so ive been working on this little 40m CW transmitter for the past few weeks, pretty basic stuff, just a crystal controlled thing based on a design i found in an old QST article from like 1987 or something. got the oscillator stage working fine, crystal is pulling to about 7.030 which is where i want it, but when i hook it up to the dummy load and watch the output on my scope the waveform looks kinda... mushy? not a clean sine, more like the top of the wave is getting clipped or flattened a bit.

im running about 12v in and the final is a 2N2222 which i know is a bit light but i was just trying to get something on the air to prove the concept. output is maybe 200mw into 50 ohms which is fine for now. checked all my solder joints, nothing cold that i can see. the low pass filter after the final is 5 element chebyshev cut for 40m, values came out of the ARRL handbook so those should be right.

not sure if the clipping is a problem with the driver stage not giving enough base drive, or if the 2N2222 is just being pushed too hard, or something with the supply. anyone dealt with this before? i dont have a spectrum analyzer so i cant check harmonics directly but im guessing theyre not great if the waveform looks like that.

yeah the 2N2222 is almost certainly your issue there, that thing really isnt meant to be a final even at QRP levels if you're trying to get a clean output. the absolute max collector dissipation on those is like 625mw and if youre getting 200mw out youre probably dissipating more than that depending on your efficiency. it'll run but it'll be unhappy and the distortion will show up exactly like you're describing.

try dropping in a 2N3904 as driver and put something beefier in the final, even an IRF510 would let you clean things up a lot and theyre like a dollar. also worth checking your dc bias on the base of the final, if the transistor isnt biased right into the operating region for class C it'll distort. for CW class C is fine but you need the drive level to actually be enough to swing it properly. the LPF should take care of most harmonics if its built right but you want the fundamental clean before it even gets there.

clipped waveform on a transmitter final almost always comes down to either too much drive or insufficient supply current. whats your power supply doing under load, have you actually measured the rail voltage when the key is down? a cheap wall wart will sag and that can do weird things to the output stage. i built a similar thing a while back using a 40673 MOSFET as the final and had this exact issue and it turned out my bench supply was current limiting without telling me about it.