N2VY: A 500 Watt HF Power Amplifier: Part 4

Caution: This project is in the design and simulation stage. Changes will be made as prototyping, building, testing, evaluation and commissioning proceeds.

This circuit is a copy of Jeff Anderson, K6JCA's directional coupler with some changes. I have chosen to use the Analog Devices ADL5904 for processing the forward and reflected voltages. It is specifically designed for protecting the inputs and outputs of RF amplifiers (among other things). This choice necessitates changing the attenuator circuit values. Here is my logic:

The input range of the ADL5904 is 45 dB from -30 dBm to +15 dBm
The highest power level in my various simulations without the need to trip the hardware circuit is about 550 watts.
I will design the attenuation chain to to set the ADL5904 input to 12 dBm for about 630 watts or 58 dBm of power.
This requires 46 dB attenuation from the output to the forward voltage port.
The directional coupler itself provides 28 dB of attenuation.
That leaves 18 dB of attenuation that I will implement in two pi attenuators, similar to the K6JCAs design.

Jeff, K6JCA, in his analysis calculates the maximum reflected power to be 56 watts with a 2:1 SWR or 75 volts peak into a 50 ohm load. This is exactly the number I get from LT Spice simulations across various loads.

I will keep the attenuation the same for the forward and reflected paths so once the power level is read into the Arduino A/D convertors, the level of mathematical gymnastics (and the associated errors and debugging time) is minimized. I may have to change my mind, we will see.

The ADL5904 has an absolute maximum at the RF input pin of +25 dBm. So, there is 13 dB of headroom between the design point and the absolute maximum rating.

I will set the hardware trip point for forward power at 550 watts or 57 dBm (which makes it closer to 500 watts but this can be tweaked during testing). After 46 dB of attenuation that is 11 dBm or 1.12 volts peak into 50 ohms. This is also consistent with simulation results (1.15 volts into 50 ohms).

I will set the hardware trip point for the reflected power at about 50 watts of reflected power or 47 dBm. After 46 dB attenuation, this is 1 dBm of power into the RF input of the chip and again, consistent with simulation.

So the forward and reflected power trip points are 11 dBm and 1 dBm respectively at the ADL5904 RF input.

Consulting the ADL5904 data sheet for setting the comparator threshold voltage without calibration (I will see if that is good enough during the testing or I will have to do some level of calibration);

For 1 dBm in the 10 to 30 MHz range: 269 and 266 mV. I will set it to 266 mV.
For 14 dBm in the 10 to 30 MHz range: 858 to 844 mV. I will set it to 851 mV.

The two ADL5904s are on the directional coupler sub assembly so there will be no need to ship RF around the box any more than necessary. The cost of this is a 3.3 volt regulator.

The schematic of the directional coupler and the signal processing circuits following it is below.