Hi Bill
I tend to agree with Dallas, however the result you achieved is outstanding, the only issues I have is the temperature and mechanical stability, I note that you are addressing this by constructing using adjustable capacitors, that you did not achieve the desired result.
You have to allow for stray coupling something that you cant visualise and will very with layout.
In the industry when required to design built a high Q filter or preselector that provided both mechanical and temperature stability on the higher frequencies (VHF UHF the window coupling technique is used ) utilizing Parallel LC, usually 4 to 6 cells achieving the passband required, the 50 Ohm match using a C divider. This method of design allows variable passband utilizing stagger tuning, the return loss off memory is less than 3 dB
The stability is achieved by mechanical construction.
Both Motorola and Philips used this technique, with Philips on PCB construction the coil shield (Can) would have a slot that aligned with the next coil.
Not so good in my mind, Motorola would use a extruded alloy block, RCA and GE radios used the same as Motorola.
BTW meant to say that ripple is not necessarily due to filter ringing, but bad matching from generator to load in that the return loss current interferes with the main frequency time domain waveform.