Precision Temperature Measurement with the ADS1248

Precision Temperature Measurement with the ADS1248
Joseph Wu Senior Applications Engineer Texas Instruments – Tucson The ADS1248 is one of our latest 24-bit ADCs, optimized for temperature measurement applications. This Live Lab will use the ADS1248EVM-PDK and ADCPro to demonstrate thermistor, RTD, and thermocouple measurement solutions. After attending this session, you will be fully prepared to discuss these measurement applications, and demonstrate them to your customers. 2009 European FAE Summit, Munich

Presentation Overview
An Overview of Temperature Elements The ADS1248 and ADCPro Precision Measurements with the ADS1248 This presentation is organized into three sections. First, there is an overview of different types of temperature elements that can be measured. Next, there is a section covering the ADS1248 and some of its special features used in temperature measurements. This section also covers the ADS1248EVM and the MMB3 motherboard and the program used to control the ADS1248EVM- PDK called ADCPro. Finally, there are several examples of systems showing how the ADS1248 can be used to make temperature measurements. In this Live-Lab presentation, we have the ADS1248EVM being run by ADCPro and we can see how these measurements are taken. 2009 European FAE Summit, Munich

What sort of temperature elements can we measure with the ADS1248?
In this section, we’ll introduce three different type of temperature elements and how they are used. Each type of element have their advantages and disadvantages. For each there are also differences in how they are setup and applied. 2009 European FAE Summit, Munich

Temperature Monitoring - RTD
Source: Advanced Thermal Products, Inc. RTD: resistance temperature detector Positive temperature coefficient Wire-wound or thick film metal resistor Manufacturers: Advanced Thermal Products, U.S. Sensors, Sensing Devices Inc. The first temperature element is the RTD RTD: resistance temperature detector – [This measures temperature through a linear resistance change with temperature (over a limited temperature range)] Positive temperature coefficient Wire-wound or thick film metal resistor Manufacturers: Advanced Thermal Products, U.S. Sensors, Sensing Devices Inc. 2009 European FAE Summit, Munich

C C A A A PRTD PRTD PRTD B B B Most RTDs are available in two, three, and four-wire lead configurations. The two-wire configurations requires fewer connects to the element which relaxes the board and electronics requirements. However, the parasitic resistance at room and over temperature of the wires combine with the low resistance value of the RTD element. The three-wire and four-wire configurations are most common in precision circuits. Both of these RTD devices can be configured with an amplifier or instrumentation amplifier so that the wire resistance at room and over temperature is significantly reduced or eliminated from the circuit’s output. These different configurations are used to minimize the effect of the parasitic resistance on the measurement. D a.) Two-wire lead configuration b.) Three-wire lead configuration c.) Four-wire lead configuration 2009 European FAE Summit, Munich

Advantages: Most Accurate High linearity over limited temperature range (-40oC to +85oC) Wide usable temperature range Advantages: Most Accurate - sensitivity available to +/-0.1°C High linearity over limited temperature range [for example -40°C to +85°C] Wide usable temperature range - [through using linearization or calculation, - 250°C to 600°C (ASTM) 850°C (IEC)] (ASTM and IEC are different standards detailing equations used to describe the polynomial equations governing the RTD behavior over temperature) 2009 European FAE Summit, Munich

Disadvantages: Limited resistance Low sensitivity Lead wire resistance may introduce errors Requires linearization for wide range Wire wound RTDs tend to be fragile Cost is high compared to a thermistor Disadvantages: Limited resistance range [typically 100Ω to 1kΩ but down to 10Ω and up to 10kΩ are available] Low sensitivity, [about Ω/°C for a Pt100 RTD – this is nominally a 100Ω RTD at 0°C] Lead wire resistance may introduce significant errors Requires linearization for wide range – [for a large temperature range for example -200°C to +850°C] Wire wound RTDs tend to be fragile Cost is high compared to a thermistor – [but it has a significantly wider temperature range] 2009 European FAE Summit, Munich

Temperature Monitoring - Thermocouple
Source: Datapaq Thermocouple: temperature element based on two dissimilar metals The junction of two dissimilar metals creates an open circuit voltage that is proportional to temperature Direct measurement is difficult because each junction will have it’s own voltage drop Thermocouple: temperature element based on two dissimilar metals The junction of two dissimilar metals creates an open circuit voltage (called the Seebeck voltage) that is proportional to temperature Direct measurement is difficult because each junction will have it’s own voltage drop 2009 European FAE Summit, Munich

