[WriteLog] WL CW keying comments

[W. Wright, W5XD]

I have some comments on the numerous postings about problems with CW
keying:.

In Writelog's Setup/Ports menu, there currently 7 different entries under
"CW keyer type". Three of the settings, "Microham" and the two "Winkey"
settings share some code with each other (because they all have the K1EL
keyer chip underneath), but none of the others share any code with each
other or those K1EL keyer settings. It is exceedingly unlikely that there is
any garbage cw software problem that is somehow shared among those settings
that have no common code.

That means that any posting on this topic that does not say what the "CW
keyer type" setting was is not helpful in narrowing down any diagnosis. If
everyone that has such a problem reports that setting, I am pretty sure we
would discover that there are several "CW keyer type"s involved, which means
that, assuming there is a software issue in WL, there has to be more than
one. But to see any pattern, the data we have must be first separated by
that setting.

All the modern keyer types also involve a COM port (I doubt many folks use
LPT port circuits on modern machines). I have noticed major differences in
device driver behavior between the COM ports that are built into PC
motherboards, and the ones that connect via a USB port. This is just a hunch
on my part, but I suspect there might be a pattern if everyone reports
exactly what hardware their COM port is. Again, the report has to be
specific to be useful. Right click on "My Computer" and choose "Manage" and
then go to the "Device Manager", and finally, browse to the COM port that
you are actually using for CW keying. (Don't guess--figure out which one it
really is by looking at WriteLog Setup/Ports). First go to the "General" tab
and report the "Manufacturer". Then go to the "Driver" tab and report the
"Driver Provider", and "Driver Version". 

If people report those configuration details, there might be a pattern among
the reports.

It has been suggested that multi-threading on multi-core CPUs might be a
culprit. I won't agree or disagree with that theory, but I will comment that
internal to WL, all CW keying code runs on the same thread as the rest of
the WriteLog, including the user interface. (This was its original design
back in the Windows 3.1 world where there were no threads, and I have never
changed it.) The OS itself, and any device drivers, have multiple threads on
multi core CPUs and that could somehow introduce bugs, but I consider that
unlikely just taking into account the fact that we're talking about
software/hardware combinations that are run in far more situations than ham
contests. 

To add my own experience to the discussion, I ran the "w5xd multi+MK1100"
setting with 1500W and two radios for 1000 QSOs this past weekend and did
not notice any keying issues. The CPU is a dual core AMD. The port I
successfully used this past weekend says manufacturer: "(Standard Port
Types)", "Driver Provider" Microsoft, and "Driver Version" 6.1.7600.16385. 

At this point I don't see any pattern in the reports that I can see how to
investigate further.

Wayne, W5XD