Calibration
Chapter 5
SCXI-1124 User Manual
5-4
© National Instruments Corporation
•
Code rounding error–When you want to output a voltage, you must write an integer bit
pattern. This can introduce up to 0.5 LSB of error. Most analog output boards do not include
this error because it is an inherent characteristic of a DAC. However, when hardware
calibration is used, the zero crossing offset can typically be trimmed to zero. You cannot do
this with the SCXI-1124 software calibration. Therefore, the code rounding error is included
in the error specification.
Your total error after calibration will be:
(1.5 LSB/4,095 LSB)*1.015 = 0.0372% of full scale
The extra factor of 1.015 is in the formula because the actual range of the circuit is 1.5% larger
than the nominal range. This accounts for the few LSB that you lose at the top and bottom of the
range. For a particular calibration, it is more accurate to use the number 4,095/(B
H
- B
L
) in place
of 1.015. You must also add any errors from your calibration equipment.
Current Calibration Method
When you calibrate a current output, use the two point calibration method described previously.
When you calibrate a current output, you cannot use 0 as a code. Because the current output
curve has a bend (see Figure 5-1), you must use two codes that are on the sloped portion of the
curve. You should use 4,095 as the upper code. For the lower code, using the code 255 is
recommended for simplicity. This code is guaranteed to be on the sloped portion of the curve,
but is still far away from 4,095. You can iterate the calibration procedure and use a code a few
bits above the B
L
, but this only improves your accuracy by at most 7 ppm of full scale.
The sources of error for current calibration are the same as for voltage calibration, plus an
additional error for using a point that is not an endpoint. This additional error is:
0.5 LSB*(4,095 - B
L
)/(4,095 - C
1
), C
1
≥
B
L
For C
1
= 255, and a typical B
L
of 37, the additional error is approximately 0.53 LSB. Your total
error after calibration will be:
(2.03 LSB/4,095 LSB)*1.017 = 0.0504% of full scale
The extra factor of 1.017 is in the formula because the actual range of the circuit is 1.7% larger
than the nominal range. This accounts for the few LSB that you lose at the top and bottom of the
range. For a particular calibration, it is more accurate to use the number 4,095/(B
H
- B
L
) in place
of 1.017. You must also add any errors from your calibration equipment.
EEPROM Limits
The EEPROM has a limited amount of storage space. Therefore, there is a limited range of
values for B
L
and B
H
that can be coded into the EEPROM. Valid values for B
L
and B
H
are:
0
≤
B
L
≤
255 and 3,840
≤
B
H
≤
4,095
