[Elecraft] Mic gain goes to zero using N1MM

[Guy Olinger K2AV]

On Wed, Sep 9, 2015 at 11:29 AM, K8TE <billamader at gmail.com> wrote:
>
> In my case, the Mic gain is normally set to 8-9 (CM500 headset).  The anomaly decreases the Mic Gain to zero.  I manually reset it to 8-9 which restores SSB power output to normal.  I haven’t tried any other actions to restore the Mic gain.  All indications I have this is a Mic Gain problem.
>

Since the mic gain, displayed in a digit format, is reported as
literally going to the digit zero, we must remember that everything
inside a dotted line within the K3 is entirely SDR. We must carefully
avoid our decades-ground-in tendency to think of troubles in our
familiar and comfortable analog fashion. I do include myself in this
habit and sometimes catch myself going retrograde, having to do a
brainwave CTL-ALT-DEL, and start over from the beginning.

This trouble is presenting inside the K3's SDR dotted line. Of the
eight "pots" on six shaft centers on the left side K3 front panel, all
those rotational functions *advise* the CPU of their state via
multiplexed one and zero state data lines. There is no gain
potentiometer inserted anywhere in the various audio paths, nor is
there a steady state control voltage from a potentiometer controlling
a linear pass transistor in the audio string anywhere, thank your
lucky stars. Remember "scratchy" audio? Just another of the analog
bugaboos happily banned forever by the K3's mostly SDR hybrid scheme.

The MIC control function uses a *shared* encoder assigned three
separate unrelated settings. If the encoder or its physical data
connection to the CPU was a problem, it would affect all three
functions MIC, SPEED and DELAY on the front panel. I have heard of the
encoder going bad and needing to be replaced. If the encoder is
miscellaneously sending a string of "decrease" encoder signals, it
should also happen in SPEED or DELAY mode. All the encoder can do is
send increase or decrease signals. You would have your CW and VOX
delays going to nothing, or your CW speed going to the 8 WPM minimum.

The CPU knows what function is currently "on the knob" and the static
values in force for the three lower left encoder functions. Exactly
one function at a time is currently assigned to the knob, and the CPU
interprets a "decrease signal" accordingly. The decrease signal
travels to the CPU over a multiplexed data line which is either there
and properly working for many diverse functions or is not for all
those functions.

Doesn't this really start to smell like a program issue? That gets you
to the next thing -- if it is a firmware bug, then the trouble is
present for ALL K3 users running the affected firmware version(s). And
we should have lots of reports because every single user of the
firmware could be experiencing the same problem intermittently.

The only program code that could isolate the trouble to a *few* users
would have to be external to the K3. Is it possible for an external
program to set the MIC gain? This of course is impossible in an analog
radio priced for ham customers, so analog thinking would not suggest
that. The trouble would have to be IN an analog radio. But since the
K3 is digital, we note that page 1 of the K3 programmer's guide has
"MG * Mic gain" cell in the command table cheet sheet. That just might
be an "Aha" moment.

If an external program intermittently sent an MG 000 command to the
K3, you would inconveniently find the MIC gain set to zero,
intermittently. Not a K3 intermittent physical problem, but a K3
*commanded* to set MIC gain at zero, a command which the K3 mindlessly
obeys, as yet unable to read a contrary indication from the mind of
the operator. Let me know when the mind-reading K7 shows up. I want
one. Get rid of a lot of cables and input devices.

I, myself, with my terribly soft and mumbly radio voice, with
extensive trials managed to get a good setting for K3 SSB MIC gain,
compression and TX EQ settings. I was directed to those settings, and
had those settings confirmed as clear, punchy and understandable over
the air, by the PVRC contest guru crowd. Predictably I haven't myself
touched those settings in maybe three or four YEARS now. They are
still where I put them. They better stay there, too. People don't hear
me nearly as well when I'm non-K3-processed. It's like turning off the
amplifier.

Repeating myself, if it's a third party program on someone's
particularly configured PC, or with a particular parallel combination
of running third party programs, then the mic gain to zero problem
would be scattered and rare. If it's in K3 firmware, it probably would
have been caught in alpha or beta testing. If if got to the general
field in a production release, Big E would have been buried in
complaints.

Myself, I would suspect something incoming on the control serial line
like "MIC 010" losing ASCII digits and arriving as "MIC 0", or mangled
to "MIC 000" or something like that. Does the K3 process a "MIC 0"
command or consider that an invalid command? I don't know. But poor
physical connections for external serial lines, or overloading USB
hubs have resulted in mangled command strings, or two programs
fighting for exclusive use of the serial line, and those kinds of
problems far beyond Elecraft's control, are as common as nails. And by
my reading of this reflector those troubles are frequently first
blamed on the K3, even if the true chances of that actually being
correct are worse than hitting the lottery.

Do follow the support advice to disconnect all the external programs
and see what happens.

The complete list of all possible combinations of third party programs
that can drive a K3 over a serial connection is an absolute witches
brew of widely scattered quality, from the sublime to the simply
awful, sometimes implemented in a hamshack with grotesque physical
arrangements. There are programs in distribution that have known
problems that the coder has no intention of ever fixing, for any
number of reasons, and some have publicly stated such. What you got is
all you are ever going to get with these. Meaning that your particular
problem could have been reported to the author a thousand times
without effect. That in a programming world that swiftly keeps moving
on. And of course this is without mentioning the ubiquitous
fallibility in PC Bios programs and operating systems.

Elecraft is your friend. Respect the Elecraft.

73, Guy K2AV