Yeah, the formula I have (Greg Noonan was my source a couple of decades ago, and I'm sure I saw the same in something by George Fix too) is definitely an approximation, I suspect precision requires a regression equation. What's probably the key thing here is that the digression starts at the same point, and tracks closely, so I'd bet both equations are approximating the same "real" formula. And also, most importantly, it does suggest that up to a pretty high gravity, the inevitable errors in your gravity readings will be far more significant than the errors in your chosen calculator. Thanks again, cheers, enjoy your evening.
(In my equation things probably go pear shaped right off the top. That 0.8192 is an approximation of the relationship between real attenuation and apparent attenuation. Another nonlinear relationship, I believe. The SG equation is obviously doing the same, differently. )