Tuning Interdigital Filters - Dishal's Method
时间：04-04
整理：3721RD
点击：
I've designed a number of 3 pole Interdigital Filters, which I have implemented using round rod / slabline construction.
The target is a filter with a Butterworth response, a centre frequency of 1275MHz, a Bandwidth of 200MHz and conventional 50ohm input and output taps. I've been tuning these 'by eye' - tweaking the filter shape and passband ripple by sweeping the filter to display amplitude across an appropriate frequency range. The filters looked OK, but where I should have had a flat passband response, I had 0.6dB ripple. I've been on a mission to get these filters much closer to the pass band design criteria.
I came across Dishal's method, and yippee, I had the first filter 'on the nose' very quickly, which also revealed that the Input Impedance was 40 + J0 (Exactly). So I jacked up the tapping point around 10%.
The first resonator came to resonance very quickly, the second one set at 90degrees and the input impedance was reported as 2 + J51. Highly delighted at this point.
The third resonator now refuses to come to resonance. The phase increments from 90 degrees to around 120 degrees then as I continue to back out the tuning screws the phase falls back, there is no tuning position where the phase can be set +/- 90 deg to the middle resonator.
If I reverse the filter (which is symetrical) the first two resonators tune properly and the third will not. (The third resonator is the last in the sequence, which tuned perfectly when it was the first resonator in the sequence.
I've tried tuning the filter low @ 1250MHz and high at 1320MHz, to see if rod length is an issue, but I get exactly the same results.
ANy ideas what is going on? - What can I try next?, I've checked calibration, cables, connectors, used different lengths of cable to flush out measurement issues.
My only thought and a dumb one at that, is to increase the length of the end rods by 1-2%, but it's the only parameter left to change.
So any guidance, gratefully received!
Regards
Steve
The target is a filter with a Butterworth response, a centre frequency of 1275MHz, a Bandwidth of 200MHz and conventional 50ohm input and output taps. I've been tuning these 'by eye' - tweaking the filter shape and passband ripple by sweeping the filter to display amplitude across an appropriate frequency range. The filters looked OK, but where I should have had a flat passband response, I had 0.6dB ripple. I've been on a mission to get these filters much closer to the pass band design criteria.
I came across Dishal's method, and yippee, I had the first filter 'on the nose' very quickly, which also revealed that the Input Impedance was 40 + J0 (Exactly). So I jacked up the tapping point around 10%.
The first resonator came to resonance very quickly, the second one set at 90degrees and the input impedance was reported as 2 + J51. Highly delighted at this point.
The third resonator now refuses to come to resonance. The phase increments from 90 degrees to around 120 degrees then as I continue to back out the tuning screws the phase falls back, there is no tuning position where the phase can be set +/- 90 deg to the middle resonator.
If I reverse the filter (which is symetrical) the first two resonators tune properly and the third will not. (The third resonator is the last in the sequence, which tuned perfectly when it was the first resonator in the sequence.
I've tried tuning the filter low @ 1250MHz and high at 1320MHz, to see if rod length is an issue, but I get exactly the same results.
ANy ideas what is going on? - What can I try next?, I've checked calibration, cables, connectors, used different lengths of cable to flush out measurement issues.
My only thought and a dumb one at that, is to increase the length of the end rods by 1-2%, but it's the only parameter left to change.
So any guidance, gratefully received!
Regards
Steve