I know nothing about the dynamic range, linearity or overload capability of the mixer, is it useful or possible for you to specify an optimum received signal strength or a range of levels in dBm that the mixer is happiest with?
Assuming there is not much internal noise level generated by the LNA or AGC would it be best (in a crowded HF ham band for example) to operate with an eye to reducing the broadband gain of the LNA and push up the narrow band gain of the AGC?
If the answer is yes, would it be better to alter the default set point of the AGC loop or go full manual? I've never played with the AGC settings in SDRuno because I only know enough about AGC curves to get myself in trouble. I guess the answer is going to be dependent on too many other variables like the number and relative strength of signals in the passband.
Now for a really stupid question, given that the zero offset DC mix product is visible in the receiver, presumably it's not part of the AGC loop?
Reason: No reason
There is no simple answer to your first question as it really depends upon whether the signals sit within the bandwidth of the pre-selection filters or not. In addition, signals that are outside of the IF bandwidth are no visible anyway, so thwere is no way of knowing their signal strength. We do publish detailed technical information for our units that gives IP3 performance at various gain settings, so it is possible to calculate IMD3 product levels for a given interference level within the bandwidth of the pre-selection filters.
The LNA will add very little noise when operating in the HF bands at the higher gain settings. When the RF gain is turned down to the lower gain settings, the LNA is in fact turned off and so noise from other parts of the RF stages and IF stages does start to become significant. The IF stages contribute very little noise and it is possible to reduce the IF gain considerably before you see any measurable impact on the receiver NF. In fact in the HF bands, I doubt even with the IF gain at minimum, you would see any loss in sensitivity. You will not see any IMD3 or IMD2 products from the IF stages until the ADC starts to clip at that point, the products will rise VERY rapidly. You can try adjusting the set point to see how you get on, but the AGC loop works on the RMS power and clipping occurs on the peaks of the composite signal within the IF bandwidth. So what matters is the peak-average ratio of the composite signals within the IF bandwidth. In our experience a set point somewhere between -20 and -30 dBFs is about right, but it does depend on how 'busy' the bands are.
Regarding your final question, the DC offset correction sits within the AGC loop, so the AGC will not respond to DC. The reason for this is that the level of DC offset can be non-monotonic with IF gain setting and so allowing the AGC loop to respond to the DC offset level could cause problem with loop stability.
Reason: No reason