Source: Agilent Reference (Cold) Junction Compensation Voltage is proportional to Temperature V = (V1 – V2) ~= α(tJ1 – tJ2) If we specify TJ1 in degrees Celsius: TJ1(C) = tJ1(K) V becomes: V = V1 – V2 = α[(TJ ) – (TJ )] = α(TJ1 – TJ2 ) = (TJ1 – 0) V = αTJ1 Measurement of the voltage across the thermocouple requires some sort of reference or cold junction compensation. This is because each junction of dissimilar metals will create it’s own junction voltage. In the above diagram, we have a voltmeter on the left and copper wires going to the thermocouple. The thermocouple in this example is made from junctions of copper and constantan. Here we have a known junction and ice bath given by J2. The temperature measurement will be at J1 Since the J2 is at a known temperature (zero degrees C), the temperature can be calculated based on the temperature coefficient for the thermocouple. In this case, the voltage measured at the voltmeter is proportional to the temperature referenced to 0 degrees Celsius. However, the ice bath can be replaced with any cold junction with a known temperature and the difference can be calculated. Most calibration tables or formulae are tied to a cold junction temperature of 0C. 2009 European FAE Summit, Munich

Advantages: Self-powered Simple and durable in construction Inexpensive Wide variety of physical forms Wide temperature range (-200oC to +2000oC) Advantages: Self-powered, no excitation necessary Simple and durable in construction Inexpensive Wide variety of physical forms Wide temperature range (-200oC to +2000oC) 2009 European FAE Summit, Munich

Disadvantages: Thermocouple voltage can be non-linear with temperature Low measurement voltages Reference is required Least stable and sensitive Requires a known junction temperature Disadvantages: Thermocouple voltage can be non-linear with temperature [for large temperature ranges] Low measurement voltages (a K-type thermocouple has a sensitivity of 40uV/oC) Measurement based on an absolute voltage means a reference is required It is the least stable and sensitive Requires a known junction temperature 2009 European FAE Summit, Munich

Temperature Monitoring - Thermistor
Thermistor: Thermally sensitive resistor Sintered metal oxide or passive semiconductor materials Suppliers – Selco, YSI, Alpha Sensors, Betatherm Thermistor: Thermally sensitive resistor Sintered metal oxide or passive semiconductor materials Suppliers – Selco, YSI, Alpha Sensors, Betatherm 2009 European FAE Summit, Munich

Advantages: Low cost Rugged construction Available in wide range of resistances Available with negative (NTC) and positive (PTC) temperature coefficients. Highly sensitive Advantages: Low cost option for less critical applications Rugged construction Available in wide range of resistances: 100Ω to 40MΩ Available with negative (NTC) and positive (PTC) temperature coefficients. [NTC is most common] Highly sensitive: [value -3.9% /°C to -6.4% /°C for an NTC thermistor] In general, I would imagine the thermistor is less likely going to be used in conjunction with an ADS1248 for several reasons. Most who want to buy the ADS1248 are willing to pay more for the accuracy of an RTD or thermocouple. The range of the thermistor is less than either of the other options. The sensitivity of the thermistor is such that the 24-bit ADS1248 is overkill. However, look for the ADS1148 coming soon! 2009 European FAE Summit, Munich

Disadvantages: Limited temperature range: -100oC to 200oC Highly non-linear response Linearization nearly always required Least accurate Self-heating Disadvantages: Limited temperature range: -100°C to 200°C Highly non-linear response Linearization nearly always required [which may limit the usable range] Least accurate [50 deg C range for 10-bit accuracy] Self heating may be a problem [Since the thermistor requires excitation and has a high sensitivity, self heating may be an issue] 2009 European FAE Summit, Munich

What can we do with the ADS1248 and its EVM?
Now that we’ve covered some of the temperature elements that are available. What can we do with the ADS1248 and its EVM? The ADS1248 was design specifically for the measurement of temperature. It has many features that make it easy to interface with the devices in the previous section. 2009 European FAE Summit, Munich

ADS1248 Block Diagram 2009 European FAE Summit, Munich
Here is a block diagram of the ADS1248. Going through the list of features, this delta sigma data converter has: Eight analog inputs, switchable with an analog multiplexer An internal, low noise and low drift 2.048V reference, and two external reference inputs An internal oscillator, but with a single pin input that can be used an external clock A low noise PGA with 8 gains from 1 to 128V/V Ten different data rates from 5 to 2k samples per second An internal VBIAS that sets up a a voltage of (AVDD-AVSS)/2, useful for setting up voltages for an unbiased thermocouple Two current DACs that can be used to set up biasing for an RTD or a bridge. Note that the IDAC current can be output to either IDAC external pins or directly to any of the analog inputs. 2009 European FAE Summit, Munich

ADS1248EVM-PDK 2009 European FAE Summit, Munich
Here is a picture of the ADS1248EVM along with the MMB3 motherboard controller. This attaches to your computer through a USB and can be powered entirely through the USB cable. In this Live Lab demonstration, we have one right here! 2009 European FAE Summit, Munich

