0
$\begingroup$

We are optimizing an impedance match with s2p models for an LNA and found that an "L" match with a serial cap and shunt inductor will optimize to provide +3 dB of output, versus using serial cap with shunt cap.

But then we noticed the output side has a DC bias for the drain, and clearly that cannot be shorted to ground through a shunt inductor!

So the drawing below works (cubes are capacitors), but we can get a better optimization if component "S6" were an inductor.

Is there a way to arrange the components so I can use a shunt inductor as component "S6" in the match but block DC to ground?

L impedance match with bias tee

$\endgroup$

1 Answer 1

1
$\begingroup$

You could put a large capacitor in series with your choice of inductor. It needs to be big enough to have a low impedance at your minimum frequency.

schematic

simulate this circuit – Schematic created using CircuitLab

With large capacitors in parallel, watch out for power-on transients that can bias the MMIC incorrectly for a short time. Do a transient simulation as well as an AC simulation.

$\endgroup$
2
  • $\begingroup$ Thanks! Is there a difference between the order of placing the capacitor or inductor closer to the RF line? To me placing the inductor closest to the RF line the way you have drawn it makes sense in terms of organization, but is there a difference or reason to choose one over the other? $\endgroup$
    – KJ7LNW
    Apr 26 at 23:21
  • 1
    $\begingroup$ Yes I think the inductor first is probably best, You will be using it to provide a fairly high impedance in parallel, probably hundreds of ohms. If you extend the line out to one side with a capacitor, there's additional capacitance to ground which will change the effective impedance of the inductor. Inductors are not perfect either, this is another whole complex subject. The best high frequency inductors are conical ones, for the same reason - they start on the "live" side with an incredibly small 0.1 mm coil and get bigger to deliver the bulk inductance. $\endgroup$
    – tomnexus
    Apr 27 at 5:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .