ADXL103: Root cause of temperature drift component

Document created by analog-archivist Employee on Feb 23, 2016
Version 1Show Document
  • View in full screen mode

What is the root cause of the "Temperature drift component (X-axis)". What is
the definition specified as "typ" compared to "max". Is "typ" equal to 1 std
dev? What I am requesting is more or less what is the standard deviation for
that component. Are there a more accurate distribution avalible?


The root cause of zero g tempco is mainly due to mechanical stress transferred
from the die attach material to the MEMS mechanical system. Not much stress is
needed. 1nm (1 x 10^-9 m) of movement of the mechanical system results in a
100mg apparent change! There is a small contribution due to electronic drift,
but not very much.

The mean zero g tempco of the ADXL203 is very close to zero and the standard
deviation is 0.08mg/°C. So the typical specification (±0.1mg/°C) encompasses
about 75% of parts. The maximum specification (±0.8mg/°C) is dictated by the
abilities of our production test system and not really reflective of the part's
capability. I've never seen an ADXL203 with tempco greater than 0.4mg/°C even
after looking at thousands of parts over several years. There might be one out
there (of the millions we've built), but I've never seen it or heard of its

The histograms shown in figures 6 and 9 are from a population of about 1600
parts from 3 manufacturing lots. I have repeated this test on newer parts
(about 2 years since my original data collection) and found similar results. So
the histograms in the data sheet are quite accurate.