10GHz filter / probes - an update

Andy G4JNT

Using short probes from SMA spills sticking about 3mm into the first and last cavities at 5mm spacing from the end plates , the modified 'JVL filter would tune up nicely just using the maximum smoke technique.  The bandwidth at 3dB points was around 15MHz (ignore values in previous post, I typed that wrong)
That is quite narrow and considering the insertion loss is only 3 - 4dB is actually pretty impressive.    ALSO, a loss like that means maximum smoke tuning is permissible - it doesn't work too well with lossless filters

Anyway, to try for a wider bandwidth I increased the probe length, and with the same spacing from the end wall  the 17mm cavity could no longer be brought onto resonance.  the probes wer eloading / detuning it too much.   So for that filter I returned to the original short probes with their low value of coupling and accepted its 15MHz bandwidth and 4dB insertion loss. 
Which is a lot narrower than I intended, but nice to have on the shelf .

The conclusion is probably that the probe coupling adds cavity tuning / loading that the in/out irises on the original 'JVL didn't.   The original iris coupling is based on shunt susceptance and shorts across waveguide are inductive.  My coupling probes are capacitive, and entering into the E field of the cavity, so a totally different effect on the tuning.  Like using coupling L or C on a lumped filter gives different resonator values

So to use the JVL route with simple probes in and out requires shortening of the first and last cavities, or a loop, or some sort of coupling that doesn't add a strong reactive element.

I may make another set of cavities if inclined to experiment, but as the filter I do have does a good enough job - even if it is a lot narrower than intended - enough said ..    A single conversion 10GHz to 28MHz IF ... I wonder...

Andy  G4JNT

Join UKMicrowaves@groups.io to automatically receive all group messages.