Help with single chip transmitter for HF

Hello,
I design a small transmitter for the 40m band and I have found this circuit http://www.qrp4u.de/docs/en/rf_ampl/ (the last circuit in the page). I have tested this amplifier and I am very satisfied. It requires 1-3mW to output 1-1.5W or so.
I would like to modify the circuit so that the "diode" transistor is replaced by an external diode and this transistor will be used as a crystal oscillator that will drive the amplifier (other 4 transistors).

Can a single crystal oscillator transistor provide 1-3mW or it definitely needs a buffer?

I have tried to replace the broadband input transformer with an identical transformer as the one in the output, but this time I connected the variable capacitor at the input winding to tune it to 7MHz. However I could not make it tune to 7MHz, I could only make it tune to the harmonic of 14MHz, so I guess more turns are needed?
But since this transformer is trifillar I fear that increasing the number of turns will affect the low impedance required by the transistors bases.

Any help is appreciated.

You can use an 'external' (discrete) diode but make sure it is thermally bonded to the transistors. It sets their bias point and compensates for increased current (= higher temperature) by dropping the fixed bias voltage.

Generally, you should aim to keep crystal oscillator power as low as possible but you might be OK at 3mW. A buffer stage, even with zero gain is never a bad thing to have after an oscillator.

Brian.

betwixt said:

You can use an 'external' (discrete) diode but make sure it is thermally bonded to the transistors. It sets their bias point and compensates for increased current (= higher temperature) by dropping the fixed bias voltage.

Generally, you should aim to keep crystal oscillator power as low as possible but you might be OK at 3mW. A buffer stage, even with zero gain is never a bad thing to have after an oscillator.

Brian.

Click to expand...

This is exactly like this circuit **broken link removed** which uses a discrete diode for the bias. Thermal coupling is good but I have not seen any problems even wthout thermal coupling of the diode (practically speaking). I have even managed to blow up a chip by overdriving it (still at 12v) so I do not think the internal thermal coupled "diode" is of very much use, practically speaking.

What I was thinking (for simplicity), was to build something like this **broken link removed** but with the driver transistor to be used as an oscillator, not a buffer. It couples this transistor using the same input transformer (but currently I have only managed to make it resinate at 20m not 40. But I am not sure if I will be able to provide 2-3mW out of a single XO to drive the amplifier.

Any thoughts/ideas/schematics?