No matter how I tried, I could not drop the second harmonic to make 43 dBc. Best I could get was about 39. That's not far out of FCC regulations and is probably well within the suppression required in other parts of the world. Probably it is just my particular board, the minor variations of the capacitors and inductors combining to just miss the mark. I haven't heard of anyone else reporting it but I will pass this along just in case. If you are using an external amplifier then I am certain that the filtering on that amplifier will solve it easily. You could also use an external low-pass filter.
I ran a simulation of the filter bandpass using the AADE software. The results are shown in simulation as the upper trace. You will notice that the suppression provided at 14.4 MHz (the second harmonic) is around 26 dB. The lower trace is around 50 dB at that frequency and is the result of paralleling L7 with a 100 pF capacitor. This extra capacitor acts like a trap for the second harmonic. I find myself installing this on most of the transmitters that I design, just to save an additional filter section.
I just soldered the 100 pF disk across that inductor on the underside of the board. I don't notice any difference in operation but I am now legal. The spectrum analyzer displays show the results before and after the modification.
Has anyone else experienced this problem? If so, here is a cheap and easy fix.