ADS1248EVM Schematic 2009 European FAE Summit, Munich
This busy diagram is the schematic for the ADS1248EVM. I wont spend too much time on it except to point out a few key features. Input signals can be applied to the analog header or the screw-in terminal blocks. The two IDAC outputs can also be accessed through the terminal blocks. There is a buffered REF5020 onboard the EVM that can be used as the external reference for the ADC. The EVM has circuitry that allows for the analog supplies to be run at either +5V or +/-2.5V. If you run this EVM as a standalone board, you must consider where to apply the power and account for that. One important thing to note is that the inputs all have some low pass filtering. If you run the IDAC through to the input pins, there will be some current drop through the resistors, which will affect measurements. For this demonstration, I have removed the resistors R4 and R5 from AIN0 and AIN1 and replaced them with 0 ohm resistors. 2009 European FAE Summit, Munich

ADS1248EVM Layout 2009 European FAE Summit, Munich
This is the basic layout of the ADS1248EVM. Blocks in red are the default jumper settings. Again, the analog header and the screw-in terminal blocks can be found on the left side of the EVM. Terminal blocs for the two IDAC outputs and a Ground can also be found on the left side. The reference and buffer can be found on the top left side of the EVM. As mentioned in the previous slide, R4 and R5 have been replaced by 0 ohm resistors. These located at between the analog header and the ADS1248 at the top of the column of resistors. 2009 European FAE Summit, Munich

ADCPro with the ADS1248 Plug-in
Here is a screen shot of ADCPro running with the ADS1248EVM Plug-in and the Multichannel histogram tool. The plug-in is installed on the left, with the histogram tool on the right. 2009 European FAE Summit, Munich

ADS1248 Plug-In 2009 European FAE Summit, Munich
This is a screen shot of just the ADS1248 Plug-in for ADCPro. Here, we can set the gain, and the data rate at the top with two pull-down menus. At the bottom, we can power down the device with a button click. On the bottom right and indicator bar shows the progress of taking a sample of data. In the middle of this screen shot is the first of five tabs for the ADS1248 Plug-in. This tab is labeled “I/O Config” with the title on the left side. Here you can control the input multiplexer. Each of the eight analog inputs (AIN0-AIN7) can be set as the AINP or AINN input to the ADC for a measurement. Additionally, each input can be tied to VBIAS (to serve as the DC bias of a floating thermocouple). Each input can also be tied to the output of one of the two current DACs. This is a handy feature for driving current to an RTD. Also, AIN2-AIN7 can be used to serve as GPIOs when selected in the column. Note, in order to use the IDAC, the internal voltage reference must be turned on. The control for this is located on the third panel labeled “Power & Ref” Below the rows of input multiplexer selections, there is a system monitor function. This controls a set of switches in the multiplexer that can be used as a system monitor. The ADC input can be used to measure offset (by tying the inputs together at midsupply), gain (by measuring the inputs to the full-scale reference), and the temperature (by measuring a proportional to absolute temperature voltage as described in the datasheet). It can also be used as a coarse measurement for VREF1/4, VREF0/4, AVDD/4, DVDD/4. Next to the system monitor enable is a menu to control a current source to detect a burnout for a sensor. 2009 European FAE Summit, Munich

The second tab of the ADS1248 plug-in is used to control all the GPIO pins (labeled “GPIO”) The center switch selects the pin as an input or output.. If the pin is selected as an output, a button click lighting the button in the left column indicates the GPIO as an output high. Without the light, the GPIO is an output low. If the pin is selected as an input, the right columns shows the input value read on the pin. If the light is lit, the GPIO reads a high, if the light is off, the GPIO reads a low. Note that before a GPIO can be controlled here, it must be enabled first. To enable the GPIO for the input pins (AIN2-AIN7), this is done on the first panel controlling the input multiplexer. To enable the GPIO for the REF0P and REF0N pins, it is done on the next (third panel) controlling the reference. 2009 European FAE Summit, Munich

This is the third tab of the ADS1248 Plug-in, labeled “Power & Ref”. It controls the reference, the IDAC, and power supply configuration. At the top of the panel, A pull-down menu controls which reference is used to make ADC measurements. VREF0, VREF1, and the internal reference can be selected. Below the pull-down menu are two clickable boxes that can be used to set REF0P and REF0N as GPIO pins. The internal reference control selects whether the reference is Off, On, or follows the START signal. For more details on this, refer to the ADS1248 datasheet. The reference voltage value can be entered into the VREF space. This value will be used to calculate the voltage value in the Multiscope tester plug-in As mentioned in previous slides, the internal reference NEEDS TO BE ON if you want to use the IDAC function. The precision IDAC current values are generated based on the internal reference voltage. The Power Supply control selects between a single supply (+5V) and a bipolar supply (+/-2.5V) mode. The EVM draws its supply voltages from the MMB3 mother board and has onboard circuitry to select between these supplies. The IDAC magnitude control sets the amount of current each of the two IDACs provide. The IDAC0 and IDAC1 routing controls allow you to specify where the IDAC0 and IDAC1 currents go: To either the IEXC1 and IEXC2 pins, or to the analog inputs. You can also turn these currents off. 2009 European FAE Summit, Munich

