Intermittent lack of audio on start-up

[w9ran]

I'm seeing inconsistent operation of the transmitter in which no audio is present, even though the RF drive is present. I confirmed this by looking at the RF signal at the gate of the MOSFET in AM mode, which shows a steady carrier but no audio. In SSB, broadband noise appears and can be heard to diminish at the end of the audio file and resume when it loops. Stopping and restarting, power cycling, etc does not restore correct operation.

Occasionally the audio will begin as it should, but I haven't been able to identify what triggers proper operation. I suspect it that no or incorrect serial data is being sent via USB when this happens even though the transmit.py program runs OK. When streaming does start correctly it sometimes will run reliably for an hour or more, although at times it stops intermittently with a time-out message for reasons unknown. Transmit.py is running on a Windows 10 PC.
Bob

[w9ran]

Jon, please ignore my comment about AM mode, I momentarily forgot there will be no phase/frequency modulation only amplitude. Instead let me focus on LSB mode, where I happened to capture the anomaly within 1 minute after start-up in this video:
https://share.icloud.com/photos/0acC_j49Yn53qOiVZF0-Xr7PA

You can see the broadband noise and hear it in the receiver, with a silent period when the audio loops. After 16 seconds, you hear one cycle with correct LSB modulation, then it returns back to noise. About 2 minutes after making this recording correct modulation appeared and continued without interruption. The receiver and spectrum analyzer are loosely coupled to the gate drive signal, which is the OSC signal from GP8 of the Pico, buffered through two 74AC00 gates wired as inverters to drive the MOSFET PA gates.

[w9ran]

A bit more information about this issue. First, it only occurs in LSB or USB mode - AM and FM start up correctly.
Second, while the wideband noise and distorted audio is present, o'scope monitoring the RF output shows very large spikes:

Easier seen and heard in this video: https://share.icloud.com/photos/0d7-_tWidtKNNKVHT_DEOMixA

AM and FM start with correct modulation instantly. USB and LSB start like this, then after a few minutes correct audio begins with no correlation to anything I can see - just spontaneous. After it becomes normal the audio and waveform are perfect and remain so until the python program is stopped.

Hope this helps identify the root of the problem.

Bob

[w9ran]

Now with the test-tone, it is possible to compare start-up of the Pico in different modes. The screens below show the signal at GPIO8 on a new Pico with no hardware attached - just the bare Pico board powered from USB.
On the right is the AM mode which shows a solid single frequency carrier vs the broadband noise on the left which is LSB mode.

On my transmitter prototype the broadband noise usually disappears and normal audio begins after some minutes but the bare Pico has not done so after 30 minutes. Very odd!