Simple RF square wave inverter question

[OgreVorbis]

So I am building a class D amplifier that operates in the RF range.
The input to the mosfet drivers of the amp take two square waves (5V TTL) where one is 180 degrees out of phase.
Currently I am using a signal generator, but I'd like to build my own driver.

To keep things easy (I am a beginner and not a real engineer), I just want to use a crystal oscillator. Using that oscillator, what is the simplest way to create the second wave that I need. I imagine I need logic inverter, but I've never made a circuit like this. What should I use? Keep in mind, I'd like it to be functional up to 20 MHz and low part count.

For a non-fixed frequency reference, I could probably use the cheap AD9850, but I really want something cleaner than that at the higher frequencies. Is there something similar to it (like better model) that I might be able to get on ebay as a module.

Thanks in advance!

[bd139]

Look at SiLabs Si5351A IC. Has two outputs which can have independent phase. Available on breakout boards for virtually nothing from china. Will go from around 1 to 150MHz and kick out square waves. You can set output current/drive in code as well.

Good tech info here:

https://www.qrp-labs.com/synth/si5351ademo.html

If you're using an arduino, then there's a suitable library here (I have used it when I was being lazy and it works nicely): https://github.com/etherkit/Si5351Arduino

[OgreVorbis]

Look at SiLabs Si5351A IC. Has two outputs which can have independent phase. Available on breakout boards for virtually nothing from china. Will go from around 1 to 150MHz and kick out square waves. You can set output current/drive in code as well.

Good tech info here:

https://www.qrp-labs.com/synth/si5351ademo.html

If you're using an arduino, then there's a suitable library here (I have used it when I was being lazy and it works nicely): https://github.com/etherkit/Si5351Arduino

Thanks for the reply. I'd really rather do this with a logic circuit than something I need to program. I am not using any microcontroller at all right now and I don't want to add one to the project.

I know a fair amount about electronics (I am an amateur), but I've just never had to do "logic" based circuits before.

I know this would add more complexity, but it would be nice if the duty cycle could be adjustable.

[nick_d]

You did not state if you want dead time?

The right way to do this with dead time is for the high side drive to have slightly higher duty cycle and the low side to have slightly lower duty cycle.

Suppose your cycle is 50ns (20 MHz) and you have complementary FETs (low is N type and high is P type). Try this:
0ns lo=0 hi=0: high drive active
20ns lo=0 hi=1: dead time
25ns lo=1 hi=1: low drive active
45ns lo=0 hi=1: dead time
50ns repeat

One way to achieve this is to generate two square waves of 50% duty cycle, identical except for a 5ns phase shift. Then AND them together to create the low drive and OR them for the high drive.

A good way to get a 5ns delay is a R/C network feeding a logic input. In fact, you only need the resistor if you are willing to rely on the stray input capacitance being relatively predictable. You could also do it using gate delays (a couple of inverters in series).

Check out the 74LVC1Gxx family.

cheers, Nick