In 1.4.1, the new keyer will be disabled. For many of you it is working
perfectly. For those of you with an ACPI 1.0 motherboard, your real
time clock will not support what we needed to get the keyer to work
reliably. With the ACPI 1.0 so called hiperformance frequency is 3.5
Mhz and you would've though this was adequate. It is not. We have
several who are generating little bursts of white noise and why is
clear. We then protect against that happening and the clock jitter and
element spacing goes all over the map on these ACPI 1.0 mobos. We need
a more careful analysis than we have given this and get reliable tests,
etc. We can't allow you to generate white noise at full power on the
air. The old keyer will be back in charge temporarily.
Frank wrote this keyer to run as a separate process entirely. We were
trying to bend it into a thread rather than a process unadulterated by
all that was going on around it. The primary difficulty was in dealing
with MOX and attempting to handle delays with minimum latency and and
insert an asynchronous thread's ring buffer in place our another threads
ring buffer and it is just all too fragile the way it is written.
Fear not, it will return after a clearer picture emerges.
For those with a compiler, you can enable it but do not expect it to be
reliable as we are in the middle of serious debugging.
Bob
