EMRFD Message Archive 15418

Message Date From Subject
15418 2019-09-04 17:18:11 mhadley157@gmail.... Bandpass filters swr
Why is it virtually impossible to see a good match (return loss) to a designed load across the pass band of a double or triple tuned BPF?

Mike G4JXX
15420 2019-09-04 18:41:22 Ashhar Farhan Re: Bandpass filters swr
The band pass filters are not strongly terminated and need to be terminated at the other end with their designed impedance. A 50 ohm resistor on the other port will make all the difference.

- f

15433 2019-09-05 11:15:04 g4jxx Re: Bandpass filters swr
I am using a design impedance of 50 ohms in and out. My signal source is 50 ohm and output is terminated in my spectrum analyzer 50 ohm input.

When aligning a filter I can get a 1:1 swr at the passband centre frequency only. Adjusting a filter for the required passband (as seen on the analyzer) then shows a poor swr, sometimes as high as 3:1. There is very little insertion loss when correctly adjusted.

I have tried link coupling and split capacitor coupling methods both calculated for 50 ohm in/out but results are the same, high swr.

Double and triple tuned filters are top coupled chebyschev design.

Mike
15439 2019-09-06 02:17:34 Ashhar Farhan Re: Bandpass filters swr
It shows this behaviour when the coupling is not optimal. How you run any simulations?

15441 2019-09-07 19:24:05 mhadley157@gmail.... Re: Bandpass filters swr
I used a triple tuned bpf program which is on the cd that came with emrfd. My problems arose when using iron powder cores and capacitor matching in/out (smd devices). Using an unloaded Q of 250 for frequencies up to 14 MHz I tried to trim the matching for 50 ohm but to no avail.

I have now restyled the filters to use rewound Toko 10k coils with link coupling. Now with an unloaded Q of 100 matching to 50 ohm over the whole of the passband is achievable. Measured insertion loss is no greater than 3db over the range of different filter frequencies.

I have yet to measure the return loss but expect to see quite a change from that projected in the original calculations.

I have not run the design through any simulation program. I have SPICE but I believe it it not best suited to run LC filter simulations.



Sent from my ASUS


-------- Original Message --------
15444 2019-09-07 21:37:13 Ashhar Farhan Re: Bandpass filters swr
Mike,
Using a lower quality inductor to get a broader bandwidth is a little like using a truck to widen your front door.
A redesign will often lead to a better filter. The important thing that using higher Q inductor gives is better out of band rejection and steeper slopes. Q minute of poking around with the emrfs software should get you sorted out.
- f