The key input is just grounded to keep the amp "on", I don't do any manual, hands on a key, CW stuff. I would like to keep the RS232 input though, as software generated, very slow Morse is a method used on LF for communication that I have previously used on a different amp.
As for the "signal", yes the 272 kHz square wave is derived from either the newly built U3S device :
https://www.qrp-labs.com/ultimate3/u3s.html
or the Kenwood Ts-590, pre amp and frequency doubler. The signal is present as soon as the software begins a TX sequence. WSPR gives just under 2 minutes continuous, OPERA has a few seconds TX, then stops for a few seconds, then on again, for a 32 minutes long TX sequence.
As the U3S seemingly has cured the randomly occurring start up oscillation or whatever it was, I may stick with it, it's cheap as chips and frees up my expensive TS-590 for SSB voice (yes, I do actually speak with other amateurs as well as use this funny digital stuff....
). As such I am going to investigate a firmware option to have 180 degree out of phase, (mirrored) outputs generated by the Si5351a device. I am thinking this might drive the FET driver IC directly? I need to look at the levels, and see if one output is held high when no signal is being created. If so capacitive coupling will still be needed.
Finally, is the groups decision final that C1 and C2 with their associated 15k resistors are there to help stop both FET's being briefly turned on together?? And if so would their continued presence be desirable between the Si5351a outputs and a driver IC, should I be able to make the idea work?
Thank you spec!! I am expecting a call from RS Component's marketing team asking why I am no longer buying dozens of FET's