looking for high performance.
I've not made a good one that wide. It usually works out cheaper to buy one.

Peter

http://www.picosecond.com/objects/AN-01e.pdf

I've done some microstrip bias-tees for approximately that frequency range.

For capacitor, use 100n, size 0402, X7R ceramics, since it has lower high-frequency losses. For inductors I used a series of two values of Epcos' B82422-A. The one soldered to the microstrip is a 2u2 or 3u3 and the one in series is 100u. You should add a blocking capacitor of 100n, 0805 to the 100u inductance if needed. I've seen no resonances in that freq range and isolation to DC port of about 20 dB.

Hope this helps!