outputs of the R1 boards. See note 2 for more details.
Grounding the node labeled "SB" in Fig 5 switches the receiver from lower sideband to upper sideband reception by changing the phase of ICO by 180 degrees. The output labeled "LPCLK" in Fig 5, is not used in the receiver. This output is used to clock active filters in the exciter in note 2.
Mixers U8 and U9 are CMOS analog multiplexers. The digital local-oscillator signals drive the select inputs of U8 and U9. The 1477-Hz cycle is divided into six phases, numbered 1 through 6, in time order. I'll describe the operation of U8; U9 operates similarly.
During phases 1 and 2, multiplexer output 0 is selected, connecting U8 pin 3 to U8 pin 1 and leaving pin 2 open. During phases 4 and 5, multiplexer output 2 is selected, connecting U8 pin 3 to U8 pin 2 and leaving pin 1 open. During phases 3 and 6, pin 1 and pin 2 are both open.
The two op-amps of U31 sum the mixer outputs. Resistors R31, R32, R33, R34, R35, and R36 must be well matched as they determine the balance of the audio double-balanced mixers. I used 1% resistors here.
Op-amp U31B, with R35 and R36, forms a current mirror. The current from R36 into U31A pin 2 equals (minus) the current, from R31 and R33, into U31 pin 6.
Op-amp U31A forms an inverting unity-gain low-pass filter stage. Q31 provides a current boost so the output can drive low-impedance headphones or a small speaker. R39 provides a dc current into U31 pin 2 so Q31 runs class A and draws power only from the o-V supply.
Noise from the +5-V supply will enter the audio signal through R39. My supply is quiet, so I have no problem, but a noisy supply may need filtering.
The audio signal at the emitter of Q31 equals the voltage on U8 pin 1, minus the voltage on U8 pin 2, plus the voltage on U9 pin 1, minus the voltage on U9 pin 2. This provides the double balance for the mixer action of U8 and U9.
All of the receiver's gain is before audio mixers U8 and U9, to minimize the audibility of the 1477-Hz tone from mismatches in mixers U8 and U9.
The receiver works well; clean sig ials are very intelligible. Some signals, of course, are transmitted distorted, and some are buried in noise or interference. Audio fidelity is not as good as that of a receiver with a wider passband.
The in-passband spurious responses show up as frequency-inverted speech. These are noticeable when listening to a CW carrier, but are hardly noticeable when listening to speech, as W1PNB reported for a Weaver exciter.6 Tweaking the gain ratio on one R1 board versus the other would reduce the level of inverted speech. The 1477-Hz tone is barely audible at full gain. The skirts are excellent, and the passband sounds just wide enough for speech.
1 Weaver, D.K., "A Third Method of Generation and Detection of Single-Sideband Sig nals," Proceedings of the IRE, Dec 1956. Note that this whole issue was devoted to single sideband!
2 Anderson, P.T., "A Different Weave of SSB Exciter," QEX, Aug 1991, pp 3-9. Note error in Fig 7: a 422Q-Q resistor should be inserted in the wire between U11 pin 7 and U11 pin 2.
3 Campbell, R., "High-Performance Direct-Conversion Fteceivers," QST, Aug 1992, pp 19-28.
4 The Radio Amateur's Handbook, 1981. pg 14-33 (ARRL: Newington, CT).
5 Anderson, P.T., "Transistorizing Surplus VFOs," QST, Feb 1989, pp 45-46.
6 Wright, H.F., "The Third Method of SSB, How It Works in Theory and Practice," QST, Sept 1957, pp 11-15. □□
Was this article helpful?