So in moving this forward some, i decided to knock up a 4th order filter based on the Dishal parameters, and then plotting out the filter to see how well theory and reality collide.

These were the settings i used in dishal based on previous experiments.

This is the filter realized, 2nf were used for the 1984pf and parallel 1nf and 580pf for the 1708pf capacitors.

Test rig was scope on output and function gen on the input, i did not worry about impedance matching 50ohm to 150ohm.

Measurements were taken every 200hz and plotted in a spreadsheet. A couple of things, firstly there are 2 peaks not sure why, second major one is the filter is 2K wide not 400hz wide, might almost work as a narrow SSB filter HIHI and the third major issue is the center freq is even lower than predicted at about 442 -> 443 Khz, no where near the 447 predicted and still miles away from 455Khz where the radio’s IF is.

This has been a very useful experiment, I think i can improve on this first effort by a long margin and I intend to go back the the drawing board and see if i can get more accurate motional parameters and also if i can pull the frequency UP somewhat with added capacitance to get this thing closer to 455Khz.

Ok, so as an addendum to this, i terminated the filter at 150 ohm, turned the power on the function gen right down to its lowest setting and this is how the filter now looks.

Hi Robert,

I just stumbled over your web page and found your real interesting experiments with ceramic ladder filters. My interpretation about the differences between the calculations and your measurements is that your assumed resonator parameters are far away from the real values (a wild guess because lack of more details).

Here some thoughts you may try out:

1) The parallel capacitance Cp of only 6pF appears to be much lower than all specs I have seen so far for these resonators. Cp is normally in the range of 100 to 200pF for ceramic resonators at 455kHz. A measurement of Cp is important.

2) If you change Cp just for fun from 6pF to 100pF in the xtal evaluation program , you will get a much lower Lm of approx. 20mH. (BTW, such an evaluation prog. is also contained in Dishal in the “Xtal/G3UUR-Method” menu).

3) A filter calculation with your cap values now reveals about the same bandwidth of 2 to 2.2Khz that you measured. The rounding of the curve maybe due to a low Q of the resonators and/or frequency differences.

Unfortunately, there is no chance to “pull” the center frequency to 455kHz, because the parallel resonance is too low (estimate <460kHz). There may be some way to overcome this (e.g. compensation of Cp), but this complicates the whole tremendously.

So, this is all guess work from my side which hopefully may help you a bit. Keep experimenting ;-))).

73 and a Happy New Year, Horst

Hi Horst, thanks for your input, I do appreciate your comments and thoughts, I will give your ideas a go later and see where they lead me. I did not know where to start with Cp and did not know how much of an effect it had on the filter. It was one of those things i sort of put on the back burner but i knew i would have to come back to it eventually to get make the filter much more accurate.