The fourth tab is the Cal tab (referring to calibration). Pressing the Self Offset Calibrate button causes the ADS1248 to perform a self calibration; the offset and full-scale register values are then updated on the plug- in screen. You may also write to these registers by typing in hex values in the New fields, then press the Set button. You may read the values at any time by pressing the Read button. 2009 European FAE Summit, Munich

The fifth tab is labeled “About”. It gives details about the EVM and software. It can be used to show the Plug-in and Firmware versions. The ID indicator shows the value of the ID[3:0] bits in the ADS1248 ID register. The Notes text may show relevant notes about the plug-in or the firmware code, if there are any. 2009 European FAE Summit, Munich

On the right side of the ADCPro Program is the space for the test plug-ins. Two of them in particular can be used for taking data from the ADS1248EVM and performing some basic analysis. The first one is MultiHisto. The MultiHisto plug-in is used to take a block of data and organize the ADC readings into a histogram. Block size is the number of data points is the set. To start taking the data, press the acquire button at the top right corner of the ADCPro window. The DC Analysis box displays several parameters of the captured readings. The StDev column displays the standard deviation of the data set. This is equivalent to the RMS noise of the signal, assuming a dc input. The Codes (pp) column shows the peak-to-peak spread of the codes; for a dc data set, this range would be the peak-to-peak noise. The Mean column displays the mean of the data set; as pointed out above, this value might be one way to measure offset. The last two columns on the DC Analysis display relate to calculating ENOB or the Effective number of bits. The ENOB column displays the effective number of bits of the converter as calculated from the standard deviation or RMS noise. The Noise Free Bits column displays the effective bits of the converter when calculated using the peak-to-peak noise. If the inputs of the ADS1248 are tied together, the noise of the part can be measured. In the ADS1248 datasheet, the noise listed in various tables and graphs can be easily duplicated. 2009 European FAE Summit, Munich

A second useful test for ADCPro is the MultiScope plug-in. Like an oscilloscope, it plots data versus time. We’ll use this for taking data in the Live Lab. The y-axis gives the voltage measured by the ADC. In order for this to tell the user the correct voltage, the VREF value must be accurately measured and entered into the VREF space on the “Power & Ref” tab of the plug-in. 2009 European FAE Summit, Munich

What type of systems can be put together with the ADS1248?
So what types of systems can be put together with the ADS1248? The subsequent slides show several setups with the temperature elements shown earlier in this presentation. In this section, we’ll spend more time on RTD configurations than either thermocouple or thermistor configurations. Many different configurations of RTD can be used depending on the desired tolerance to line resistance. 2009 European FAE Summit, Munich

2-Wire RTD Measurement 2009 European FAE Summit, Munich
The 2-wire RTD connection is the simplest method using an RTD for a remote temperature measurement. It is a two- wire measurement because there is a single drive and single return for the RTD connection. The resistors circled in red are the parasitic line resistances. It represents the potential error in the measurement. In many cases the line resistance can be a large contributor for the error, because the sensor is far away from the sensing electronics. In many cases, the RTD is a low value resistor in the PT100, this is nominally 100ohms at 0C. If the sensor is far away from the ADC, the line resistance may be a significant error, especially because the sensitivity of the device is only ohms/C. Shown above, a current from one of the IDACs is sent through the input pins directly into the RTD. The return current goes through RBIAS, which biases the RTD above ground so that the measurement in in the common-mode input range of the ADC. The voltage across RBIAS also acts as a reference voltage. One of the ADS1247/8 current sources is connected to one of the terminals of the RTD lines by setting the appropriate bits in the IDAC0 register. The value of the current can be chosen by setting the ISELT bits in the IDAC0 register. The internal band-gap reference MUST be turned on by setting VREFCON bits in the MUX2 register. The internal reference has to be turned on for the IDAC to function even though the reference to the device is supplied externally. On the EVM we set can measure the RTD with AIN0 and AIN1. IDAC0 is used at AIN0, while IDAC1 is turned off. The reference resistor RBIAS can be read by REF0. A jumper from REFN0 to ground and a wire jumper from REFP0 to AIN1 will connect the reference. The voltage developed across the RTD is measured by also connecting the RTD through the PGA to the ADC. The voltage measured across the RTD is proportional to temperature (determined by the RTD’s characteristics). RBIAS value is selected according to IDAC current source setting. The reference to the device is also derived from the IDAC. The appropriate external reference has to be selected by setting the VREFSELT bits in the MUX2 register. RBIAS determines the reference voltage to the ADC as well as the input common mode of the PGA. The reference as well as the input to the device is a function of the IDAC current in this topology. This ratiometric approach guarantees more Effective Number Of Bits (ENOBs) as the noise in the IDAC reflects in the reference and as well as in the input and hence tends to cancel off. The effect of the IDAC current temperature drift also gets canceled off in this ratiometric topology. The major limitation of the two wire method is the voltage drop across the line resistances add up to the voltage drop across the RTD and hence the sensor cannot be very far away from the measurement setup. For best performance with the ratio-metric approach no filtering capacitance should be added to either the signal path or the reference path. The IDAC current mismatch drift does not matter since there is only one current path. 2009 European FAE Summit, Munich

