The article The oft asked question of how much an LNA improves a 70cm weak signal station solicited some comment on optimal configurations. This article deals with the notion / Rule of Thumb that optimal LNA gain is just sufficient to offset line losses.
This article explains with graphs the relationship between Signal / Noise degradation (see Signal to noise degradation (SND) concept) and LNA gain in the configurations discussed in the original article. See The oft asked question of how much an LNA improves a 70cm weak signal station for documentation of the scenario assumptions.
The critical value for SND is a personal choice, but for the purpose of this discussion, let's choose 1dB. That is to say that the S/N at the receiver output is less than 1dB lower than the ultimate that could achieved with the antenna system given the external noise environment.
The total line loss in the example configurations was 2.6dB. The model assumes that LNA Noise Figure is independent of LNA Gain, though in the real world, there is typically some small dependence.
Often the choice of LNA Gain drives the choice of a single stage or two stage LNA, which has cost implications.
Terrestrial external noise – 495+5K
Above is a chart showing SND vs LNA Gain. It can be seen that as LNA gain is increased, SND improves rapidly with a knee around 15dB LNA gain above which SND improvement is slower.
For SND lower than 1dB, one would choose LNA gain of more than 12dB for this scenario.
Satellite external noise – 40K
Above is a chart showing SND vs LNA Gain. It can be seen that as LNA gain is increased, SND improves rapidly with a knee around 20dB LNA gain above which SND improvement is slower.
For SND lower than 1dB, one would choose LNA gain of more than 18dB for this scenario.
Very high terrestrial external noise – 5000K
Above is a chart showing SND vs LNA Gain. It can be seen that as LNA gain is increased, SND improves rapidly with a knee around 5dB LNA gain above which SND improvement is slower.
For SND lower than 1dB, one would choose LNA gain of more than 2dB for this scenario.
The bigger question for this scenario is whether there is worthwhile benefit in using an LNA at all. If very low LNA Gain is sufficient, no LNA is probably nearly as good… it is just not needed and likely to have significant downsides with negligible benefit.
What does it mean?
Even for this simple practical scenario where all else is held constant, the SND response to changing LNA Gain depends on the external noise level, and the response is not a simple linear one, but one with a distinct knee.
It is not safe to simply use as much gain as possible / available as high gain amplifiers are more prone to IMD, and more likely to overload the following receiver creating IMD.
Excessive gain is not a safe solution.
In high noise environments it takes little effort to achieve low SND, in fact an LNA is probably not warranted.
At the other extreme, the very external low noise level of a satellite path means that a suitable LNA is vital to low SND, and it probably needs high gain to achieve low SND, even for an LNA with very very low Noise Figure.
It is complicated, and it is a multi-dimensional problem… and that is where the G/T spreadsheet may assist.
Rules of Thumb may not be reliable.
Be wary of over simplified rules, this is a complicated problem for linear systems, then candidate solutions need to be tested carefully to see whether they are unduly affected by IMD etc.
Since the kind of improvements made to achieve a high performance weak signal receive system tend to be expensive, a little time to make measurements and on desk studies to understand the problem and design a solution may be a good investment.