kees,?

these filters are very fickle about the exact impedances at both ends. i hope you measured them with strong 50 ohms source and termination

- f

On Sat, Mar 18, 2017 at 5:17 PM, Steve Hunt <steve@...> wrote:

Mike,

I did the same using RFSim99, so the shift in the measured frequency response must be down to capacitor tolerances, inductor values being out, or, conceivably, stray capacitance added by the PCB.

It's surprising just how much variation there can be in the inductor values. In addition to the Al tolerance of +/-10%, the winding spacing makes a big difference. In Micrometals application note they show the inductance of 10 turns of #20 wire on a T50-17 varying between 0.22uH and 0.39uH depending on how the turns are spaced!

Also I note in the QRP labs table that at least two different required inductor values are realised with the same number of turns on the same core, so that introduces another "approximation factor".

Steve G3TXQ