Test and Measurement Forum

Logging data from a Wheatstone Bridge

Post discussions on projects you are working on

Logging data from a Wheatstone Bridge

Postby Gerry » Wed Aug 03, 2016 10:01 am

We often get requests from customers wanting to log data from resistor based sensors that have very low sensitivity, such as strain gauges. So, what follows is a run through of how to acquire data from them using Pico Technology data acquisition products.

The proven method to maintain accuracy with low level resistive transducers without noise corruption, and without the need of a voltage reference, is to use either an Anderson loop, or Wheatstone bridge. The Anderson loop is relatively newer, and not as well supported (there is a large body of background information available on the internet for Wheatstone bridges) and it will have similar obstacles to overcome to those in a Wheatstone bridge, so we will focus on the Wheatstone bridge.
The most straightforward way to implement a strain gauge in a Wheatstone bridge circuit is to use a bridge completion module, or load cell which comes pre-configured with the gauges and cabling. Typical requirements for an excitation voltage to power these are 10V DC, ideally from a constant current source. The PicoLog 1000 series data loggers and ADC 20/24 data loggers have a constant voltage source for providing 2.5V of power to circuits (e.g. to establish a bias for converting double ended outputs to single ended). However, because of the typical resistor values used, the Wheatstone bridge configuration would be too big a load for the limited current output, so a separate excitation supply is required for powering the bridge circuit.

The Wheatstone bridge combined with it's excitation voltage generate a signal output that has 2 signal voltage points and a different reference voltage point relative to its excitation voltage. The difference between the 2 signal voltage points is the differential voltage, and the difference between the reference point and the lowest signal voltage is the common mode voltage. So, it must be either interfaced to a Data Acquisition device that also uses differential inputs, or the Wheatstone bridge signal output must be conditioned so that it can be interfaced to a Data Acquisition device that doesn't use differential inputs.

Using Single Ended inputs

1000 series PicoLog and associated Small Terminal Board
Our PicoLog 1000 series data loggers have inputs that are only single ended, and have an associated Terminal board (https://www.picotech.com/accessories/mi ... inal-board) that can be used for interfacing and signal conditioning.
In a Wheatstone bridge the voltage range of the resistor used, as a transducer for applied pressure, only changes over a very small range in comparison to the total output voltage of the bridge circuit. So, this means that there is a large common mode voltage, which must be removed to be able to interface the signal output to a PicoLog 1000 series data logger with a large enough input range, to give the sort of detail required for measurement.
One idea to remove the common mode voltage would be to associate the data logger ground with a reference point, relative to the bridge signal output, that is low enough so that the whole signal remains positive, and high enough to bypass the common mode voltage (or as much as possible). It is possible to do this if the excitation supply to the Wheatstone bridge has outputs that can be floated from ground, or uses a battery with a long enough lifespan for data logging sessions. A resistor divider could then be used to establish a voltage equivalent to the –ve output of the bridge (or as close to it as possible). However, this method would introduce gross errors in the data (the resistor divider will introduce an additional current path, which means that even with a constant current source, the current through the bridge will vary with varying load, affecting accuracy, and because of the resistor values that would be required in the divider in order to straddle the common mode voltage range, there will be large changes in the reference point voltage relative to the changes in the bridge output voltage, affecting resolution) and it would also introduce common mode noise at a relatively high level in comparison to the signal level.

So, the best way to interface the bridge signal output to a PicoLog 1000 series input, and remove the common mode voltage (and noise) would be to effectively turn a single ended input channel of the data logger into a differential one by inserting a differential amplifier, between the bridge and the data logger. The differential amplifier can either be built and implemented, or bought (as an industrial signal conditioning amplifier module) to avoid some circuit layout and soldering work. If the amplifier draws less than 10mA it can be powered from the PicoLog 1000 series 2.5V voltage source, otherwise a portion of the Excitation voltage from the power source for the bridge, stepped down to 2.5V, could be used (or all of the excitation voltage can be used but the output swing of the differential amplifier should then be limited to 2.5V).

So, to setup up the PicoLog 1000 series data logger with a differential amplifier module, the Excitation power supply needs to be connected to the excitation inputs of the bridge, and the amplifier power supply to the differential amplifier supply inputs. Then the -ve signal output from the bridge needs to be connected to the inverting input of the differential amplifier and the +ve signal output from the bridge to the non-inverting input of the amplifier. Next, using the small terminal board the +ve signal output of the amplifier needs to be connected to the Channel 1 input of the terminal board, and the -ve output of the amplifier to a ground terminal of the terminal board.

For a worked example we will use a load cell that has a sensitivity of 2mV/V at a rated load of 1Kg (so for a 10V supply that would be a strain sensitivity of 20mV/Kg). We will have a differential amplifier gain of 250, and encounter a maximum of 0.5Kg of strain in our application (giving a maximum output of 20x10^-3x250x0.5 = 2.5V from the amplifier for maximum strain).

