[ §2 · calculators · feedback strength ]

Feedback strength

This exists nowhere else. It says at what primary current the feedback loop takes control away from the self-oscillator: that is, whether the oscillator gets long enough to drag the coil onto the upper pole.

CT ratio N:1burden, total · Ωlead inductor · µHfrequency · kHzclamp threshold Vf · Vprimary current, peak · A

impedance |Z| 5.93 Ωfeedback strength 37.054 mV/A (0.0371 V/A)clamps open at ~18.9 A (11.8 % of peak) known window 20.000 – 35.700 mV/A low edge: davekni, simulation high edge: Rafft, working hardware ← weighs more Inside the window, nearer the top.Not much margin on the ‘the oscillator managed it’side: if it settles on the lower pole, drop the burden first.

formula

|Z| = √(R² + (2πfL)²)        strength = |Z| / N   [V per amp of primary]
clamps open at I = V_f · N / |Z|

limits

The metric itself is sound: |Z|/N, and comparing two builds by it, is arithmetic.

The absolute takeover current is not. What is computed is the moment the clamps open, which is not the same as the proven moment feedback wins. For that you need the amplitude of the oscillator’s ramp at the comparator input, and nobody has measured it.

The reference window is wide. davekni’s simulation (zero bridge delay, almost no lead) gives 0.020 V/A; Rafft’s working hardware at 49° of lead gives 0.036. The factor of 1.8 between them is the honest uncertainty. Rafft also runs a different comparator (MAX913) and R5 = 6 kΩ. Weight Rafft higher: it is iron, not simulation.