This is an old revision of the document!
ADALM1000 Calibration Procedure
The ADALM1000 device can be calibrated by the user to achieve more accurate measurements and signal generation. The calibration procedure requires a Digital Multimeter (DMM) and a resistor with a value anywhere between 2.5Ω and 25Ω.
The device is calibrated by taking a set of measurements according to the procedures described in the sections below. The measurement results must be hand written into a calibration file that will be afterwards used by the calibration software to compute all the calibration parameters and write them into the device.
Calibration support is available only stating from Pixelpulse v0.88 and firmware version v2.06 (You don't need to look at the firmware unless you want to - it's only added here for completeness).
Make sure to update the firmware on your device and Pixelpulse before starting the calibration procedure.
Calibration file
The calibration files contains <reference, value> data pairs for all the measurement and source channels. Below is an example of how a default calibration file looks like.
# Channel A, measure V
</>
<0.0000, 0.0000>
<2.5000, 2.5000>
<\>
# Channel A, measure I
</>
<0.0000, 0.0000>
<0.1000, 0.1000>
<-0.1000, -0.1000>
<\>
# Channel A, source V
</>
<0.0000, 0.0000>
<2.5000, 2.5000>
<\>
# Channel A, source I
</>
<0.0000, 0.0000>
<0.1000, 0.1000>
<-0.1000, -0.1000>
<\>
# Channel B, measure V
</>
<0.0000, 0.0000>
<2.5000, 2.5000>
<\>
# Channel B, measure I
</>
<0.0000, 0.0000>
<0.1000, 0.1000>
<-0.1000, -0.1000>
<\>
# Channel B, source V
</>
<0.0000, 0.0000>
<2.5000, 2.5000>
<\>
# Channel B source I
</>
<0.0000, 0.0000>
<0.1000, 0.1000>
<-0.1000, -0.1000>
<\>
The calibration file must contain for each channel at least the data points shown in the example above. It is possible to add as many data points as desired, the only constraint is that the measurement for offset calibration (0V or 0A reference) has to be the first data pair in the data set. A data set for a channel is delimited by the </> and <\> tags. The order of the data sets in the calibration file must be as in the example and the calibration file must contain all the shown data sets otherwise it is not valid.
The example calibration file can be downloaded from the link below. This file assumes perfect device operation - no offset or gain errors, and can be used to reset your device calibration in case an inaccurate calibration file was loaded into the device.
If an inaccurate calibration file was loaded into the device this can be erased by writing into the device the default calibration file.
When you want to re-calibrate your device, before starting the calibration process write into the device the default calibration file and power cycle the device. This will erase the previous calibration and will ensure correct re-calibration.
Storing the calibration data into the device
To store the calibration data into the device:
The smu binary also supports displaying and resetting calibration for all supported, attached devices via the following:
Make sure that Pixelpulse is closed when running the smu binary.
To acquire the smu binary, use the libsmu installer below in addition to installing the correct architecture support for Microsoft's Visual C++ Redistributable for Visual Studio 2015.
Instructions on how to build, install and use libsmu can be found here.
If you have multiple ADALM1000 devices make sure that just one device is plugged in when trying to store the calibration data.
Calibration procedures
The procedures below describe all the required steps to perform and record the calibration measurements. The Voltage Source of the ADALM1000 is very precise and normally does not require calibration so it can be left to the default values in the calibration file.
