A CW transceiver for 30 Meters using 72 components

By KD1JV

For the spring 2010 FDIM (Four days in May) design contest, the ARCI (Amateur Radio Club International) decided on seeing if anyone could design a rig using 72 or less parts. What made this design a challenge was the fact the rig had to be a transceiver, the receiver had to have single signal reception and only 1 IC could be used. The design goal of my version was to come up with a reproducible design, using all common, easy to get and inexpensive parts. It also had to have a reasonably sensitive receiver and a transmitter with a bit more than flea power output. I meet these goals, the receiver having a MDS of about 0.5 uV - not great but usable - and a transmitter which puts out about 3/4 of a Watt. Nearly all the parts can be sourced from Mouser, the exception being the transmitter crystal and the toroid cores.

Five rigs were entered into the contest. I'm sorry to say my rig didn't win the contest (likely due to the receiver not working because of a lead shorting to the case I didn't notice after packaging it). I did however get an Honorable mention certificate.

..

The Circuit:

A number of different ideas for a receiver were tried before the final configuration was hit upon. It was quickly evident that making a rig with acceptable performance using only 72 parts was not going to be easy. I kept ending up with 80 to 90 parts. Everything finally came together after I hit upon the idea of using a LM386 as a combination product detector and audio amp. I'm not quite sure what made me think to try using a LM386 in such a fashion and was amazed when it actually worked! I'm not exactly sure how this works, but my theory is thus: If you look at the internal configuration of the LM386, you will see that the two input transistors share a common constant current sink. I believe the RF signals applied to the amplifier inputs are mixed in this current sink and the resulting audio beat note is amplified as if it were connected to one of the amplifier inputs. Some experimentation showed that the BFO signal had to be applied to non-inverting input pin 3 and had to be of very low amplitude. Therefore, pin 3 is left floating and stray coupling is used to inject the BFO signal. The amount of BFO injection is controlled by physically moving R17 closer or farther away from pin 3.

RF by-passing of the LM386 is important and must be done physically close to the chip. C18 across pins 1 and 8 improves the gain of the amplifier. MDS sensitivity of this product detector/audio amp is about 50 uV for a signal applied to the low impedance winding of T1.

The rest of the receiver is pretty standard. A super VXO provides pretty much full band tuning of 30M for the receiver. The mixer is comprised of series connected j-fets, which mimic a dual gate mosfet in operation. A simple two crystal, one cap IF filter provides reasonably good selectivity and opposite sideband suppression. An IF amplifier is required to get reasonably good sensitivity. Muting during transmit is achieved simply by using a MOSFET to switch a low value resistor to ground from the IF amplifier's base. This reduces the gain of the IF amp without completely turning it off. Side tone during transmit is achieved by tuning the receiver to the transmit frequency.

Because most of the 72 parts we were allowed to use are used up in the receiver, the transmitter had to be as simple as could be. This meant crystal control. Although directly keying the oscillator would have saved a few parts, using a PNP keying switch allows some wave shaping of the rise and fall time of the oscillator and simplifies controlling the QSK and mute switching. The single transistor output amplifier provides 700 mW of output power. The output low pass filter attenuates all spurs to -50 dBc or better. This is achieved by adding C22 across L5, which makes a trap at the second harmonic.

Improvements:

The schematic notes that the internal cap across the T1 secondary is removed and a 390 pfd cap used to tune the secondary to 4.194 MHz. This was done because if the internal cap was left in the IF can, it would have to have been counted as an individual part and that would have made the total 73 instead of 72. It really makes more sense to leave the cap in the IF transformer and use a 330 pfd cap to retune it instead.

2N7000's can be used instead of BS170's, but note that the Source and Drain pins are reversed between the two types. Therefore, when using 2N7000's, put them in the board 180 degrees from the part outline shown on the board.

Adding RIT would improve the usability of the rig significantly. This is easy enough to do and adds about 11 parts. Having RIT would allow keeping the receiver tuned to the transmit frequency to hear a side tone and allow tuning in another station which might not be exactly on your frequency. A suitable RIT circuit is shown below. The RIT voltage divider needs to be supplied with a regulated voltage. This could be either a zener diode and resistor or a three terminal regulator. A voltage of 8 to 9 volts would be suitable for the supply. The way this circuit works is during receive, Q11 is in an off state, allowing V1 to adjust the voltage across a tuning diode, D3. When the transmitter is keyed on, the mute voltage goes high and turns Q11 on, shorting out the RIT control, V1. The voltage across D3 is now centered at half the regulated voltage because R18 and R19 on either end of the RIT control are equal values.

No number is given for D3, as a number of different kinds of diodes will work, with varring degrees of tuning range. 1N4001 diodes will work, as will zeners with a reverse voltage of 20 or more volts. LED's also make good tuning diodes.

Construction:

I built my rig on a pcb as shown. Dead bug on ground plane should work just as well. The tuning cap C17 is a polyvaricon, which can be taken out of a junk AM/FM radio or bought new from www.qrpkits.com. An air variable would work just a well (if not better, especially if it has a reduction drive built in) just being somewhat larger. A to scale board layout can be down loaded by clicking here. This is a through board view for directly printing on toner transfer film. The image gets reversed when you iron the print out onto the board.

Tune up:

T1 and T2 will need to be adjusted for best receiver sensitivity. BFO trimmer C14 is adjusted for best opposite side band suppression. The position of R17 in respect to U1, the LM386 will affect receiver sensitivity. Tilt the resistor over towards U1 until a high pitch audio oscillation starts, then move the resistor away from the chip until the oscillation stops.

Parts layout: Actual size : 3.3" x 2.4"