Oh. This is not nice. We have a third harmonic peaking way up there near 90 dB. This in turn sets our SINAD which is a sum of distortion and noise to the same value. With CD's noise floor being -96 dB, we have distortion products which would peak above that.
This preamplifier was measured by stereophile which stated it had exceptional performance. I am not seeing any signs of that, nor where there enough documentation in said measurements for me to try to fully replicate them.
To be fair, the preamplifier has a rated output of 1V and is clearly designed to achieve its lowest THD at levels typical for power amplifier sensitivities of 1-2V (as was/is the norm). The Stereophile plots also show this to be perfectly true. You are testing it at 4V, 4 times its rated specification. Of course the THD rises at 400% of rated spec.
Specification is 200mV for 1V out. You are feeding in 4V. 20 times its rated sensitivity for full rated output. A CD at full level is 2V.
Also, the preamplifier has a gain of 14dB. Even with the gain pots backed right off (-10dB) there is still a gain of 4dB. How then did you test at 4V in and 4V out without winding back the master volume pot from maximum? Once the master volume pot is not at maximum, your tests cannot be compared to Stereophile's (vol max (as is standard testing protocol), input shorted etc). The Stereophile 1:1 (unity) 2nd plot shows the FR deviations, but all other tests are at full volume position and adjusting the input level.
Also, just as an aside, why a 300ohm load? It's not a headphone amplifier- why compare it to one? The line outs are 60ohm rated per leg for the balanced and 60ohm for the single ended. A typical load would surely be several k ohm.
The Stereophile tests do show exceptional performance in my opinion.