Andy,
You are correct that it is mostly an emphasis to your description but the points, although subtle, are I think of important general applicable engineering description in the design of all receiver front end amplifiers, low noise (LNA) just being a specialized description of some. My first point is that the first stage is not necessarily existing for the purpose of having anything to do with the bandpass filter. In the special case of the Sawbird, yes the existence of the filter will lower the amplification output of that first stage device somewhat.. However, in all front end RF amplifiers the purpose is the same, that is to establish sufficient meaningful gain to the intended signals in a desired range of frequencies that will allow further detection and manipulation. One of the consequences of that first device is that it establishes not only the initial gain for the rest of the amplifier but also, and just as importantly, it establishes the “noise figure” for the rest of any amplification that takes place afterwards. This is crucial for the designation of an amplifier as “Low Noise”.
As I pointed out, in the special case of the Sawbird with its filter immediately following that first stage is that it is almost irrelevant to the general functionality of the first stage as would be found in all front end amplifiers.
Now in the case of the Sawbird the filter following the first stage does lower the resultant gain of its output, but it is almost inconsequential. That first gain of the first stage may be as much as 10-15 dB’s minimum and then is perhaps lowered by maybe 1-3 dB’s at most by the filter. You still have at least a resultant 11-12 dB minimum gain remaining out from the first stage. In many amplifier designs that is actually what the actual gain is designed.for, but the main goal here is “to establish the lowest noise level possible” from that first stage. You can always make up the lost gain by following stages without worrying too much about the noise as it was mostly established by that first stage.
Finally, the second stage normally found in many, if not most front end amplifiers, even those without a bandpass filter, gives a further signal increase by now establishing another amplification factor of 10-20 dB’s for the remaining amplifier and yes, in the case of the Sawbird here you can make up the at most the 1-3 dB’s lowered from front first stage by the filter.
As far as the coax loss, well that is not taken into account by the design for the LNA as it is an unknown variable to the LNA designer. It is up to the user’s application which is why always, it is recommended to the user to keep the loss to a minimum by utilizing the best possible coax cable and keeping the run as short as possible.
I wrote this first from my understanding over years of pouring over many RF amplifier circuits diagrams, mostly microwave for front end and transmitting, It has always been a fascination to me of their design and operation in microwave receivers and RADAR applications for satellite reception and ham radio use. So I didn’t cheat here. I then did the now thing and asked the Co Pilot AI what it thought. Kind of elated by what it responded to.
All following from CoPilot AI:
Considering insertion loss
I need to figure out the typical insertion loss of a SAW filter between stages. The user wants to know the dB reduction of the first stage output with the filter in place. From what I know, SAW filters usually have an insertion loss of around 1–3 dB in the passband. For example, NooElec mentions a total gain of 40–42 dB, and typical MMIC stages might have around 20 dB each. So, the filter could cause about a -2 dB loss. I’ll keep it concise but still explain the concept. Good question—that “mystery loss” between the two stages is exactly where the SAW earns its keep.
Short version: in the SAWbird H1, the ceramic/SAW bandpass filter between the two LNAs typically costs you on the order of 1–2 dB of gain in the passband, maybe up to ≈3 dB worst case, compared to a design where the first LNA feeds the second directly.
A bit more structure:
- Total module gain:
The SAWbird+ H1/H1m datasheets quote about +40–42 dB gain at 1420 MHz for the whole module (two LNAs + SAW filter). - Typical per‑stage LNA gain:
The MMIC LNAs used in this kind of design are usually in the 18–22 dB gain range each at L‑band. Two such stages in cascade, with no filter in between, would naturally land you around 36–44 dB total. - What that implies for the SAW filter:
Since the measured total gain is still around 40–42 dB, the SAW filter’s insertion loss in the passband must be relatively small, with ≈2 dB being a very plausible “typical” number. - Why that’s acceptable for noise performance:
Because the first LNA has high gain, the effective noise penalty of putting a 1–2 dB loss after it is modest (Friis helps you here). You pay a little in overall NF, but you gain a lot in out‑of‑band rejection and RFI suppression, which for 21 cm work is usually a very good trade.
So, if you imagine a “plain” two‑stage LNA with no filter, then drop in the SAW between stages, you can think of the first stage’s output—as seen by the second stage—as being reduced by roughly 1–2 dB in the 1420 MHz passband, with ≈2 dB a good mental default.
Best,
Adrian