2-Wire RTD Measurement Advantages: Disadvantages: Simple
Ratiometric – IDAC current drift is cancelled Noise in the IDAC is reflected in both the reference and the RTD Disadvantages: Least Accurate Line resistance affects the measurement The filter must be removed on the EVM. For the two-wire setup for an RTD measurement, we can list the advantages and disadvantages. For the advantages: It is simple. It uses two inputs and an IDAC, and with another resistor the setup is finished. With the eight inputs of the ADS1248, you could string seven RTD’s in this fashion and get measurements for each. The measurement is ratiometric. You don’t need an absolute measurement, and the IDAC current drift is not a factor because only the ratio between the RBIAS and the RTD resistance matters. As a consequence, the noise in the IDAC is reflected in both the reference and the RTD. Again only the ratio between RBIAS and the RTD matters. For the disadvantages: This is the least accurate way to setup the RTD. The line resistance affects the measurement. As mentioned earlier, the line resistance can be a large factor in the error. Minor changes can help reduce this error (and they will be discussed later). Since series resistance will hinder the measurement, the filter must be removed on the EVM. If you look at the schematic, there are two 47ohm resistors in the input signal path. They must be removed or they become part an IR drop in the measurement. If you are using AIN0 and AIN1 as the inputs, then R4 and R5 must be removed and replaced with a short (or a 0ohm resistor). 2009 European FAE Summit, Munich

2-Wire RTD Measurement Setup
2-Wire measurement sensitive to series resistance R4 and R5 removed on EVM Plug-in: PGA Gain = 1, Data Rate = 20 Block Size = 128 AINP = AIN0 < IDAC0 AINN = AIN1 Reference Select = VREF0 Internal Reference = On IDAC mag = 1000uA IDAC0 = AIN, IDAC1 = Off VREF = 1V ≈ (1000uA x 1k) Board: RTD: Black, Green: AIN0 RTD: White, Red: AIN1 Reference Resistor: AIN1 to GND, 1k Jumper: GND to REF- Wire: AIN1 to REF+ Here are some notes on the setup of the two wire RTD demo: On the EVM R4 and R5 are removed (or else the current from the IDAC reacts the the additional resistance from the input filtering Set the PGA Gain = 1 Set the Data Rate to 20 and the Block Size to 128 (you don't want to wait too long to get a result) AINP = AIN0, set IDAC0 to come to this input, AINN = AIN1 Select the reference as REF0 The internal reference should be on for IDAC configuration The IDAC magnitude should be 1000uA, IDAC0=AIN, IDAC1=off. RTD is nominally about 109 ohms at 25C, while the RBIAS is 1k ohms. The ADC should read about 109mV as a measurement depending on how accurate RBIAS is. During this demo, I had a 4 wire RTD used to look like a two wire RTD. The RTD goes across AIN0 and AIN You will need a jumper to short REF- to ground and a wire to tie REF+ to AIN1 2009 European FAE Summit, Munich

Example: RTD: PT100 IDAC = 1mA RBIAS = 1k Each line resistance = 0.5
2-Wire RTD Measurement A PT100 has about a 0.384 change for each 1oC of change Example: RTD: PT100 IDAC = 1mA RBIAS = 1k Each line resistance = 0.5 We get: Reference 1mA x 1k = 1V ADC Measurement: 1mA x (100 + 0.5+ 0.5) = 101mV Input is within ADC common- mode input range Here we show an example with values of a 2-wire RTD measurement. In this example, the RTD is a PT100. The IDAC used is 1mA going into the RTD and returning through an RBIAS of 1kohms. Here I’ve added a line resistance of 0.5 ohms each. In this case the reference will be 1mA x 1kohms = 1V The ADC will measure, 1mA x 101ohms = 101mV. Here the input is within the input common mode range of the ADC, because it’s set up by the IDAC current reacting with RBIAS. There is one important thing to take out of this measurement. Since all the current goes through the RTD, the line resistors, and RBIAS, and since the input current to both then inputs and reference are small, the RTD is a direct ratio of RBIAS. If the value of RBIAS is accurately known, then the ADC reading is a ratio of the the RTD resistance and the RBIAS value. In this example, the reading never even needs to be converted to a voltage. This is a rather simplistic version of what is shown as a demo, however it does show the necessary calculations for setting the system up. Remember that a platinum RTD at room temperature will be roughly 109 ohms. 2009 European FAE Summit, Munich

