Is My Calibration Good?
A quick way to check calibration is to take readings from internal ground channel AIN15 or any channel jumpered to GND. This is the midpoint of the AIN system, so if the device has changed and is out of calibration it will likely show up at this convenient voltage. Use Kipling or LJControlPanel to see the readings as you control Range and Resolution Index. The accuracy specifications are the same for the U6 and T7, and the -Pro versions of either: LabJack T7 Noise and Resolution Specs
- You only need to check one channel as they are all multiplexed to the same AIN circuitry.
- You do want to check each range of interest (±10V, ±1V, ±0.1V and ±0.01V), as each has unique errors.
- You do want to check both converters on a -Pro, so ResIndex 8 and 9 on a U6-Pro or T7-Pro.
- This should be done with the device stabilized at room temperature.
- To avoid confusion, this should be done with no other connections except comm/power.
NIST Traceable LabJack Calibration Service
Order CalibrationIf that ground test shows readings within specifications, you could then proceed to check some other voltages. Ideally a voltage around 10% and 90% of the range you are testing (e.g. -8 and +8 volts for the ±10 volt range), but a couple voltages from DAC0 (e.g. +2 volts and +4 volts) are fine also. Use any stable source to provide the test voltage, use a reference voltmeter (must be substantially more accurate than the LabJack specs) to measure the actual voltage, and then use Kipling or LJControlPanel to note the U6/T7 measured voltage. If all errors are within spec, you (or your calibration lab) can issue a new calibration certificate based on this verification of the current calibration. A 2-point verification is considered sufficient, as that is the minimum to notice a shift a slope or offset.
Note that verifying / calibrating the smaller AIN ranges can be very difficult because the noise of many voltage sources will be substantial compared to the error band of the U6/T7. ResolutionIndex can be used to reduce noise, but often additional oversampling and averaging has to be used to determine the average value of the source signal. You want to use enough oversampling such that the noise (difference between multiple averaged values) is reduced as needed, but want to average over the minimum required time period so you are not introducing actual changes in the source signal. The noise testing applications can be handy for this.
U3, U12, T4 verification: Same idea as above, with a couple details. For unipolar ranges (e.g. 0-2.4 volts), 0.0 volts might not be valid and thus you need 2 voltages in the 10% to 90% range. Every channel on the U12, and the high-voltage channels on the U3 & T4, have per-channel signal conditioning and thus each must be verified individually.
U12 calibration: The U12 is calibrated against its own reference voltage, which is a 2.5V signal that appears at the CAL terminal. The end of Section 3.7 of the U12 Datasheet describes how to do a self-calibration in the field, and in order to call this an absolute calibration you just need to measure the CAL voltage with a proper reference meter and confirm it is 2.49375 - 2.50625 volts.