Hi Steve there are a number of factors that you will need to take into
account in deciding where to start I would say that a crystal mixer is
adequate for 3sec dots QRSS and marginal at 10 secs and not sufficiently
stable really for 30 sec dots. Mixer VXOs need the crystals to be driven
fairly hard to get a reasonable swing even on 136. My own preference is for
a couple of Colpits circuits, rather than the logic inverter version. The
latter can drive the crystal so hard they go off at an overtone (yes even
2MHz  AT cut crystals). The circuit I used has been 'stolen' from LMR FM
radio design of the 70s and intended to start the crystal over a fairly wide
temperature range in mobile use.
http://www.alan.melia.btinternet.co.uk/mixervxo.htm

There are also some simple design equations for a single ended Class E PA
running a few hundred watts....air tested by Finbar EI0CF and Mike GW4HXO
but only put on a dummy load here.

Actual stability does not matter if you are above the noise level i.e. not
more than about 20dB down on audible morse. Sometimes a slightly drifting
signal can stand out from the weak Loran sidebands. That is providing it is
a constant strength and not popping in and out of the noise with propagation
peaks and troughs.

For very long distance work where the listener is working blind, you will
need to be able to specify you frequency to at least 1Hz and probably
slightly better. If you drift you may start in the clear but 'run into'
Loran lines. I missed receiving Larry on the day he was first received by Ko
because I rejected an interesting looking signal because it was about 1.4Hz
off the expected (I think ). Ko wasn't sure of the exact frequency and saved
everything and the reception was verified by John G4CNN after some image
enhancement. For 30sec dot over a long distance it is advisable to be able
to hold 0.1Hz for about 90minutes.

Personally I would build a mixer for starters, its easy and gives you a
source to enable you to play with the PA and the TX aerial which will be the
major part of the projects. Even work a few locals ?? well up to 1500 miles
on QRSS3!!   From your location I would consider the possibility of a DDS,
which could then be programmed to send Jason or other more esoteric modes.
Getting the first signals across is the challenge, after that it gets easier
(I will get tomatoes thrown at me at the HF Convention this year for that
comment !!) I have listened over many months to stations getting stronger
and stronger as they gain experience and improve their set up.

Others will probably have different viewpoints !!

Good luck, I look forward to listening for you when the big day comes.

Cheers de Alan G3NYK
[email protected]
http://www.alan.melia.btinternet.co.uk