Voltage measurement calibration
Voltage measurement calibration requires at least 2 data points - 0V and 2.5V. The steps below describe the calibration procedure:
On the ADALM1000 device connect the CH A and CH B inputs to GND
In Pixelpulse set the mode for both channels to Measure Voltage
Using Pixelpulse read the voltage measurements for the two channels
Replace into the calibration file in the
# Channel A/B, measure V sections the
<0.0000, 0.0000> data pairs with
<0.0000, gnd_value>, where
gnd_value is the voltage measurement read in Pixelpulse
On the ADALM1000 device connect the CH A and CH B inputs to 2.5V
Using Pixelpulse read the voltage measurements for the two channels
Using a DMM read the actual value of the 2.5V
Replace into the calibration file in the
# Channel A/B, measure V sections the
<2.5000, 2.5000> data pairs with
<dmm_value, pixelpulse_value>, where
pixelpulse_value is the voltage measurement read in Pixelpulse and
dmm_value is the 2.5V measurement taken with the DMM
Voltage source calibration
Voltage source calibration requires at least 2 data points - 0V and 2.5V. The steps below describe the calibration procedure:
In Pixelpulse set the mode for both channels to Source Voltage Measure Current
In Pixelpulse make sure that the Repeated Sweep mode is not active
From Pixelpuse source 0V on channels A and B
Using a DMM measure the voltage outputs of CH A and CH B relative to GND
Replace into the calibration file in the
# Channel A/B, source V sections the
<0.0000, 0.0000> data pairs with
<pixelpulse_value, dmm_value>, where
pixelpulse_value is the voltage source value in Pixelpulse and
dmm_value is the measurement taken with the DMM
From Pixelpuse source 2.5V on channels A and B
Using a DMM measure the voltage outputs of CH A and CH B relative to GND
Replace into the calibration file in the
# Channel A/B, source V sections the <2.5000, 2.5000> data pairs with
<pixelpulse_value, dmm_value>, where
pixelpulse_value is the voltage source value in Pixelpulse and
dmm_value is the measurement taken with the DMM
Current measurement calibration
Current measurement calibration requires at least 3 data points - 0A, 100mA and -100mA. Channels A and B must be calibrated sequentially. The steps below describe the calibration procedure:
On the ADALM1000 leave both CH A and CH B open
In Pixelpulse set the mode of the channel to be calibrated to Source Voltage Measure Current
In Pixelpulse read the current measurement for the channel to be calibrated
Replace into the calibration file in the # Channel A/B, measure I sections the <0.0000, 0.0000> data pairs with <0.0000, gnd_value>, where gnd_value is the current measurement read in Pixelpulse
On the ADALM1000 device connect the CH A or CH B to one end of the resistor
Connect the other end of the resistor to the current input of a DMM
Connect the GND of the DMM to the ADALM1000 2.5V pin
In Pixelpulse source a voltage value on the channel to be calibrated so that the current measurement is close to 100mA
In Pixelpulse read the current measurement for the channel to be calibrated
Read the current measurement on the DMM
Replace into the calibration file in the # Channel A/B, measure I sections the <0.1000, 0.1000> data pairs with <dmm_value, pixelpulse_value>, where pixelpulse_value is the current measurement read in Pixelpulse and dmm_value is the current measurement taken with the DMM
In Pixelpulse source a voltage value on the channel to be calibrated so that the current measurement is close to -100mA
In Pixelpulse read the current measurement for the channel to be calibrated
Read the current measurement on the DMM
Replace into the calibration file in the # Channel A/B, measure I sections the < -0.1000, -0.1000> data pairs with <dmm_value, pixelpulse_value>, where pixelpulse_value is the current measurement read in Pixelpulse and dmm_value is the current measurement taken with the DMM
Current source calibration
Current source calibration requires at least 3 data points - 0A, 100mA and -100mA. Channels A and B must be calibrated sequentially. The steps below describe the calibration procedure:
On the ADALM1000 device connect the CH A or CH B to one end of the resistor
Connect the other end of the resistor to the current input of a DMM
Connect the GND of the DMM to the ADALM1000 2.5V pin
In Pixelpulse set the mode of the channel to be calibrated to Source Current Measure Voltage
In Pixelpulse source 0A on the channel to be calibrated
Read the current measurement on the DMM
Replace into the calibration file in the # Channel A/B, source I sections the <0.0000, 0.0000> data pairs with <pixelpulse_value, dmm_value>, where pixelpulse_value is the current source value in Pixelpulse and dmm_value is the measurement taken with the DMM
In Pixelpulse source 100mA on the channel to be calibrated
Read the current measurement on the DMM
Replace into the calibration file in the # Channel A/B, source I sections the <0.1000, 0.1000> data pairs with <pixelpulse_value, dmm_value>, where pixelpulse_value is the current source value in Pixelpulse and dmm_value is the measurement taken with the DMM
In Pixelpulse source -100mA on the channel to be calibrated
Read the current measurement on the DMM
Replace into the calibration file in the # Channel A/B, source I sections the < -0.1000, -0.1000> data pairs with <pixelpulse_value, dmm_value>, where pixelpulse_value is the current source value in Pixelpulse and dmm_value is the measurement taken with the DMM
Semi-automated Procedure
The ALICE 1.1 DeskTop Software includes a semi-automated method to calibrate the ADALM1000 using the AD584 precision 2.5 V reference from the ADALP2000 Analog Parts Kit. Just wire up an AD584 for a 2.5 V output, powered from the fixed 5V supply from the ALM1000 and follow the steps as outlined in the ALICE desktop user's guide. No DMM needed (except to maybe verify the output of the AD584).