The single most informative thing in your data is something the script's warning text glosses over:
RIAA can't be the cause here, because RIAA would rotate both channels by the same amount, and your two channels differ by 50°.RIAA at 399 Hz contributes roughly 55° of phase lag, identically to L and R. If it were in the chain, you'd expect φ_L ≈ φ_R ≈ (true phase − ~55°). Instead you have +127.9° and +178.4°. That asymmetry has to come from something channel-dependent, which points squarely at cause 2 for channel L.
Here's the decomposition that makes everything click. Split each channel's FM deviation into the in-phase (real, cos φ) component, which carries the actual VTA information, and the quadrature (sin φ) component, which is tracing distortion:
- Ch L: 41.8 Hz × cos(127.9°) = −25.7 Hz in-phase, +33.1 Hz quadrature
- Ch R: 29.0 Hz × cos(178.4°) = −29.0 Hz in-phase, +0.8 Hz quadrature
The in-phase components are nearly identical. Once you project out the junk, both channels agree, which is exactly what you'd expect since there's only one physical stylus angle. Channel R is essentially a clean measurement: φ = 178.4° is just "in-phase but inverted," i.e., a legitimate negative-sign error, not a fault condition at all. Channel L has the same underlying signal plus a large quadrature contamination that rotated its phase to 128° and inflated its apparent deviation to 41.8 Hz.
This has two uncomfortable consequences for the summary panel:
- The L channel's +4.01° reading is an artifact. The script appears to take sign/magnitude from the raw deviation rather than the in-phase projection, so L's answer is corrupted by the quadrature term. Using the script's own implied calibration (~10 Hz per degree, from 29.0 Hz → 2.89°), the corrected L in-phase deviation of −25.7 Hz gives roughly −2.5°, in close agreement with R's −2.89°.
- The "AVG VTA 17.06°, no adjustment needed" verdict is probably wrong. It's the average of one valid reading (13.6°) and one contaminated one (20.5°), and they cancel by coincidence. The consistent story from both channels after correction is a VTA of about 13.6 to 14°, roughly 2.5 to 3° below θR = 16.5°. If that holds up, you'd want to raise the tonearm pivot, not leave it alone.
On your note that "both channels say distortion is low": the green checkmark is looking at the 399 Hz second harmonic (−45 dB), which is a different quantity from quadrature FM of the 3960 Hz carrier. The two don't have to agree. And in fact the rest of your data quietly corroborates that L is the worse channel: AM IM% of 5.58 vs 4.06, d2H of 6.23% vs 5.20%. Something is producing asymmetric tracing behavior on the left groove wall contact. The usual suspects are azimuth error, anti-skate side force, VTF at the edge of spec, or asymmetric stylus wear.
