LT3476 spice model results not agreeing with actual circuit results. The Spice model predicts a stable 1 MHz repeating duty cycle (75% SW closed) with only slight variation to maintain current at the target 335 mA target.
Looking at actual results from the circuit, provides very different results where the SW voltage, as measured at the connection to the inductor, shows an irregular pattern where the SW is on duty cycle > 95% for 5 us and ~7.5 us with no switching activity at all. The irregular switching activity shows significant variation in the target 350 mA drive current trace shown below it, generally aligned in time.
Model schematic shown below. Any ideas what is causing the difference between Spice and actual circuit performance? Any experiments or data that would help resolve the issue?
Scope showing proper duty cycle and no skipped pulses.