That Kang T/R switch is interesting, gets around the need for a PIN diode.
As in the stock rig, pin 2 of T7 drives the LPF and the antenna via two 0.1uF caps in parallel, or 0.2uF.
At 7mhz, the impedance of the two 0.1uF caps is 1/(2*pi*freq*farads) = 1/(2*3.14*7e6*0.2e-6) = 0.1 ohms,
so 500 times less impedance than the 50 ohm impedance of the transmit LPF, and thus of no consequence..
During receive, he drives the base of Q1 through 47+4.7 pF of capacitance, or about 50pF,
At 7mhz, an impedance of 1/(2*pi*7e6*50e-12) = 455 ohms.
That 50pF is still there during transmit, so about 10% of the transmit energy is getting shorted to ground through
that 455 ohms and the two back-to-back diodes (the diodes prevent Q1 from blowing out on transmit).
Also, the 455 ohms means that the receiver gets a little bit less signal than on a stock rig
So extremely weak signals might be harder to hear, but will have less distortion receiving very strong signals.
So the 47+4.7pF caps are chosen to be a good compromise between transmit and receive performance at 7mhz,
I'm OK with losing 10% of my transmit power..
On the uBitx we want this to work from 3.5 to 30mhz.
Perhaps could do this by mechanically switching in different values of capacitance to replace that 47+4.7pF
depending on the operating frequency. Or go with one of the 1n4007 PIN diode T/R switch designs.