I am working on GF-11 transmitter restoration and having problems with
CW keying (ON-OFF keying). The original design has provisions for MCW
mode, but not for ON-OFF keying.

I tried various keying tricks with control grid bias and screen grid
bias keying in output stage (pair of 837 tubes in push-pull), and the
net result is small frequency shift of about 100Hz.
The original circuit does not have a buffering stage from oscillator
output to PA input, and anything I do in PA stage changes the loading
for the oscillator output, which in turn shifts the frequency of the
oscillator circuit.
Any suggestions?

I also have one, or possibly two spare plug-in coils for GF-11 Tx, if
somebody needs.

73,
Vlad, kb9olm (email: kb9olm AT yahoo DOT com).

I remember the GF-11, but not real familiar with it. It's similar to a
number of military sets designed during the mid Thirties. And most often
were of a simple MOPA (master oscillator,power amplifier) circuit.

With no buffer stage between the oscilator and final stage, I would think
that it would be really tough to obtain a really clean CW signal.

I'm sure you're familiar with the ARC5 command sets of a bit later
vintage. They were of the basic MOPA type. I recall that some guys claimed
that they got clean keying by keying the PA screen grids and oscillator
plate together.

Are you running the set from a regulated power supply? That might help.

For real fun, try to modulate the 837's. I'm sure you will wind up with an
un-happy mix of AM and FM at the same time.

I ran MOPA Command sets in the Fifties, and always added a buffer stage
(12A6 tube) between the osc and final. Mostly on 75 meter 'phone.

Are those 837's really in push-pull? Those small military rigs usually had
parallel finals.

And bottom line, a hundred cycle shift to me sounds really cool and adds
character to a transmitter from that era. Especially nice on Straight Key
Night.

What is the oscilator tube? Maybe a triode like a 10,or something
similiar.

I have a nice BC-230 transmitter of the same era, and I never even thought
of trying it on CW. It uses triodes all the way through. A '10 driving a
'10 with a pair of VT-25's in parallel as a plate modulator.

John