Tuning Interdigital Filters - also Dishal's Method

[Darktrax]

Dishal's method is easily searched, and has been mentioned in this forum before. Essentially, by temporarily short-circuiting all but the first sections of a filter, and then tuning for maxima and nulls while sequentially removing the short-circuits, an exact proper tuning can be done quickly.

The method is very easily applied in simulation, but is also used by deliberately making devices with tuned circuits shorted out, the short being removed during adjustment by laser trimming, or simply cutting.

Direct hand-adjusting without this trick can give apparently near-correct passband results, even though it ends up using a combination of capacitor values that are very different from the ideal. Also, it takes much time, and there is a certain skill in recognizing which maxima and minima to adjust, and how they "tilt".

Current link--> Dishal's Method

What I seek is any detail on a variation of this method where during sweep testing, both the passband and the input reflection are displayed. The filter is connected via a quite long 50-Ohms high quality cable, and a section of "stretchable" line - a variable length coax of telescopic sliding tube. There is no need for direct intervention inside the filter with short (or open) circuits.

The sweep display would show multiple waves and nulls. The method involved setting a marker frequency, adjusting the coax length to bring a null onto the marker, then tuning the filter for some maximum (or minimum). Then the stretchable line would be altered to the next quarter-wave point, and another capacitor tuned. As the final capacitor is tuned, the passband comes up correct in a very striking impressive way.

I would be grateful if anyone has clearer details of this procedure, and how it works - many thanks if you can help.

 

[biff44]

I have seen a paper that uses time domain signals to tune each resonator individually. Try googling on that.

 

[Darktrax]

I have seen a paper that uses time domain signals to tune each resonator individually. Try googling on that.

Good call biff44 - many thanks.
That method features using a display of the time domain response. I am guessing this might be derived from the (reverse?) Fourier Transform of the more usual sweep frequency display, it possibly being a feature of the (expensive!!) kit.

From some direct experience of learning to adjust multi-section microwave filters, it is an extraordinary skill. A Vector Network Analyser has the couplers built-in, but even with scalar displays from a sweep generator, and using a directional coupler to get at the reflection characteristic, it is enough to let you set up many filters perfectly.

When a filter has many tuning and coupling adjustments, it can take a long time unless there is some logical ordered procedure to do it from the beginning. Trying to tweak a nearly fully adjusted filter so easily leads to making the whole thing worse!

If this thread collects all the know-how on how to do this, then I will be pleased and many might benefit.