It's interesting how John Rice was talking about post production results. Traditionally make 10 parts, break up 3 and statistically work from there. In 3D printing though every part can be 'wildly' different, so statistical analysis doesn't cut it. Only by inspection of every single part can real inspection be done, and the only way to do that is in-situ monitoring. As we know, Sigma uses in-situ monitoring to make sure the build is within the parameters chosen by the manufacturer to qualify an acceptable build.
This concept is becoming more prevalent.
1. How every 3D machine will produce different results every time due to the physical nature of the building process, thus rendering statistical process control ineffective. If each part is different, then each part needs to be inspected somehow. Traditional methods don't do this.
2. How important it is for the part to be the constant, not the machine, again the opposite of traditional manufacturing.
This is the big difference from traditional manufacturing, but if 3D is to move into production, these key points to be addressed. As he mentioned Sigma is leading this technology by a decent margin. It may all be slow, but at this stage the alternative options are very few and far between, certainly at any commercial level like PrintRite3D is available, let alone being independent of the machine makers themselves, which as has been mentioned before is important.
Can anyone tell me a better way to inspect every part that is commercially available now? This is a crucial question. If we go towards production how else will every part be inspected? CT scans are not practical for every part. Present OEM inspection systems collect the data but lack the alorithms to make that data actionable, and are not independent of the machine.
Every parts material properties will be different. How else to inspect?
anyone?
