Tuning filters with a NanoVNA

[mstangelo]

If I buy a kit is it possible the tune the filters with a NanoVNA with the final transistor in place. In DL2MAN's video he mentions you can tune the harmonic reject filter but not the bandpass because the final output is not 50 ohms. What is the output impedance of the final transistor in resistance and parallel capacitance?
I can probably build an attenuator to match it to 50 ohms.

Mike N2MS

[LILO]

It makes no sense.
There is no bandpass to tune, all you have to do is to put lpf notch on 2d harmonic for maximum suppression. That's all.

[dl6sez]

Hello OMs,

as LILO stated 100% correct, it just makes no sense!

73 de Chris

[kf7yrs]

From the video, I understand it the way Mike described. I think class-e amplifiers require a tuned circuit to convert the square wave output to a sine wave. One of the inductors, and a few of the caps, serve that function and DL2MAN stated in the video that he could not tune that one with the VNA since it was not 50 ohms. He basically adjusted the turns on that inductor until the output power and efficiency were where he liked them to be, about 5W and >70%, as I recall.

The other inductor and its accompanying caps are a notch filter aimed at removing the first harmonic. DL2MAN used the VNA to adjust the inductor and move the notch to the correct frequency.

I found adjusting both inductors (/ band) to be a less than trivial process. At first my radio was outputting less than 2-3 W per band even with the wires spaced as widely and evenly as possible. Squeezing the windings together lowered the power, so I ended taking out turns. This got all bands to 5W or more and ~70%. I had to take 2 turns of the 80m band and 1 off most of the others, a bit of a pain but worth it in the end.

After getting the power and efficiency to where I wanted them, I worked on the notch. I had previously adjusted them but now they were off. Before I readjusted the notch, I looked at the output on a VNA (with a 30dB attenuator inline!) and saw that the harmonics were down at least 45dB even though they did not look ideal using the DL2MAN-VNA method. I left good enough alone

Lee

[mstangelo]

Sorry I did not make my question any clearer but KF7YVS got it right. What makes aa amplifier a Class E is that the tramsistor has to switch on when the output voltage is zero. Refer to this presentation:

https://web.ece.ucsb.edu/Faculty/rodwel ... ClassE.pdf

The output series bandpass circuit has to be resonant but meet the above criteria. It would be nice to have the simulated output so that one could tune with the NanoVNA to get in the approximate range.

Mike N2MS

[dl6sez]

Hello OM,

what should this simulation show? Every FET you use differs slightly in spec that's the main reason why one has to adjust also Lx1 to match the FET(s). This is also the reason why it is not used in commercial transceivers. You can only match 50 Ohm exactly at a single frequency point that's also clear.
So for me your simulation idea is still useless.

73 de Chris

[SP9TKW]

I don't remember where I found suggestion, which I proposed to use 4:1 balun at the input side of RF Board and connect S21 to antenna output from RF board. That setup allowed me to tune LPF.

Info I found, stated that input impedance on RF board is about 10-12ohm, so 4:1 would match it to the 50ohm from nanoVNA, antenna connection is 50ohm by default.

I got more than 5W on every band on classic bands, which I made from initial batch of *Rxxxx* boards. I removed all LPF caps on that board and installed caps as per specification. I also used new cores.

Efficiency was between 78% and 82% depending on the band.

I will be building new TRX shortly, so that will give me a chance to confirm method again. The initial (tr)uSDX was sold to SOTA activator going on 6mths trip. Up till now, I made 9 activations with over 500 QSOs using that rig.

I hope that helps.