Precision of measurements?

[skim]

I'm curious if someone could offer an explanation regarding the precision of the diamond measurement tools. I understand a sarin machine is used to measure, but I've witnessed discrepancies that I cannot explain (or rather find an answer for). Is it generally understood that the "precise" looking numbers have variances?

I have 2 scenarios

minor discrepancy:

• original measurement done by AGS (unknown year, but based on serial number, it does not seem "old")
• new measurement completed by AGS today. (it was re-certified)
• differences (all minor, but different nonetheless):
  • gained 0.001 carat
  • table off by 0.2%
  • height off by 0.01mm
  • star off by 1 (don't know the unit of measure)
  • pavilion percentage off by 0.1%

significant discrepancy:

• original measurement done by AGS 2004
• new measurement done using AGS and GIA tech (from a well known vendor here) this month. Both show the "Sarin" logo on the app screenshot, so I assume it was using that tech for both.
  • At the time, I didn't scrutinize the numbers, and I guess I just trusted that it was done right (i.e. the machine/tool was properly calibrated, "they know what they are doing") - since, as I mentioned, the vendor is well known/respected here.)
  • Note: I paid the vendor a nominal fee to have the diamond shipped and analyzed (and returned since the analysis findings were not to my liking).
• differences
  • Measurements
    • Original: 6.67 - 6.72 x 4.09
    • AGS tool: 6.63 - 6.68 x 4.08
    • GIA tool: 6.67 - 6.72 x 4.073
  • CA/PA
    • Original: 34.7 / 40.8
    • AGS tool: 34.8 / 40.8
    • GIA tool: 34.58 / 40.69
  • Table
    • Original: 56
    • AGS tool: 53 <-- this is not a typo
    • GIA tool: 56.1
  • Depth
  • Original: 61.1
  • AGS tool: 61.3
  • GIA tool: 60.83

Thanks,
S

 

[Karl_K]

You can drive yourself nuts expecting to get different machines to agree.
Because they never will.

There 3 major suppliers of scanners each of which have several different models and accuracy levels.
Some of the machines report things different than others.
Each brand and sometimes each model within a brand uses different methods of rounding and averaging for the report.
Each brand and sometimes different models in the same brand calculate different measurements and measure others.
Even with the same brand and model there are different accuracy settings. Fast/standard/precise for example.
Then how well and often it is calibrated is a huge variable.

The things that I see that is hmmmm worthy is the 53 table and 34.58 / 40.69.
The table is the hardest measurement for the machine to get right because it is sitting on it but that is more than a bit off.
Ask them about them.