SW40+ Receiver  Front End
This note describes the circuits between antenna and mixer U1(SA612). Circuit operation during receive as well as during transmit is described since the very large transmit signal modifies circuit operation. The circuitry is divided into functional blocks: a good way of separating function for debugging. Each module serves a different purpose.
RX frontend schematic
PI filter
    The first circuit "module" is the five-element low-pass filter involving C39(470pf), L4(1uH), C38(1000pf), L3(1uH), and
C37(470pf). You should see that this circuit is symmetrical: it looks the same from either end. Indeed, the transmitter propagates its signal from right-to-left out to the antenna, while the receiver uses the same circuit in the other direction. In any case, it is a low-Q filter that attenuates signals above 7MHz, while passing all below 7MHz.
    Ideally, this filter should be terminated in a 50-ohm resistance. Due to its low-Q, this requirement is somewhat relaxed, and we shall see that the receiver doesn't comply here. Circuit function is more appropriate to the transmitter, to attenuate harmonic output. But why not take advantage of its filtering action for the receiver?
T/R switch
    The next circuit "module" blocks most of the transmitter power from overpowering (and destroying) U1. We want most of the transmitted power to go out the antenna, not into the receiver. Yet we want all the received signal to pass through this circuit with little attenuation, into U1. Circuit components here include C40(47pF), D7 - D10, and RFC3(10uH).
 T/R scope waveforms  Consider what happens while transmitting. A very large amplitude waveshape (not a sinewave) appears at C40. This point is common to both receiver and transmitter. That is, when not transmitting, the very small received signal must pass this same point on its way into U1.
    The transmitted signal is of such large amplitude, that D7- D10 conduct 7MHz. current to ground. Current is limited by the impedance of C40...480 ohms at 7MHz. While conducting, these diodes look like low-impedance to ground. However, their forward-bias voltage of 0.6v each mean that there is still a 7MHz. signal across them.  But instead of 30v p-p it is clipped to about 3.3v p-p ("red"scope trace). The RFC3 choke further attenuates the transmitted signal so that the R.F. GAIN pot sees even less ("green" scope trace). However, from there, the voltage amplitude is boosted by T1 (tuned to 7 MHz.) on its way into the mixer due to its step-up turns ratio.
    The critical parameter of the mixer's input transistors is their reverse breakdown voltage between base and emitter. For R.F. transistors, this voltage limit is about three to five volts. Anything less won't cause damage.
    While receiving, the transmitter is idle, and looks like a very small capacitance in parallel with C37. At 7 MHz., antenna signals pass through the PI filter unmodified and appear at C40 with no attenuation. Since the amplitude of received signals is so small, D7-D10 do not conduct, and they appear as a tiny (insignificant) capacitance to ground. This allows C40 and RFC3 to work in concert as a series-tuned circuit resonant at 7 MHz.
    The series resonant circuit sees about 50 ohms on its antenna side (C40) and about 100 ohms on the mixer side (RFC3). So its loaded Q is low: about 3. For such a low Q, exact trimming is unnecessary: resonance is "close enough" to 7MHz. At this frequency, reactance of C40 and RFC3 cancel, leaving an almost direct connection between C37 and the R.F. Gain pot (and to the link winding of T1 too).
    You might ask how large an antenna signal would have to be before D7-D10 began conducting? If amplitude across each diode rises to perhaps 0.6v, diode resistance may start to load the series-tuned circuit of C40/RFC3. A Q of three means that the diodes see three times the antenna amplitude. So that reduces the diode conduction limit to 0.2v. With two diodes in series, maximum peak voltage would double to 0.4v, and we could withstand a 0.8v p-p antenna signal before conduction begins.
7MHz. tuned Transformer (T1)
    The PI filter, and the series-tuned T/R switch  provide inadequate selectivity  to reject unwanted mixer input signals (particularly at the image of 10Mhz). T1 is a tuned bandpass filter that attenuates signals both above and below the 40M band. In addition, it also steps up the low antenna impedance to more closely match mixer input resistance of 3000 ohms. Here are the turns ratio of windings and taps: Let's see how the mixer impedance matching works out...
First, what's the SA612 input impedance? The data sheet says "unbalanced: 1500 ohms". However you should note that neither input (pin 1 nor pin 2) are bypassed to ground. So the input tends to be balanced at twice this value, and the 11-turn winding section of T1 sees 3000 ohms. Assuming that flux linkage within T1 between the 2-turn link and the 11-turn winding is 100%, then the impedance ratio is (11/2)2 or 30.25.  So the 3000-ohm mixer load looks like 99 ohms at the link winding.
    Since the series-tuned T/R switch has nearly zero impedance at 7MHz., the PI filter sees 99 ohms as well. And so does the antenna. Measuring with a noise bridge verifies that at T1's 7MHz. resonance, input Z is indeed 100 ohms. Now this is not a very good match to a 50-ohm antenna system. Why did Dave not use the full 17-turn winding instead? This would give about a 41-ohm impedance at the link...a closer match to a 50-ohm system.
    There are a number of possible reasons: As a matter of fact, you could improve the robustness of the front-end by changing to the 6-turn link section rather than the 11-turn section: connect SA612's pin 1 to the other end of the transformer (requires cutting a P.C. trace). The price you pay is poorer sensitivity: if you have a good antenna, it is a reasonable price to pay.