SW40+ Audio Circuits A general-purpose dual operational amplifier (NE5532) is used to amplify audio to drive headphones. Most of the rig's gain occurs here. These audio circuits are detailed in many texts of op-amp applications. This note concentrates on aspects unique to the SW40+ application of these common circuits. Once again, circuitry is described in blocks: separated into power amp, mute gate and differential preamp.

Audio Output Amplifier
One of the two NE5532 op-amps is configured as a band-pass filter, and can drive headphones to a fairly high sound level. Its highest gain of 33 occurs at a frequency of 816 Hertz. Q is about three, giving a bandwidth of 270 Hertz. This bandwidth complements that of the crystal filter. It also reduces wideband hiss resulting from high audio gain that might be otherwise irksome. The components that determine frequency response and gain are R10(22k), R13(1MEG), C25(820pf) and C26(.0022uf).

audio frequency response Because the output at pin 7 is at a DC potential of +8v, audio must be coupled to headphones through a capacitor, C27(47uf). A "ballast" resistor is included, R14(10 ohms), to prevent strange reactive headphones from giving the op-amp a hard time.
This amplifier has a very low output driving impedance (less than one ohm), and will try to pump current into low-Z phones (or an inadvertent short-circuit) until internal current-limiting circuits kick in (maximum load current is 38 ma). Current limiting "clips" the positive-going and negative-going audio peaks, resulting in distorted, harsh audio.
The RF Gain control should be set so that audio level is below the current-limiting threshold. Set this way, current-limiting acts very nicely to clip the odd noise peak. You could call this a poor man's noise limiter.

With no feedback, this amplifier would have very high gain (over 10,000). The four feedback components mentioned above limit the maximum gain to 33. Don't think that the extra gain is "thrown away",it is diverted to other purposes:

distortion is reduced
output is very low resistance, able to drive any load
input impedance is linear and predictable
gain & frequency response is determined by passive (external) components
operation is independent of supply voltage, bias voltage and temperature

Reducing excess gain with feedback is a powerful way to improve desired amplifier characteristics - most of the amplifiers in this rig (RF as well as audio) use this technique. You can thank feedback for making sure the theoretical and measured frequency response shown above agree so closely.

The Mute Gate

Q1 is a switch. When it is "open", its impedance from source to drain is very high (many megohms). A small fraction of audio is allowed to leak from U4a to U4b through R9(4.7MEG) so that a "sidetone" can be heard. When it is "closed", the FET appears as a small resistance (roughly 100 ohms from drain to source), and audio is conducted through it from U4a(pin 7) to R10. You should be able to see that R9 and FET switch are effectively in parallel for audio signals.

Why is this FET not an amplifier? Its drain and source are at the same DC voltage. This means that there is no standing DC bias current: the only current passing through this FET is due to AC signals.
But doesn't the AC signal appear between source and gate, as in an amplifier? Consider the DC potential there as well... When the FET is "open" a full -8 volt DC potential on the gate overpowers any small AC signal on the source -nothing gets through. When the FET is "closed", the gate and source are connected together through R8(1MEG) keeping the gate at the same DC potential as the source. However, C24(0.1uf) bypasses the gate, so yes, audio signals on the source could affect the FET-switch operation. However, as long as AC signals are small (less than a few hundred millivolts), resistance between source and drain remains linear, and insignificantly small. You can't use a gate like this at high audio levels (at headphone levels, for instance) because of the requirement of small drain-to-source AC voltage.

The capacitor C24(0.1uf) is required to keep the FET gate in its open state after transmitting for a few hundred milliseconds to allow the receiver to "recover" and not pass an audible thump on key-release. C24's voltage is dumped very quickly through D5 to cut off audio very quickly, preventing an audible thump on key-down. Sequencing transmit/receive switch over this way is critical for seamless audio with no clicks or thumps.

Audio Preamp, U4a

U4a is configured as a differential amplifier. A regular amplifier's input uses ground as its voltage reference. A differential amplifier has two inputs (neither one grounded): output is proportional to the difference between these two inputs. A good differential amplifier ignores any signals that are common to the two inputs, only amplifying differences. In this case, differential inputs are at R2(10K) and R3(10K).

    This kind of amplifier is more complex, requiring more parts than a simple amplifier. Why did Dave choose it?
    The product detector U3 provides two opposing-polarity outputs ideally suited for differential amplification. This means that audio amplitude is effectively twice as big. With so much audio gain following, audio hiss due to op-amp noise is significant. We should take advantage of all the signal available: a single-ended amplifier could only use one or the other output of U3, not both.
    There is a more subtle reason for using the differential configuration that also involves noise. U2 is a rather noisy voltage regulator - while its output is a constant +8v, AC variations (at 800Hz) are significant. So U3's supply (pin 8) contains audio noise that propagates through internal 1500 ohm resistors to its outputs at pin 4 and pin 5. Since this noise is common to both outputs, a differential amplifier will ignore it as a common-mode signal. Remember, U4a will only amplify differences between the two outputs of U3.

Gain of U4a is mostly set by the ratio of two resistors (R7 / R2). That's 510k/10k or a gain of 51. Differential gain is twice this value (102).
U4a is also a simple low-pass filter, with R7(510k) and C23(150pf) providing a roll-off frequency of 2080 Hertz. Diodes D3 and D4 clip large amplitude signals, keeping output signals small enough that the mute gate can handle them. These diodes also help keep U4a stably biased, even for very large input signals.
Audio from the product detector is fed into the differential amplifier through coupling capacitors C20(0.1uf) and C21(0.01uf). C20 and R2 act as a high-pass filter with cutoff frequency of about 350 Hz. C21's effect on frequency response is insignificant since U4a's input resistance on this leg is so high.