I tried to simplify Lucs concept even further by

• Removing the analog instrument, measuring bridge balance with the ADC of the micro controller. As the ADC port is high impedance, I added an 1K resistor in place of the instrument.
• Placing all components of the analyser on one PCB. As the counter code is fairly simple, I just moved everything into an ATMEGA8 cpu.
• Removing the gain pot. As the "virtual" instrument can not be overdriven, gain can be easily simulated in software.

So far the results have been very promising. In order to get a precise approximation of the original analog meter, I wrote the code in a way that instead of just setting a character to a black box, I defined 5 custom characters on the HD44780 display which allow me to display the meter in a very fine resolution (80 pixels instead of 16). This works really well. As the original meter did not have markings either, I felt this is a more than adequate replacement.

The frequency display works nicely too. I made the mistake first to believe it would be sufficient to feed the output frequency directly into the prescaler clock. This works but when in balance (50||50), the bridge draws too much current to keep the voltage high enough for the counter to work. Same as in Lucs original design I then added a buffer stage.

In general, the design works but requires the operator to understand what he is measuring since the square wave that is used for measurement produces "ghost resonances" at harmonic frequencies. Those would be at different impedances though so they can be identified.

The ANTAN B meter is a dirt cheap way to measure antenna impedances. The most expensive components were the multiturn potentiometer and the LCD bezel. Knowing its limitations, the meter is a great tool. The only cheaper alternative would be a noise bridge which works on a similar principle but requires a receiver.