A 3-wire RTD application is shown above. In this 3-wire example two selectable current sources are used to provide symmetry and compensate for the line resistances in the RTD wiring. The ADS1247/8 current sources are connected to the two channels connecting to the two RTD terminals by setting the appropriate bits in the IDAC1 register. The value of the current can be chosen by setting the ISELT bits in the IDAC0 register. The internal band-gap reference has to be turned on by setting the VREFCON bits in the MUX2 register. The internal reference has to be turned on for the IDAC to function even though an external reference channel is used. The voltage measured across the RTD is proportional to temperature (determined by the RTD’s characteristics). RBIAS value is selected according to IDAC current source setting. RBIAS determines the reference voltage to the ADC as well as the input common mode of the PGA. The reference as well as the input to the device is a function of the IDAC current in this topology. The noise in the IDAC reflects in the reference and as well as in the input and hence tends to cancel off. This ratio-metric approach guarantees more ENOBs. The effect of the IDAC current temperature drift also gets cancelled off in this approach. Only the IDAC current mismatch drift matters. The limitation of the two wire method has been avoided in this topology and hence the sensor can be very far away from the measurement setup as long as noise coupling into the wire does not degrade the noise performance. This setup utilizes less than half of the dynamic range of the ADC as the input to the ADC is never negative. 2009 European FAE Summit, Munich

3-Wire RTD Measurement Advantages: Disadvantages: Simple
Input line resistances cancel Sensor can be farther away Ratiometric – IDAC current drift is cancelled Disadvantages: IDAC current and drift need to match For the three-wire setup for an RTD measurement, advantages are: It’s still simple, and compared with the two-wire RTD measurement, we’ve only added one current source. If the line resistances are equal and the current sources are equal, the effect of the line resistances cancel each other and only the RTD measurement remains. If the measurement can tolerate line resistances, then the sensor can be farther away from the ADC. It’s still a ratiometric measurement Disadvantages: The IDAC current and drift need to match. Luckily, the IDAC currents match very well (typically within 0.03% of full scale) 2009 European FAE Summit, Munich

3-Wire measurement far less sensitive to series resistance Measurement illustrated with 47 of series resistance Change reference resistor to 499 Plug-in: PGA Gain = 1, Data Rate = 20 Block Size = 128 AINP = AIN2 < IDAC0 AINN = AIN3 < IDAC1 Reference Select = VREF0 Internal Reference = On IDAC mag = 1000uA IDAC0 = AIN, IDAC VREF = 1V ≈ (1000uA x 1kW) Board: RTD: Black, Green: AIN2 RTD: White: AIN3 RTD: Red: AIN5 Reference Resistor: AIN5 to GND, 499 Jumper: GND to REF- Wire: AIN5 to REF+ Here are some notes on the setup of the three wire RTD demo: Set the PGA Gain = 1 Set the Data Rate to 20 and the Block Size to 128 AINP = AIN2, set IDAC0 to come to this input AINN = AIN3, set IDAC1 to come to this input Select the reference as REF0 The internal reference should be on for IDAC configuration The IDAC magnitude should be 1000uA, IDAC0=AIN, IDAC1=AIN. Again, the RTD is nominally about 109 ohms at 25C. However, the RBIAS is 499 ohms (note that we have twice the current going through RBIAS so we have half the resistance to maintain the reference value). The ADC should read about 109mV as a measurement. You will still need a jumper to short REF- to ground and a wire to tie REF+ to AIN1 Here, I am using AIN2 and AIN3 as the measurement input. If you remember, these inputs still have the 47ohm resistor in series. As long as the resistances match, the system will do a good job rejecting the effects of this series resistance. 2009 European FAE Summit, Munich

3-Wire RTD Measurement Example: We get: RTD: PT100 IDAC1 = IDAC2 = 1mA
RBIAS = 500 Each line resistance = 0.5 We get: Reference (1mA+1mA) x 500 = 1V ADC Measurement: 1mA x (100 + 0.5  1mA x 0.5 = 100mV Again we’ll use the PT100 RTD. Here, we use both IDACs at 1mA such that they go to either input. RBIAS is lowered to 500ohms, the line resistances remain at 0.5ohms. In this case, we still get 1V at the reference. However, since each IDAC pushes the same amount of current into the line resistances, the values cancel. This leaves the current from the first IDAC reacting with the RTD. The result in this example is 100mV, which is just IDAC1 reacting with the RTD alone. Again, this is a simplistic version of the measurement. The previous slide of the setup notes tells you more about the actual calculation. 2009 European FAE Summit, Munich

3-Wire RTD Measurement However:
A PT100 has about a 0.384 change for each 1oC of change 0.384 x 1mA = 384uV However: If the IDAC currents or line resistances do not match, there can be errors in cancellation. ADS1248 IDAC currents are matched to 0.03% typ. With 1mA IDACs, the mismatch is 0.3A In previous example, error is 0.3A x 0.5 = .15uV The error in line resistance mismatch can be higher in comparison! However, If the IDAC currents or line resistances do not match, there can be errors in the line resistance cancellation. Remember we’re looking at sensitivity of about 384uV/C of change. In the last example, the IDACs were at 1mA, so the typical mismatch is 0.3uA. 0.3uA on 0.5ohms is 0.15uV Or if the mismatch in the line resistance is 0.5ohms reacting with 1mA, this is 0.5uV The point is that in any given system, I think I’m more likely to have problems with 0.5ohms of line resistance mismatch than 0.03% of IDAC mismatch. 2009 European FAE Summit, Munich

3-Wire RTD Measurement with Hardware Compensation
This 3-wire topology allows complete utilization of the input dynamic range of the device by adding a compensating resistor RCOMP in the second arm of the RTD as shown in the figure. The voltage drop across RCOMP subtracts from the voltage drop across the RTD. RCOMP is chosen such that it is equal to the RTD resistance at the middle of the temperature measurement range. By adding RCOMP, we compensate the measurement so that the center of the temperature range is a 0V measurement. This also allows for using the PGA function to gain up the input voltage. 2009 European FAE Summit, Munich

Same Benefits and Problems as the typical 3-wire measurement Advantages: Centers the measurement so that the center temperature is at 0V Easier to use a larger PGA gain Disadvantages: IDAC current mismatch is gained up by RCOMP as well as the line resistance Since this is the same as the 3-wire RTD measurement, it has many of the same advantages and disadvantages. With hardware compensation, the temperature can be centered so the measurement is at 0V. This also means that we can use a larger PGA gain for the measurement. However, if there is an IDAC current mismatch. This is gained up by the RCOMP as well as the line resistance. 2009 European FAE Summit, Munich

3-Wire RTD Measurement with Hardware Compensation Setup
110 resistor added as hardware compensation Centers the measurement around 0V so that more gain can be used. Plug-in: PGA Gain = 128, Data Rate = 20 Block Size = 128 AINP = AIN2 < IDAC0 AINN = AIN4 < IDAC1 Reference Select = VREF0 Internal Reference = On IDAC mag = 1000uA IDAC0 = AIN, IDAC VREF = 1V ≈ (1000uA x 1kW) Board: RTD: Black, Green: AIN2 RTD: White: AIN3 RTD: Red: AIN5 100 resistor AIN3 to AIN4 Reference Resistor: AIN5 to GND, 499 Jumper: GND to REF- Wire: AIN5 to REF+ Here are some notes on the setup of the three wire RTD with hardware compensation demo: Set the PGA Gain = 1 Set the Data Rate to 20 and the Block Size to 128 AINP = AIN2, set IDAC0 to come to this input AINN = AIN4, set IDAC1 to come to this input An RCOMP of 110ohms goes from AIN3 to AIN4. Note that AIN3 is not selected and no current flows into that input. Select the reference as REF0 The internal reference should be on for IDAC configuration The IDAC magnitude should be 1000uA, IDAC0=AIN, IDAC1=AIN. The RTD is nominally about 109 ohms at 25C. However, the RBIAS is 499 ohms (note that we have twice the current going through RBIAS so we have half the resistance to maintain the reference value). The ADC should read about 109mV as a measurement. You will still need a jumper to short REF- to ground and a wire to tie REF+ to AIN1 Here, I am using AIN2 and AIN4 as the measurement input. Again, the series 47ohm resistors are rejected from the measurement if they match. The important part of this setup is that the RCOMP sets the reference measurement to 0V so that the ADS1248 can better use the PGA gain. In this example the 110ohm roughly matches the 109ohm resistance of the RTD at room temperature. 2009 European FAE Summit, Munich

