An amateur radio project using the uSDX hardware and controlled with a Teensy 3.2. This project uses the Teensy Audio Library for a large part of the DSP processing and the prototype was built using the QRP-LABS QCX+ transceiver kit.
The concept and a part of the code is based upon the original work of PE1NNZ and can be found here: https://github.com/threeme3/QCX-SSB
There is a thread on groupsio for this project:https://groups.io/g/ucx/topic/alternative_processor_project/85821954
The uSDX is a minimilist hardware design where the software does as much of the work as possible. The key part is the EER ( envelope elimination and restoration ) transmitter. The transmit audio is split into I and Q audio, the amplitude is sent to a PWM modulator of the final amplifier and the frequency part is sent to the Si5351. The transmitter chain does not need to be linear yet can still transmit SSB.
The Teensy 3.2 was chosen for its 5 volt tolerant inputs and the Teensy Audio Library. (It should be noted that PJRC ( makers of Teensy ) offer an Audio shield that would provide superior audio to this minimal hardware design. The codec contains a gain controlled analog front end and anti-aliasing filters.) The Teensy audio library operates at 44khz and magically streams audio between objects. It runs on a low priority interrupt and processes blocks of audio that are 128 samples long ( about 3ms long ). The use of the audio library has some advantages and disadvantages. With one or two hours of work, one can completely redesign the whole radio which I have done many times. One disadvantage is that the sampling of audio with the two A/D converters completely tie up the A/D hardware and analogRead commands can not be used ( using the audio shield does not have this issue ). One advantage is that the Teensy can appear in Windows/Unix as a USB soundcard and adding this feature is trivial, a setting in Arduino to compile it in.
The receiver uses the Weaver method of decoding SSB. The vfo is not placed at zero beat with the desired signal but instead is placed right in the middle of the signal. The I and Q audio is low pass filtered at 1/2 of the desired audio bandwidth. It is then mixed with a complex BFO also at 1/2 the bandwidth in frequency ( if bandwidth is 3000 hz then the BFO is at 1500 hz ). The result is added or subtracted to get LSB or USB.
The transmitter re-uses a couple of the receiver audio blocks for the microphone input. The microphone, usb, or the sidetone can be selected as the transmit audio source, and the custom object MagPhase processes the audio and calculates the EER data for the Si5351 and PWM modulator.
The QCX was built as outlined in the original PE1NNZ documentation( link above ) with some small differences. A pullup was added on the signal RX, a pulldown was added on signal KEY OUT, a pulldown was added on signal SIG OUT. A FET (BS170) was added to switch the microphone signal ( DVM IC2 pin 25 ) to Teensy A3.
( note sometimes I refer to AREF as VREF in the pictures )
I did not install C4 C7 as adding them caused processor noise to enter the receiver on a previous project.
The QCX+ is wired so it works with both the PE1NNZ ATMEGA 328 program and with the Teensy 3.2. The headphone audio circuit is duplicated but doesn't need to be. I used a duplicate circuit as that was easier than finding another Jxx point on the QCX to bring the SIDETONE net up to the developement board.
I have since removed the black DVM wire and the purple wire to the DAH net and separated the microphone from the paddle. A jack was installed at the CAT position and wired to a microphone preampliphier circuit installed in the footprint of IC10.
