A Simple Modular DSB TX

There are many designs for simple and effective CW transmitters. There are rather few for simple phone transmitters of any kind. (Some novel designs can however be found on JF1OZL's Web page.)

The simplest way of getting on the air with phone is to convert a single transistor crystal controlled TX into an AM transmitter by putting a carbon mike in series with the poer supply. This is illustrated by a desin in the G-QRP Scrapbook (p??). There are several disagvantages to this approach, the main practical one being that depth of modulation is very low, I would estimate about 15%, and so the effective audio power is only going to be a few milliwatts. It does work after a fashion however, and short distance contacts can be made.

A more elegant approach is to use DSB. At QRP levels this is a perfectly respectable mode even on narrow bands such as 40m and 17m. It is more efficient than AM and not that much harder to construct.

Alternative Approaches

All methods of generating DSB follow the same principle, viz a mixer with local oscillator at the signal frequency and audio input. However, there are two rather different approaches in practice.

High level modulator

A diode mixer, which can be anything from a single diode in an unbalanced arrangement (which would give very poor carrier suppression) to a commercial 4 diode packaged unit such as the SBL-1, is fed with audio and carrier at some tens of milliwatts. There is typically 7-10dB of loss in the mixer. A milliwatt level TX can be made by feeding audio to a diode mixer either from a carbon microphone circuit or an audio amplifier such as the LM386. The lowest active component count for such a transmitter is one transistor. See JF1OZL's design. Howver, less usual components such as corbon microphones and transformers are likely to end up being more expensive than semiconductors or even ICs, so a low component count does not necessarily lead to a cheap rig!

A more conventional design was given in Sprat by ZL2BMI. Although this uses a carbon mic a dynamic or electret mic may be substituted.

Low level modulation

An active mixer of the 602/612 family can be driven directly with a microphone and has a built in oscillator. It has a typical gain of 17dB and so should be capable of providing a milliwat level single chip TX.

The design below uses a ceramic resonator, either 3.395MHz or 2MHz for 160m, which can be pulled over a useful portion of the respective bands.

Cm acts as a high pass audio filter. A value of 1microF is suitable, but can be between 100nF and 10microF.

The `crystal' is a ceramic resonator, either 3.695MHz for 80m or 2MHz for top band. For top band a 22 microH series inductor should enable the oscillator to cover most of 1.9 to 2.0MHz.

Cv is any conveniently available twin gang capacitor. A large airspaced 2 x 500pF device will give coverage of much of the 80 phone section, typically from 3.6 to 3.74MHz, a compact `polyvaricon' rather less.

Linear Amplifier

The amplifier is a straight copy of ZL2BMI's circuit in Sprat 84. With drive direct from the NE612 I can get about 300mW out, adequate for `local' qsos, and maybe a bit more if conditions are good and you're persistent!

A description of the PA is here


I built this in `modules', see photos, starting with the exciter. This is best in a screened box to ensure carrier supression. It was built on a piece of perforated board. The version in the picture uses only a single variable capacitor (the left hand half of Cv) which gives only about 40kHz of pulling range, but was what I had to hand.

The amplifier was built using (very) ugly construction. The BD135 heatsink is a piece of thin copper sheet (all metal, not pcb material) fixed with a suitable bolt. This must NOT touch an earther case or ground plane.

Everything is linked up using phono plugs and sockets. A beginner can build the exciter, connect up the microphone, put a short piece of wire on the output, tune in on a receiver and hear himself speaking. In fact, this constitues a single chip TX with an output of a few hundred microwatts. It should be possible to use it very short range qsos!

Further Development

The amplifier is capable of 1-2W output on 80m if the drive were increased. Another driver stage could be used. It may also be possible to increase the output from the NE602/612 by putting DC bias on pin 1, an idea just seen in Sprat 110.

One thing I'd particularly like to do is to make a 12m version. This idea arises from some experiments with two 25MHz crystals in parallel (and they're cheap enough!) along with a series inductor of about 2.2 microH. This will pull over the whole of this narrow band. However, the current linear amplifier runs out of puff well below these frequencies and I have yet to get round to finding or concocting a more suitable design.

I definitely must get round to this, because there aren't many simple designs for this band, and it is one on which QRP phone can be very successful.