Class E Amp - Searching for the holy grail

I spent some time at work bench again this afternoon exploring how a Class E amplifier might work.

I started off by using a Dan Tayloe N7VE example of a "No Tune Class E" amplifier. I tweaked a couple of things for parts that I have on hand but the design is about 90% the same as what is in his presentation.

In my testing I am using a signal generator as my source and then sending the output to a directional coupler with a -20dB port for my microwatt meter and the other port into a dummy load. I am also using a my two channel digital oscope to be able to watch the before and after waveforms. In this testing I am not concerned about bandpass or lowpass filters since this is a prototype that will be talking to a dummy load. (In the real world it would need to be filtered since it is going to have harmonics that are in excess of the legal limit.)

Initially I did not bias the gate but I found that I did not have enough "drive" to turn on the 2N7000 effectively. The idea of the Class E amp is the FET is operating a switch, on or off in a semi-square wave pattern. The drain capacitor labeled C1 is supposed to be helping to charge and discharge and make the switch turn-on hard and then turn-off again.

In a tuned Class E amp L2 & C2 are a tuned circuit to create a fly wheel effect. In this "No Tune" example we are giving up efficiency/power since we want to tune across seveal bands.

I was able to get some gain but not nearly what I was expecting. Since I am pretty early into the testing and learning process this is an incremental step. It was interesting to see how the wave form changed and when the phase started to shift. When you reach full Class E mode the input and output sine wave will be out of phase. 

The images below were captured about 15 minutes before the data table (above) was created so the numbers maybe slightly different. The images from the scope are specific to ham bands since that was my area of interest. (The freq counter on the scope is close but not the exact frequency of the signal generator.)


I have a couple of other ideas in schematic form on paper at my desk. I am going to try to build a couple more prototypes and see how they perform. I might need to work on something band specific before I try to work out the wind band aspect of this.

I would like to be able to cover 40m, 30m, 20m, and 17m. I think that higher bands like 20m and 17m are going to be a challenge without a band specific C2 + L2 combination or a different FET.   I also need to look at this on my spectrum analyzer to see how the harmonics are looking.

Here are a couple of interesting links:

73 de NG0R