To setup the software for the values used in this example, you need to set up input scaling for the input channel so that a maximum output of 2.5V at the data logger input, corresponds to the 500g maximum strain value from the gauge in the data log file. So, start PicoLog and set-up the recording, and sampling, as required. Next, from the 'Settings' menu option, select 'Input channels' to open the 'Converter details' dialogue box, expand the drop down list for 'Converter type and select 'PicoLog 1012/1216', then click on OK to open the 'PicoLog 1X1X measurements' dialogue box for your model of data logger. Click on 'Add..' to open the 'Edit PicoLog 1X1X Measurement' dialogue box. Type something like "Pressure" for the name, click on 'Options', and then 'OK' to 'Save changes to channel settings...' and open the 'Parameter options' dialogue box. Click the check box to 'Use Parameter Formatting', and type in the units specified for sensitivity (e.g. 'g' in our example), then in the text boxes for 'Scaling for graphs' type in the absolute minimum and maximum values expected for your units of pressure during your data logging session (0 and 500 for the example). Click on 'Scaling', and from the drop down list for 'Scaling Method' select 'Table look-up'. Click in the window at the bottom and type the following:
[minimum voltage from the amplifier] [corresponding minimum pressure value]
[maximum voltage from the amplifier] [corresponding maximum pressure value]
So for our example we would enter the following into the window:
0 0
2.5 500
(note the space between the numbers on each line) This makes PicoLog equate 0 V to 0 grams, and 2.5 V to 500 grams.

You can then click on OK, to exit out of the open windows to complete the setup. You would now be able to log the maximum difference voltage in 10-bit resolution using the PicoLog 1012, or 12-bit resolution using the PicoLog 1216.

So, when you create a data file and start the recording, you would be able to open the graphing tool and see just one difference waveform scaled in units of grams.

Using Differential inputs

ADC 20/24 data logger and associated ADC 20/24 Terminal board
Our ADC-20/24 data loggers have inputs that can be either single ended, or paired and separated from a ground reference so that they can be used as with differential signals, and have an associated Terminal board (https://www.picotech.com/accessories/mi ... inal-board) that can be used for interfacing and signal conditioning. Using these data loggers to interface to a Wheatstone bridge, provides the most straightforward setup procedure and best resolution and accuracy. The hardware setup is as follows:
Connect the +ve excitation power lead to the +ve excitation input of the sensor
Connect the 0V lead to the -ve excitation input of the sensor
Connect the +ve signal output of the sensor to the Channel 1 input of the terminal board
Connect the -ve signal output of the sensor to the Channel 2 input of the terminal board
Because the ADC 20/24 has Galvanically Isolated Digital and Analogue grounds, the ADC 20/24 inputs, if not connected anywhere, will give fluctuating readings of several millivolts as they drift according to the charge accumulated and dissipated at the input. So, to tie the reference point of the analogue-to-digital converter in the ADC 20/24 to a stable voltage level, you can connect the Analogue Ground connection of the ADC 20/24 Terminal board (marked AG) to both the -ve output of the sensor, and the channel 2 input of the Terminal board.

Using the same application example specified in 'Using Single Ended inputs', start PicoLog and set-up the recording, and sampling, as required. Next, from the 'Settings' menu option, select 'Input channels' to open the 'Converter details' dialogue box, expand the drop down list for 'Converter type and select 'ADC-20/ADC-24', then click on OK to open the 'ADC-2X Channels' dialogue box for your model of data logger. Select Channel 1, and click on 'Edit' to open the 'Edit ADC-20 or ADC-24 Channel' dialogue box, type something like "Pressure" for the name, select a conversion time of between 60ms & 660mS (depending upon what resolution you require - as detailed on page 11 of the user guide, see here: https://www.picotech.com/download/manua ... -guide.pdf) & select +/-39mV as the Voltage range (as we need maximum sensitivity for the low level un-amplified sensor output). Finally tick the box for "Differential input enable" (this effectively disables channel 2 to from being used independently from channel 1).

So, to complete the setup, for the differential input example scaling file we would enter the following into the window:
0 0
0.01 500
Note that the input stage of the ADC-20/24 data logger converts the 2 differential signals to 1 single ended signal before they reach the Analog-to-Digital Converter, so the actual digitized values used in the scaling file are the same values as those converted by the separate differential amplifier in the previous example, i.e. 0 V equates to 0 grams, and 10 mV to 500 grams. For our worked example you would now be able to Log the maximum difference voltage in a resolution up to 17-bit resolution (10mv on a 78mV scale) using the ADC-20, and 21-bit resolution using the ADC-24. However, actual maximum sensor output signals could be significantly more than this which would increase the effective resolution.

PicoScope 3425
Typical applications for sensors in Wheatstone bridge configurations don't require high speed sampling, but where the maximum sample conversion time for a data logger is not fast enough, the PicoScope 3425 (see here: https://www.picotech.com/oscilloscope/3 ... 5-overview) can be used as the acquisition device. The PicoScope 3425 has differential inputs that can also be configured as single ended.
However, it has a lower input sensitivity than the ADC-20/24 (+/-100mV as opposed to +/-39mV) and a converter resolution of 12-bits, which means that, the eventual resolution of the data is effectively 8-bit.

