i haven't followed the numbers, math, and goals very closely. so i'm taking a lot on faith. brian's statement about the density difference works out to about 10%. again, i'm assuming everything being said by everyone is accurate. brian thought that density difference was crazy. to me, it's not that surprising. density accuracy is going to depend on the weight and volume accuracy. so a combined diff of 10% doesn't seem that bad to me. if you look at the resolution of scales, the fact that they may not be calibrated, and so on. they really aren't very accurate. especially as the load capacity increases. as for volume, i was originally assuming that you were using a number calculated by the cad software. then it seemed maybe you were measuring volume. either way, depending upon how complex the part is the theoretical volume vs real volume could be causing issues.

not sure if you have ever been to a standards lab, where they measure all sorts of things. density, volume, and so on. it's incredible how much goes into just this topic. you can spend a lifetime just in calibration and measurement. it's super expensive and time consuming. so the 'we aren't going to mars' aspect is important to keeping cost down. so going from 10% error down to 1% starts getting very expensive and time consuming. i do think it's worthwhile to check into the simple and cheap things that could give you problems though.