Text preview for : dsol11.pdf part of Fluke dsol11 Fluke 720A doc dsol11.pdf



Back to : dsol11.pdf | Home

Design Solutions 11
November 1999

Testing Linearity of the LTC2400 24-Bit No Latency TM
A/D Converter
Help from the Nineteenth Century

by Jim Williams

Introduction follower is required to unload the KVD without introducing
significant loading error. Now, our KVD looks like Figure 3.
Verifying the linearity of the LTC2400 analog to digital
converter requires special considerations. Typical
nonlinearity is only 2ppm (0.0002%). Bench testing this 10k 2k 400 80
necessitates some form of voltage source that produces INPUT

equal amplitude output steps for incremental digital in-
puts. Additionally, for measurement confidence, it is de-
sirable that the source be substantially more linear than
the 2ppm requirement. This is, of course, a stringent
demand and painfully close to the state of the art.
The most linear "D to A" converter is also one of the oldest:
Lord Kelvin's Kelvin-Varley divider (KVD), in its most
developed form, is linear to 0.1ppm. This manually switched
device features ten million individual dial settings ar- 80
OUT
ranged in seven decades. It may be thought of as a
80
3-terminal potentiometer with fixed "end-to-end" resis-
tance and a 7-decade switched wiper position (Figure 1).


SEVEN-DECADE SWITCHED
R = 100k WIPER POSITION PERMITS
SETTING TO 0.1ppm LINEARITY


DSOL11 F01


COMMON DSOL11 F02
Figure 1. Conceptual Kelvin-Varley Divider
Figure 2. A 4-Decade Kelvin-Varley Divider. Additional Decades
The actual construction of a 0.1ppm KVD is more artistry Are Implemented By Opening Last Switch, Deleting Two
Associated 80 Values and Continuing