[GreenKeys] model 14 ktr

[John Nagle]

On 10/16/2011 9:02 AM, greenkeys-request at mailman.qth.net wrote:
Eric writes:
>
> Well, thanks to the generosity and transport effort of George Hutchison
> (thanks again, George!) I have spent half the night messing with this model
> 14 keyboard typing reperf.
>
> After basic adjustments, here's what's good:
> about 98% accuracy printing/punching from the loop
> about 98% accuracy sending to the loop
>
> Here's what's bad:
> about 40% accuracy printing/punching what I type.
...
>
> There are surely also some spots that still need oiling as occasionally I'll
> get wrong characters sent for a particular key until I hit the key a few
> times. Need to look up the codes and see which bit or bits are sticking.
 >
> eric

    If you're that far along, you should be able to get perfect
operation.  98% accuracy is sub-par.  Local operation should
be above 99.99% accuracy.

    What are you sending to on the "loop"?  Some electronic device with 
a UART, or at least an amplifier?  Presumably something that has
more tolerance for a bad signal than your Model 14's selector
mechanism.

    A good first step is to put a scope on the output and look at the
keyboard waveforms.  Consider putting a small resistor in the loop
and measuring the voltage across it. Then you're measuring current,
which is what matters here.  This will show you if one of the
keyboard contacts has a problem.

    Get a contact burnishing tool (about $2) and clean up the keyboard
contacts.  If there's high-resistance carbonization in there from
previous arcing, the contacts may not be passing enough current.
Put the little abrasive surface between each pair of contacts and
rub it back and forth a few times.  Don't use sandpaper or an
emery board for this; a burnishing tool won't damage the contacts
or leave a conductive residue behind.

    Send RYRYRYRYRY to the machine and watch the code bars go
back and forth. They should consistently go from 10101 to 01010.
When the machine is stopped at a point in the cycle where the code
bars are not locked, you should be able to slide them back
and forth with a finger touch.  You have a perforator, which
adds an additional load on the code bars.  Something may be
binding in the perforator linkage or punch block.

    For each typebar, you should be able to pull the typebar
down to the ribbon, release it, and it should return to
its parked position with no binding.

    If anything binds in the above tests, cleaning and oiling
is indicated.  My web site shows how I do cleaning and oiling
in some detail.  These machines work much better once all
dirt and old lubricants have been cleaned out.  Hopefully
you won't have to do the de-rusting process, which should
be avoided unless necessary.

       http://www.aetherltd.com/refurbishing14.html

    After all that, if you still have problems, get out the
adjustments manual.

	http://www.aetherltd.com/manuals.html

    Early Model 14 machines have no RF suppression at all; they
predate widespread use of radio.  So some filtering and snubbing
in the keyboard circuit may be indicated.  1000pf (a 500V ceramic;
there's a big spike to damp) in series with 100 ohms across the
selector magnet works well for me.

					John Nagle