If I turn the AGC off in HDSDR it still appears to be on when I open up the plug in. Likewise if I turn the AGC back on in the plug- in it still appears off when I go back to HDSDR.
Similarly changing the AGC threshold in HDSDR (with AGC on in both HDSDR and the plug-in (which I assume means the LNA gain reduction threshold)does not seem to interact with the settings in the plug-in.
Any advice on what is happening here. I am focused on using the RSP in the 0-60 MHz range.
By and large the RSP is working very well. This is just a quibble.
These two AGC loops don't interact directly, they are two separate loops. The AGC within the ExtIO window is used to ensure that the signal level at the IF output doesn't overload the ADCs. This particular AGC loop works on the entire composite signal power defined by the IF bandwidth, which might well contain multiple signals. The AGC loop within HDSDR will act on the individual signal selected via the software.
You may not get optimum performance when using AGC within the ExtIO window. It is really only a rough guide to the gain settings that you might want to use. The best strategy is to use this AGC loop to get the gain roughly to where you want it and then turn AGC off and then fine tune the gain manually for best SNR. An alternative approach is to leave the AGC on, but adjust the ADC setpoint. The setpoint is the level below the ADC full-scale that the RMS composite signal is set to. The default is -30 dBfs, but you may find improvements by reducing this to -40 dBfs.
Reason: No reason
This leads me to a couple more questions.
1 In HDSDR what exactly does the S-meter indicate? Is it the magnitude (rms) of the total signal after the ADC (i.e. everything) or the just size of the tuned signal after the demodulator?
2 Why do you recommend such low set points for the ADC (x30 -x100 down from full scale). I am particularly curious given that the set point includes everything i.e. wanted and wanted signals. Is it linearity, proofing against transient or something else?
Thanks for your help
Unfortunately, we are not experts on HDSDR, so we don't know for certain whether the S-meter operates on the full signal or only the signal selected by HDSDR. It would seem logical that it ought to act only on the selected signal. What I can say is that we believe to get meaningful values from the S-meter, it is necessary to calibrate if first. if you don't know how to do this, try searching this or the independent Facebook forum, or posting a note on this subject and there is a good chance that you will find someone who can help you.
Regarding the ADC setpoint, the signal level measurement performed is after a set of digital filters that follow the ADCs. This means that the composite signal power at the ADC inputs will be higher than that measured at the output of the digital filters. Secondly, you have to consider the difference between the peak signal and the RMS signal, which in a multi-channel environment can be very large. Add to this the effects of fading and the fact that the wanted signal may have a significant peak to average ratio itself and the AGC loop is necessarily slow and you find that you need a good level of headroom on the ADCs. The key thing is to ensure that there is no 'clipping of the ADCs on the signal peaks.
There is no simple answer as to what this level of headroom should be though as for example, if you are receiving a VHF broadcast FM signal with an IF bandwidth of 200 KHz, you can afford to run less headroom than if you are using an 8MHz bandwidth and the HF bands.
The final consideration that might mean the gain needs to be reduced even more is if there is a very large out of band signal that might be causing compression in the LNA/mixer. This would not be detected by AGC loop and so the to obtain acceptable performance, it may be necessary to reduce the LNA gain, but this can only be determined by trial and error.
Reason: No reason