It's a bit late to go into this tonight, but I think I see what has happened.
What I think you have managed to do is tune the early stages of the transmitter to (2 * 10.695) Mhz = guess what, 21.390Mhz
Nothing broken but the tuning, which is wildly out.
Rule #4: Be aware that simple RF power meters only show how much power there is. They don't care what frequency it is on.
To get the tuning back into the right region, use a receiver to monitor the output on the middle channel, middle band and initially tune to increase the indicated signal on the receiver, ignoring what is happening on the power meter. That way you can be more sure that you are tuning the coils to peak on the right frequency.