Based on maukka’s numbers I projected the following data for a Convoy Biscotti driver and an SST-20. Does this look right? The 1/10th resolution lumens are directly from his data, the mutliple decimal resolution lumens are from a 6th degree polynomial I created in Excel based on his numbers. Clearly the fit isn’t perfect as your approach 0mA but I bet it isn’t too far off (estimate 0.5lm at 2.8mA).
