[Elecraft] Straight Key on my K2

[Don Wilhelm]

David,

Before blaming the firmware level (you apparently have an older K2), check
the values of R1 and R2 in your K2.  If they are the older value of 470
ohms, change them to 220 ohms.  That may help the situation.

BUT...
The real problem is due to contact bounce at the key contacts.  The firmware
cannot detect the closure on both the dot and dash lines at the same time,
but when one is detected, the other is checked to see if it is closed too -
now if the contacts are bouncing, at the exact time the opposite line is
checked, it occurs at a time that just happens to be during a higher voltage
part of the bounce cycle.

For those not familiar with contact bounce, it happens with any switch on
initial contact closure - if one were to look at the closure with a 'scope,
there will be periods when the contacts are actually closed and times when
they are open (when viewed on a microsecond scale) - only after that initial
contact bounce has settled down are the contacts really closed full-time.
We perceive a switch or key as fully closed instantly because our brains and
senses do not work on that small a time frame, but it makes a big difference
to electronics which do work at microsecond increments.

You will find autodetect more reliable if you drive the K2 keying from an
electronic keyer or computer.  It is hand keys that cause the problem.  The
extent and frequency of failure will depend on the key contacts - large
soft-metal plated contacts (gold or silver) should be better than keys with
brass or nickel plated contacts (but they will still bounce).  Old mercury
switches and relays were a solution to this contact bounce problem.  If you
can find a mercury wetted reed relay, use the handkey to activate the relay
coil and use the relay contacts to drive the auto-detect diodes and I
believe you will achieve much more reliable operation, but then that is
putting a lot more components into what should be a simple arrangement.
Sometimes simplicity does have its disadvantages.  There are electronic
methods to combat the contact bounce problem and the really reliable ones
usually involve set/reset latches that set on the contact closure and are
hit with a reset pulse sometime later (up to 10 milliseconds later would be
a reasonable choice for hand key speeds) - the set must override the reset,
so choose the latch logic carefully.

Of course, the easy solution is to use an external keyer and connect the
hand key in parallel with its output - the K2 can be permanently set to INP
Hnd for that setup.

73,
Don W3FPR


> -----Original Message-----

>
> Many thanks, Ron.  I wonder whether this is something that could be
> cleared up when the controller is upgraded?  Mine is version 1.09.
>
> As it is at present, our K2s are clearly not working as intended.
>
> 73 de David G4DMP
>
> In a recent message, Ron D'Eau Claire <rondec at easystreet.com> wrote ...
> >This issue (getting alternate dits and dahs when using a hand key with
> >autodetect) came up years ago - five or so IIRC. It was finally
> pinned down
> >to a polling issue in the K2. That is, that even though both ring and tip
> >were pulled to ground simultaneously, the controller in the K2
> didn't always
> >recognize that fact, depending upon what else it was managing at
> the moment.
>
>