The AM Forum

THE AM BULLETIN BOARD => Class E Forum => Topic started by: ka1tdq on October 03, 2016, 11:32:22 AM

Title: Class E and CW
Post by: ka1tdq on October 03, 2016, 11:32:22 AM
From previous discussions about class E and CW, keying the drain voltage is the way to go but it produces key clicks. How bad would it be? I'd like to build a two FET 40 meter CW rig and key the drains with a mercury wetted relay, for its speed.


Title: Re: Class E and CW
Post by: steve_qix on October 03, 2016, 10:02:39 PM
It can be done .... but  ;)

1) You have to shape the drain voltage or you will produce mondo key clicks.  You can use an L/C, or a MOSFET source follower, and shape the gate voltage.  The source will do whatever the gate is doing in a source follower.

2) If you don't key the drive, you will get backwave.  There is a lot of feed through with MOSFETs.  I've actually jammed a frequency once by only keying the drive, and leaving it on for a long time when I was testing.  Had no idea the signal was getting through so well !

The drive can be keyed unshaped (or lightly shaped) - and probably should be.  You want the drive up before the drain is keyed on, and you want the drive to stay on until the drain voltage falls to 0.

Title: Re: Class E and CW
Post by: ka1tdq on October 03, 2016, 10:19:54 PM
Well, that's just too difficult.  How do you expect people to be able to build that?   :)

I guess I'll just stick to modulating class E stuff.  Maybe a nice ricebox is in my future to fix my CW hankerings. 


Title: Re: Class E and CW
Post by: VE3ELQ on October 04, 2016, 09:14:30 AM
You could try keying the "enable" lines on the drivers, that may work.
73s  Nigel

Title: Re: Class E and CW
Post by: KD6VXI on October 04, 2016, 09:39:16 AM

A sdr with predistortion might be the ticket.   As well as a great receiver.


Title: Re: Class E and CW
Post by: ka1tdq on October 05, 2016, 07:12:47 PM
I never thought of keying the enable lines.  I always thought that you needed to have drive applied to the gate before you applied drain voltage in class E.  And during de-key you needed to remove drain voltage before you removed drive, or else bad things would happen.  Sort of like crossing the streams in Ghostbusters.  

Would keying gate drive work with drain voltage constantly applied?  (key clicks aside)

Jon (

Title: Re: Class E and CW
Post by: VE3ELQ on October 05, 2016, 08:02:37 PM
With no drive applied to the FET gates they are in cutoff and will draw zero current with the Vdd applied, no contactor needed.  The Enable lines are active high, so pull them to ground with a resistor about 1K or so then key them to Vcc (+15 volts) to enable the drivers and make RF. Should work but may produce key click harmonics.

73s  Nigel

Title: Re: Class E and CW
Post by: Tom WA3KLR on October 05, 2016, 08:30:34 PM
If you just turn-on and cut off the drive to the FETs, the drain voltage will swing much higher than it does in steady carrier operation.  If you don't have the voltage margin, the FETs will be destroyed.

 I plan to have a dc modulator mode for CW where the modulator voltage waveform is essentially trapezoidal with the keying and perhaps some other passive filtering also.   This will keep the FET drain voltage in the normal swing region.

Title: Re: Class E and CW
Post by: K1JJ on October 06, 2016, 12:10:42 PM
Hi Jon,

Over the years I have softened the keyed waveform for many CW rigs to cure them of key clicks.    By experimenting with various R, L or C  components in series or shunt with the keyed stage elements or wherever, we can usually get a nice smooth rise time and somewhat tapered decay. (make/break)

Look at the RF output on a scope while you send thru a series of dashes with the keyer. Adding in the R/L/C to the right spot will usually do the trick. Experiment with values and positioning.

This works well for tubes. But for FETS, Tom's voltage precautions above might be the ticket.  Though I know you usually run lower drain voltage on your rigs, so maybe you can get away with a simpler passive component approach.


Title: Re: Class E and CW
Post by: steve_qix on October 07, 2016, 08:11:16 AM
Never never key the drive with drain voltage applied.  This is a big no-no.

Also, you will generate severe key clicks even with over shaping.  This is because the gain of a MOSFET is so incredibly high.  The class E amplifier is a switching amplifier, and trying to operate it in any sort of linear mode (which overshaping the drive is) will just lead to problems (parasitics, etc.).

It is really not difficult to use a drain voltage shaping device.  The absolute easiest way to do it is probably with a source follower MOSFET (like an analog series modulation would use), and apply shaping to the DC voltage applied to the gate.

I quickly sketched this out.  It is NOT completely thought out, but gets the general concept and idea across.

Title: Re: Class E and CW
Post by: ka1tdq on October 07, 2016, 10:19:05 AM
That'll take me a while to analyze that one. Could you just do a quick-and-dirty method of just sticking a large choke in the drain supply?

I like oversimplification. Just like Heising modulating my 400 watt rig rather than using PWM.


Title: Re: Class E and CW
Post by: W1DAN on October 07, 2016, 01:40:51 PM

Here is a keyed power supply article:

Mates with their their 40m 500w CW transmitter I built years ago from this article:

For my AM rig, I leave the HV on at all times, and inhibit RF drive and the PDM generator's enable line. With no drive to the PDM generator, the PDM power transistors are at cut-off and so is the class E RF stage. I leave the main PS on, and ramp up the PDM generator's enable pin with a simple RC circuit.

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands