I built the Pixie 2 and have had a lot of fun with it. I was lucky during testing as it happened to be a very active contest weekend, so now I have had contacts with 15 different countries with it. In addition to the side tone modification described later, my Pixie also has increased gain and some filtering. Also, the crystal was replaced with a 3.59 MHz ceramic resonator in series with a variable capacitor, so that the frequency could be pulled all over the CW portion of the 80 m band. Frequency agility is the key to obtaining many contacts with QRPp equipment, in my experience. The ability to tune is much more important than higher power. The usual mode of operation is to answer a call, and being locked to a single crystal’s frequency severely limits the number of contacts.
I do not remember where I first learnt that the bypass pin of the LM386 (pin 7) can be used both for muting and as an input, but I cannot claim that I have discovered it myself. However, it is my hope that the combination with the Pixie is new to readers of SPRAT, and also that it will give some ideas for other rigs using the LM386.
Reduced Sensitivity to Broadcast Interference
The main cause of broadcast interference (BCI) in the Pixie is nonlinearity in the LM386 so that it detects strong signals like a crystal detector. The root cause of this seems to be the way the LM386 is fed from the power supply. In order to accomplish muting, the power supply is fed to pin 6 via a 1k resistor (see Fig. 1), but this starves the amplifier for current and causes nonlinearity. At my location I can easily hear how the interference decreases and often disappears as I short circuit the 1 K resistor.
|Figure 1. Audio section of Pixie 2|
The fix is to use the under-utilized pin 7 of the LM386. It is designated as a bypass connection, but it can also be used for muting. Both grounding and connection to Vcc will mute the amplifier. The fix for the Pixie is to feed pin 6 via a small resistor say 10 - 100 ohms from Vcc, disconnect the positive side of the diode from pin 6 to the key and move it to pin 7, see Fig. 2. By the way, I have used 1N4148 instead of 1N914 without any problems, and often R1 can be omitted, but at the risk of oscillation.
|Figure 2. Audio section with improved muting|
Side tone circuit
The second shortcoming of the Pixie is the lack of side tone. I have borrowed G4GVM's circuit for the FOXX-3 in SPRAT 98 and adapted it for use in the Pixie, see Fig. 3. The transistor can be any general purpose NPN type with a fairly high current gain like BC109, 2N2222 or 2N3904. The variable resistor sets the level, start with it in the mid position and adjust for a comfortable level.
|Figure 3. Audio section with sidetone circuit|
In this circuit pin 7 is used as an input. The original Pixie 2 has pin 7 free, but in some variants, pin 7 is bypassed with a capacitor to ground. If it is only 0.1 uF as in the Knightlite SMiTe version of the Pixie, the side tone input will work anyway (and also the muting circuit of Fig. 2). In Fig. 3, the mute circuit is completely removed since the LM386 needs to amplify the side tone during transmission, and therefore this circuit also has increased BCI immunity. The original Pixie circuit uses the charging/discharging of the input coupling capacitor C1 (0.1 uF) into the 50 k input resistance of the LM386 as a primitive source for a side tone. With a real side tone, these clicks are in my opinion more annoying than useful, so optional click reduction can be achieved by reducing C1 to 0.01 – 0.022 uF.
In conclusion, two simple modifications to the Pixie 2 have been shown. The first one reduces sensitivity to broadcast interference, it has no side effects and it is accomplished without increasing the number of components. The second circuit also adds a side tone circuit. Both of the modifications use pin 7 of the LM386, either for muting or for input.
(Originally published in SPRAT No 113 (Winter 2002) pages 18-20.)