Amplifying a VHF Square Wave

[jgalak]

A discussion on another thread got me thinking.  I am building an APRS tracker using a clock chip as my rf source and modulator.  The output is an FM modulated square wave at 144 MHz and 3.3v (biased to 0-2.5V).  Obviously it will be low pass filtered veloce going to the antenna to kill the odd harmoics.  But it needs to be amplified - I'd like around 250-500mW, ideally,

My original plan was to use an off the shelf MMIC IC, but some comments on the other thread mentioned Class E amplifiers and amplifying the square wave itself.  I started reading about them (can't say I understand most of it yet), and it looks like a lot of the work and components needed for a Class E are specifically for dealing with the harmoincs.  It's almost like they are taking the input sine wave, turning it square, and then back to sine.  Which got me wondering:

Is there a good way to just amplify the square wave to begin with?  Seems like it should be easier and more efficient than amplifying a sine.  Maybe a push-pull with no worry about crossover distortion?  Obviously low pass filtered afterwards.

I've thought about using a logic level shifter, but the ones that go to 144MHz seem to be limited to 5V, which is a bit lower than I'd like.

Thoughts?  Are there standard ways to amplify a square wave? 

[rf-fil]

A well designed class C will usually give you something like 60% to 70% efficiency. With Class E, you can get up to 90+%. But, at your power level, the actual improvement, in absolute terms, is probably not worth it. Unless it's a learning exercise. You'd be reducing the heat dissipation by something like 100mW.

Class E amps have less gain compared to a class C amp with the same device. They are harder to design. And they need a good match at the output - so they are only suited to situations where the load impedance is well controlled. This might not the case for a mobile APRS tracker.

Btw, RF Power Amplifiers for Wireless Communications by Steve Cripps is one good book to learn about RF PA design.

[jgalak]

I'm wondering if a real amplifier topology is overkill altogether.  It's a signal that moves between 3.3v and 0v.  Can I amplify with a single transistor (bjt or fet) acting as a switch?  No need for linearity, it's on or its off.  Only needs to switch about 10V.  Just need one that's fast enough to handle 144MHz.

Wpuld that work?