That's in fact a different problem. You already have a variable gain element and want to achieve a better level constancy.
In this case, you are asking for frequency independent level detector operation, the right node to measure the actual output level and control loop accuracy.
By the way, you splitted your questions related to this circuit into at least 4 different threads. Why?
I thought this way would be easier for the searchers of the forum to find the specific information after a long time, based on the topic. If you suggest me not to do it, I won't. Thanks.
Now about the circuit, would it worth it to try different transistors than the bf494 so see if I get a flatter response above 16mhz?
It is not a problem of the discrete mmic amplifiers as I have tested the response at the output of the second bf256.
Frequency mixers requiring +7dbm LO, can still mix at half power with increased loss and at more power maybe, so that is not much of a problem, but It would me nice to achieve a flat response also on the upper bands.
Also, I am not sure if the response has to do with the coil types or not...
By the way, I have also tried to switch different coils electronically. The addition of a diode in the before or after the coil really distorts much the signal. I have thought of another approach, which I need to try. Use a PNP transistor between vcc and each coil and control their base electronically. Clean and stable DC will pass through the transistor if it is enabled, to the coil. I do not know if this will cause any distortion to the waveform, but it behaves like another psu regulator (like the BD436), so I do not see why this should cause any problem there.
I have also tried to replace the capacitor with a bb212 with no success. Much distorted waveform again. Something to try next...