The third probe run puts numbers on both. TxBufLen lags by 100 to 150 ms, about
one character at 45.45 baud: twenty-one characters pushed at 2496 ms read 0 at
2497, 2547 and 2597, then 26 at 2647. It moves once per symbol transmitted, so
readings 50 ms apart repeat.
It read 26 for 21 characters because it counts Baudot symbols, and the message
carried two digits: a shift to figures and a shift back each time, plus one at
the start. That is the case the cap was needed for and I had not seen. A contest
exchange is mostly digits, so the engine always has more to transmit than the
clock thinks, and the clock alone would run ahead of it on every QSO.
Slack of three characters covers both, so nothing changes in the pump. The
count reaching 0 does not cut the end of a message off either: it emptied 350 ms
before the last character was decoded back and the engine held the transmitter
up for another 900 ms.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RtspmWmS7f8kUvcyaHpRWZ
ericek111
7ae60be4a0