So, to setup the hardware, the connections are even simpler than those of the ADC-20/24, as follows:
Connect the +ve excitation power lead to the +ve excitation input of the sensor
Connect the 0V lead to the -ve excitation input of the sensor
Connect the red lead of the differential probe to the +ve signal output of the sensor
Connect the black lead of the differential probe to the -ve signal output of the sensor

Although it has a much more straightforward hardware setup, it has a more convoluted software setup. We need to use our PicoScope 6 software as the 3425 differential scope does not work with our PicoLog software. PicoScope 6 doesn’t have scaling files but it does have custom probes which you can set up in a similar way to customize the scale for the measurement being performed. To set up a custom probe for our example, click on the ‘Channel options’ icon (in the 3rd menu bar from the top it’s the one with the channel label and arrow) and click on the icon for ‘Probe’ with the ellipses on it to open the ‘Custom Probes’ window. Click on ‘New Probe’, then ‘Next’, and check the ‘Use the custom unit defined below’ radio button. Enter ‘grams’ and ‘g’ for the full and short name of the units, click next and check the ‘Use a look-up table’ radio button, and click on the ‘Create a Look-up Table..’ long button to open the Look-up table Scaling window. Next click in both fields of the first two rows in the window below the units, and enter the following:
0 0
0.01 500
Next Click on OK to exit to the Custom Probe Wizard window and click on the ‘(Advanced) I will manage the Custom Probe Ranges Manually’ radio button, to prevent PicoScope from selecting anything other than your manual choice (which will need to be the highest sensitivity setting of +/-100mV for maximum resolution). Click on ‘Next’, then click on ‘New Range’, and check the ‘Use this hardware input range’ radio button to enable the +/-100mV range choice (as you can see this uses 5% of the range). Finally click on ‘Apply’, click on OK twice, and then Next twice to get to the ‘Custom Probe Identification’ window. Enter ‘Wheatstone Bridge’ for the name and a suitable description, then click on Next to exit the Probe setup. Now you can click on OK to select the Custom Probe and exit the list of probes.
This makes PicoScope equate 0 V to 0 grams, and 10mV to 500 grams. For our worked example, although only 5% of the scale will be used, you would still be able to Log the maximum difference voltage in just under 8-bit resolution (10mv on a 200mV scale, which you would need to zoom vertically to see). However, actual maximum sensor output signals could be significantly more than this which would increase the effective resolution.

As PicoScope 6 is primarily analysis software for data based around a specific event, as opposed to PicoLog, which is logging software for data capture to a file over a long time period (long in comparison to PicoScope 6), it will ordinarily save data to a memory buffer and then stop. So, we need to set up the PicoScope to automatically dump the buffer contents to a file and, once it has been written, restart capturing data to save to the next file. This can be set up as follows:
Under the ‘Tools’ Menu select ‘Alarms’, click on the down arrow to expand the drop down ‘Event’ list, then tick the check box for ‘Capture’. Next click on ‘Add’, expand the drop down list for ‘Action’, select ‘Save current buffer’, then click on the button with ellipses in the ‘File’ window’ and navigate to where you want the files to be stored, name the files (which will have incrementing numbers at the end of the name when more than 1 file is created), and then click on ‘OK’ to exit back to the ‘Alarms’ window. Next click on ‘Add’ again, expand the ‘Action’ drop list, and select ‘Restart capture’ then ‘OK’ to exit back to the ‘Alarms’ window. Click on ‘Apply’, and ‘OK’ to complete the file auto-save setup, then on the menu bar at the bottom of the display, expand the drop down list for ‘Trigger’ and select ‘Single’. If the trigger has been setup to occur correctly, the software will now auto-save the buffer to files until the triggers stop.

There will be a gap between capturing the last sample, then saving the buffer to a file, and re-arming the trigger, then capturing the first sample for the next file (dependent upon the amount of data to be saved). This therefore limits the length of time over which the contiguous data can be captured and saved in a file. The limit is determined by (a) the time that it takes to capture one complete buffer of data, or (b) the maximum time period over which PicoScope 6 can buffer data to be displayed on the screen (which is limited by the longest available Time-base of 13 hours and 53 minutes), whichever is the shorter. It is possible to capture data from the PicoScope 3425 for longer periods if using the Software Development Kit (see here: https://www.picotech.com/library/oscill ... nt-kit-sdk) to create a streaming data application.

Using Double ended inputs

Using a non-differential PicoScope for the logging means that you would have to interface the bridge to both negative and positive going (double ended) inputs. Contrary to what is normally expected, using double ended inputs is actually a disadvantage because (a) you have the same issue as using single ended inputs, i.e. you would need to use an amplifier to avoid reduced accuracy and resolution, and increased noise, and (b) you have to limit your amplifier selection to one providing a +ve or -ve swinging output voltage, or use an amplifier with a single ended output, and be limited to the +ve half of the PicoScope input (losing half of the range). There are also more disadvantages, and as this is a relatively poor solution we won't continue with any more setup details for it.
Gerry
Technical Specialist
Gerry
PICO STAFF
PICO STAFF
 
Posts: 62
Joined: Mon Aug 11, 2014 11:14 am

Return to Projects

Who is online

Users browsing this forum: No registered users and 0 guests