We have in the past been accused of being paranoid with our curve tracer matching.

But you can see in the distortion measurement posted above, especially how low second harmonics is.

This can only be achieved when you curve trace under operating conditions.

And then match to 3 coefficients in a 6th order polynomial of Id vs Vgs.

The 0th order, i.e. Vgs only ensures lowest DC offset.

The 1st order, i.e Ygs, ensures equal current sharing.

Only when the 2nd order coefficient is also identical that you will get perfect even harmonic cancellation.

As another example, TI said in one of their documents :

https://www.ti.com/lit/an/sloa054e/sloa054e.pdf
**6. Reduced Even-Order Harmonic Distortion**
Expanding the transfer functions of circuits into a power series is a typical way to quantify the distortion products.

Taking a generic expansion of the outputs and assuming matched amplifiers, we get:

Vout+ = k1Vin + k2Vin2 + k3Vin3 + . . . , and

Vout– = k1(–Vin)+ k2(–Vin)2 + k3(–Vin)3 + . . . . Taking the differential output

Vod = 2k1Vin + 2k3Vin3 + . . . , where k1, k2 and k3 are constants.

The quadratic terms gives rise to second-order harmonic distortion, the cubic terms gives rise to third-order harmonic distortion, and so on.

In a fully-differential amplifier, the odd-order terms retain their polarity, while the even-order terms are always positive.

When the differential is taken, the even order terms cancel. Real life is not quite this perfect.

Lab testing of the THS4141 at 1 MHz shows that the secondharmonic at the output is reduced by approximately 6 dB

when measured differentially as compared to measuring either output single-ended.

The third harmonic is unchanged between a differential and single-ended measurement.
So if you use a BB Chipamp in bridge mode, you only get 6dB H2 cancellation.

Unless you also measure and match them.

Cheers,

Patrick