Analog vs Digital

[McTeeth]

Hello everyone,

So...I did our 1st impression scanning on our E1 and printed the models/die on our Asiga Max. Our milled emax (designed on the scanned impression) fits tight on the stone model vs the printed model (Whip Mix Ivory OS). We have success with Asiga models with Trios impressions. I see slight dimensional differences between printed die vs stone (that's probably stating the obvious)
Where do I start?
- Is it the scanner?
- Is it the die interface?
- Is it the printer?
- Is it the resin?

Thanks,
Sean

 

[CoolHandLuke]

this is a job for Metrology.

so, job 1 is to scan your printed model and compare it with the scan of your stone model. download GOM inspection software (free) and use it to Align both scans, and then use the built in comparison tools to compare one scan to another. for this to work simply import one scan as CAD body.

job2, scan your emax and compare with the designed tooth.

this way you will be able to visualize the aspect of error, and pinpoint where it began to corrupt in the process. this way you will also be able to visualize how one swing in the +/- accuracy can eviscerate the quality of a case. scan +/- 20um (twice) and built in error of the printer (30um) gives you a total allowable error of 70um, so in your comparison in GOM ignore any data between -70 and +70 um

the same for your milled emax. but i dont know what you are using to mill so ive elected to not include an example on the grounds i think i've made it clear already.

 

[McTeeth]

this is a job for Metrology.

so, job 1 is to scan your printed model and compare it with the scan of your stone model. download GOM inspection software (free) and use it to Align both scans, and then use the built in comparison tools to compare one scan to another. for this to work simply import one scan as CAD body.

job2, scan your emax and compare with the designed tooth.

this way you will be able to visualize the aspect of error, and pinpoint where it began to corrupt in the process. this way you will also be able to visualize how one swing in the +/- accuracy can eviscerate the quality of a case. scan +/- 20um (twice) and built in error of the printer (30um) gives you a total allowable error of 70um, so in your comparison in GOM ignore any data between -70 and +70 um

the same for your milled emax. but i dont know what you are using to mill so ive elected to not include an example on the grounds i think i've made it clear already.

Dam...I best be learnin some new stuff

 

[JayH]

This is assuming, of course, that the poured model is the gold standard and that if anything is off in your tests, it can't possibly be the poured stone model. Maybe your die-stone expanded more than true?

 

[CoolHandLuke]

he said he did impression scanning

so there would be a model from the impression scan, and a pour.

given its tight on the pour ...

 

[JayH]

I read it correctly...

Finish your thought so I can follow your reasoning

 

[CoolHandLuke]

if its tight on the model then something happened during pouring, or some deficiency was observed in the impression scan, like it missed a corner of a die (happens often)

this translates directly to the mill, if the mill was 100% accurate it will also miss the corner. this will create an emax that fits on the theoretical impression, but not the stone.

if the mill was not 100% accurate, but like most emax mills pretty poorly adaptive to internal contours, this too should be visible by metrological scan comparisons.

if however, the model and the impression scan line up perfectly, we can know for certain its a milling issue. if the scans do not line up perfectly, we know it was a scanning issue. if its a little of column A and a little of column B, well its a learning curve.

 

[JayH]

Gotcha. Now I understand what you're saying.

If the scans don't line up perfectly, we don't know it was a scanning issue though - it could still be error in the stone expansion.

While impression scanning with line-of-sight scanners will probably never be 100%, one of the best reasons for a doctor to go IOS is, in my opinion, it eliminates any error caused by die stone expansion/contraction and associated environmental factors. The IOS scan should always be more accurate. But because labs have been pouring stone forever that change of standard is not often thought about.

 

[CoolHandLuke]

true enough, in a general sense, but IOS still is photogrammetry based. stitching algorithms being what they are, 0.5um over 10mm square of accuracy, turns into 0.2mm discrepancy over an arch and gets worse with more data, good for general purpose dentistry, not good for precision manufacturing. singles and small bridges dont count as precision design because they only fall within the 10mm square window, not the 75mm arch. \

but yes, you are completely correct, stone expansion being what it is, its not likely the scan and the model will line up at all,which should be an indication its time to get new stone or revise mix/pour practices with an automatable solution like a wasserman auto-dispensing scale. practices like not flipping impressions, and using plastic sheet.

 

[JayH]

...stitching algorithms being what they are, 0.5um over 10mm square of accuracy, turns into 0.2 mm...

Do you have a reference for these numbers?

It's been years since I've had a formal math course but assuming I'm interpreting what you say correctly, I only come up with a max 0.071 mm error over 100 square millimeters. ?

 

[CoolHandLuke]

its numbers i calculated many moons ago in another thread

suffice to say its compound error, it is not a single line of photos with even spacing, its uneven and its in X and Y directions at tilt and varying focal lengths.

photos in a straight line, or taken with an overarching photo as control data, are helpful controlling accuracy, but no IOS ever so far, has this.

trios will claim its 8um accurate, but they don't say how many photos it takes to throw that accuracy off.