Class E amplifier design transistor problem

[Minchuu]

Hello!

I am designing a Class E amplifier. Its input signal is a BPSK signal below of around 4vpp and I need it to be amplified at least 15.4 dBm.

Furthermore, it needs to be able to drive a 50 ohm load. I used the Class E component calculator to solve so that R = 50 because that's my given load.

This is my schematic:

but the result named [Output] is distorted. In fact, it even distorted my BPSK signal.

I was thinking if maybe my transistor was the problem, but I've tried most of the transistors here and it still didn't work. Do you guys have any idea? I chose Class E because it's the most convenient because of the calculator (and because it can achieve 100% efficiency theoretically and I will only simulate anyways), but maybe it's impossible to achieve my goal with Class E? Or maybe the transistor is the problem?

Thank you very much!

 

[barry]

I don’t think you can apply a 4 V signal directly to the base of a transistor like that. Think about it:Vbe of the transistor is about 0.6V, B-E is essentially a diode. As the voltage rises, the base will just suck more and more current until something breaks. At the very least, put a resistor in series with the base. Further, if your signal is centered around zero volts you’ll need a capacitor or some other biasing scheme to accommodate that.

 

[BigBoss]

If you apply 4Vpp to B-E Junction, the model of the BJT will not be valid anymore.Models have sustainable limits, not infinite.
In additional to, why Class-E amplifier ?? You got a power consumption constraint ?The transistor which you have chosen may not be an appropriate one because this transistor is designed for A-Class Small Signal, Low Noise Amplifiers.
If your BPSK Signal Source is able to drive 50 Ohm Load, connect it directly to the Antenna because 4Vpp@50 Ohm will give 16 dBm Power Strength.

 

[Minchuu]

Further, if your signal is centered around zero volts you’ll need a capacitor or some other biasing scheme to accommodate that.

Putting in the resistor did help, thanks for that. Unfortunately, biasing the amplifier requires polar modulation, something too advanced for a newbie like me. I'll try to explore other topologies. Thanks for the feedback!

If you apply 4Vpp to B-E Junction, the model of the BJT will not be valid anymore.Models have sustainable limits, not infinite.
In additional to, why Class-E amplifier ?? You got a power consumption constraint ?The transistor which you have chosen may not be an appropriate one because this transistor is designed for A-Class Small Signal, Low Noise Amplifiers.
If your BPSK Signal Source is able to drive 50 Ohm Load, connect it directly to the Antenna because 4Vpp@50 Ohm will give 16 dBm Power Strength.

Yep, I get it now. I think the Class E is really not the right one for this project. Class D has been recommended to me previously but I thought that since this has a calculator it would be easier. That's what I get for going for the easy way out, I guess. Thanks for the feedback!

 

[dick_freebird]

500kHz is way out from where I've seen people
care about Class E (granted, my only real exposure
to RF was working at a RFIC company that was
chasing cell phone front end sockets, so 1-2GHz;
Class E is good for when your transistors do not
have much linear gain left @ freq, but still can
switch bang-bang).

Output tank is key to removing all the harmonics
but output tank can't be both that effective, and
also agile. A question for you to chase is, is Class E
suited to phase-shift keying (where output filter
acts as a "flywheel") or what trades must be made
between spur / harmonic suppression, and phase
shift keying agility (symbol rate).

In simulation you can impose kooky base voltage,
current will be limited by the modulated Rb and
you may fall off the beta-vs-Ic curve if you're above
knee current but that ought only to happen after
you have switched the load low. Beware saturation,
though - if your RF transistor models that. You might
end up just pumping the base and stuck low, if you
do saturate the BJT (and the model has anything
like realism, for that).

Back in the day I worked with 500M-1GHz fT BJTs
that would have saturation recovery times exceeding
1uS. 1uS "recovery time" vs 1-10nS turnon time, and
that's without trying to jam amps of base current
into a small signal transistor.