I did some testing on my v4 board and found interesting correlation (though not able to verify mathematical accuracy):
The lower the signal at TP1 in CW transmit, the lower the output power. This would have nothing to do with the LPFs either in the front end prior to the transmit amps, nor on the output.
This went on until above 14mhz, where the power starting dropping with out a decrease in level at TP1.
To me this suggests that there is something in the way that the design unbalances the first mixer to allow for the "bleed through" of the changed frequency of Clock#2.
I then used a constant tone through my mic for the ubitx using a signal generator connected to a powered speaker. I kept the postion of the mic constant and changed the bands. With this test, the 7.2mhz signal had more power than the 14.3mhz signal. This suggests that the transmit power that others and I have been measuring is not indicative of the SSB power.
Again, this points to the CW process of unbalancing the first mixer has having side effects that impact power on some of the bands.
I should also note that the signal I measured with my scope was not at all sine wave clean in CW mode. It was almost square. Would also explain the large amount of harmonics that I have measured with my inexpensive RF Explorer SA.
At this point I will be going back and coming up with a better way to inject a 1 or 2 tone audio signal into the uBitx to then measure the harmonics.
Has anyone else seen this type of behavior?