# LTSpice: why does my .MEAS 3dB calculation disagree with the cursors on the plot?

I'm exploring the .MEAS command with the following simple BJT Class-AB amplifier:

And am getting the following results from my .MEAS commands in the error log:

vout_max: MAX(mag(v(vout)))=(27.4778dB,0°) FROM 10 TO 1e+010
vout_3db: vout_max/sqrt(2)=(24.4675dB,0°)
f_3db: freq=(142.323dB,0°) at 1.30662e+007


The results for vout_max and vout_3db seem correct, they agree with the plot to at least 2 decimal places.

I'm very puzzled by the results for f_3db however. It appears to call V(vout) crossing the -3dB point at 13.06MHz, where the plot (screenshot below) has it at 19.41MHz.

• Why are these numbers so different?
• Have I written these .MEAS commands wrong?

I expected these to agree to at least several decimal places given they're both calculated from the same simulation run.

• None of your results match the schematic. This is not a Class AB amplifier, rather it is a unity voltage gain common collector(aka emitter follower) bipolar amplifier with an overall current gain < 2 due to 1000/500 ratio. – Tony Stewart Sunnyskyguy EE75 Nov 21 '19 at 1:00

AC analysis produces complex numbers with 'real' and 'imaginary' parts that carry the magnitude and phase information in a vector. The function MAG() is used to extract the magnitude.
You have done this for vout_max, but not for V(vout), so the comparison will give unexpected results. Change V(vout) to MAG(V(vout)) in the last .MEAS command and it should find the correct frequency.