Having noticed some serious performance issues with the first and second mixer filter stages, I identified the problem to be poor design choices and a lack of gain in the first IF stage.
I have split the first IF mixer and filter and added a dual gate mosfet variable gain amplifier on the output side of the 10.7MHz roofing filter, put the 455KHz mixer and SSB filter on a separate PCB and the difference in performance and quality was instantly noticeable.
The noise floor is much quieter and the the gain distribution is much more even across the IF strip, strong signals are no longer raspy and the AGC seems to respond better on strong signals.
I also optimised the PCB layout to make the overall size around 30% smaller than the first prototype front end stage.
The added benefit of the variable gain first IF amplifiereamplifier means that I have two independent IF gain controls, allowing my to optimise the radio at runtime – I can also control the first IF amplifier with the AGC line by adding a small diode in series with AGC input line.
Here are a couple of pictures showing the new design.