In uSDX's SSB mode, the audio signal is first digitally processed into two audio signals that are 90 degrees out of phase (I/Q signal).
The phase component (frequency change rate) and amplitude component of the SSB are then calculated; the oscillation frequency of the Si5351 is determined from the phase component and varies at approximately 5 kHz (almost the maximum frequency change rate of the Si5351a).
The amplitude component is controlled by an 8-bit PWM from the ATmega328.
I am not sure about the digital mode of the (tr)uSDX, but I believe it basically uses this SSB mode.
?
In this case, it is basically the same as ADX, where the oscillation frequency of the Si5351 is varied by measuring the period of the audio signal and calculating the audio frequency.
Therefore, the amplitude component is completely ignored. ?
Recent systems such as WSJT-X slowly increase the amplitude at the start of transmission to reduce spurious signals, which is not supported by this method.?
The analog comparator of the ATmega328P is fast enough to obtain a relatively accurate value simply, and the ADX finds the period and calculates the frequency in this way.
On the other hand, with digital audio signals, it is necessary to find the period more accurately than the sampling interval (in this case, 1/11520 second).
For this reason, the time at the zero cross point is calculated by proportional distribution from the values before and after passing through the zero point.?
The period can be determined from the times of the zero cross points.
In this case, the negative to positive and positive to negative points are used to calculate the period, respectively.
If all the audio data can be received correctly, the frequency can be measured with an accuracy of less than 1Hz in a 10ms measurement.?
Thanks,
Hitoshi, JE1RAV