We get: Reference (1mA+1mA) x 500 = 1V ADC Measurement (0oC): 1mA x (100 + 0.5)  1mA x (100 + 0.5) = 0mV ADC Measurement (100oC): 1mA x (138.4 + 0.5)  = 38.4mV Example: RTD: PT100 IDAC1 = IDAC2 = 1mA RBIAS = 500 Each line resistance = 0.5 RCOMP = 100 Again, we’ll use the PT100 RTD. Just as in the standard three wire RTD configuration, we use both IDACs at 1mA such that they go to either input. RBIAS is lowered to 500ohms, the line resistances remain at 0.5ohms. In this case, we still get 1V at the reference. With this measurement, the matching currents will react with the RTD and the line resistance in one path. In the other path, the current will react with RCOMP and the line resistance in the other path. If RCOMP is equal to the RTD, and the line resistances match, then all values are cancelled resulting in a 0V measurement. RCOMP is chosen for a reference temperature. The measurement will be based on the change from this reference temperature. Again, this is a simplistic version of the measurement. The previous slide of the setup notes tells you more about the actual calculation. 2009 European FAE Summit, Munich

The 4-wire RTD application provides the highest level of accuracy as it isolates the excitation path of the RTD from the sensing path. The ADS1247/8 current sources are connected to one of the terminals of the RTD lines by setting the appropriate bits in the IDAC1 register. Line4 of the RTD has been tied to the IOUT1 terminal of the device since line4 is not a sensing line, thus saving an input channel in the device for other sensors. The value of the current can be chosen by setting the ISELT bits in the IDAC0 register. The internal band- gap reference has to be turned on by setting VREFCON bits in the MUX2 register. The internal reference has to be turned on for the IDAC to function even though the reference to the device is supplied externally. The voltage developed across the RTD is measured by also connecting the RTD through the PGA to the ADC. The voltage measured across the RTD is proportional to temperature (determined by the RTD’s characteristics). RBIAS value is selected according to IDAC current source setting. The reference to the device is also derived from the IDAC. The appropriate external reference has to be selected by setting the VREFSELT bits in the MUX2 register. RBIAS determines the reference voltage to the ADC as well as the input common mode of the PGA. The reference as well as the input to the device is a function of the IDAC current in this topology. The noise in the IDAC reflects in the reference and as well as in the input and hence tends to cancel off. The IDAC current drift does not matter as we follow the ratio-metric topology. The IDAC current mismatch drift does not matter since there is only one current path. For best performance with the ratio-metric approach no filtering capacitance should be added to either the signal path or the reference path. This setup utilizes less than half of the dynamic range of the ADC as the input to the ADC is never negative. 2009 European FAE Summit, Munich

4-Wire RTD Measurement Advantages: Disadvantages:
Most accurate, line resistances are no longer a problem Sensor can be far away Ratiometric measurement No IDAC drift component Disadvantages: Need to use external IDAC pins Only two IDAC pins available The advantages to the 4-wire measurement are: It is the most accurate. Line resistances are not a problem because of the low input current into the ADS1248 Sensors can be very far away from the ADC itself This is a ratiometric measurement and there is no IDAC drift component to worry about There disadvantages are: You need to use the external IDAC pins There are only two IDAC pins available 2009 European FAE Summit, Munich

Return to G=1 1k reference resistor Most accurate measurement Plug-in: PGA Gain = 1, Data Rate = 20 Block Size = 128 AINP = AIN3, AINN = AIN4 Reference Select = VREF0 Internal Reference = On IDAC mag = 1000uA IDAC0 = AIN, IDAC1 = Off VREF = 1V ≈ (1000uA x 1kW) Board: RTD Black: AIN2 RTD Green: AIN3 RTD White: AIN4 RTD Red: AIN5 Reference Resistor: AIN5 to GND, 1k Jumper: GND to REF- Wire: AIN5 to REF+ Here are some notes on the setup of the four wire RTD demo: Set the PGA Gain = 1 Set the Data Rate to 20 and the Block Size to 128 RTD source wire goes to AIN2 with IDAC0 going to AIN2 AINP = AIN3 AINN = AIN4 RTD return wire goes to AIN5 as does the RBIAS Select the reference as REF0 The internal reference should be on for IDAC configuration The IDAC magnitude should be 1000uA, IDAC0=AIN, IDAC1=AIN. Again, the RTD is nominally about 109 ohms at 25C. As in the first measurements the RBIAS is 1kohm. The ADC should read about 109mV as a measurement. You will still need a jumper to short REF- to ground and a wire to tie REF+ to AIN5 Here, I am using AIN3 and AIN4. This measurement should have virtually no effects from the series resistance. Here the IDAC currents do not go through the series resistances and the parasitic resistances are not seen because the input currents are so small. 2009 European FAE Summit, Munich

4-Wire RTD Measurement Example: We get: RTD: PT100 IDAC1 = 1mA
RBIAS = 1k Each line resistance = 0.5 We get: Reference 1mA x 1k = 1V ADC Measurement: 1mA x 100 = 100mV Error is differential input current times the line resistance Here is an example of the 4-wire RTD measurement. The PT100 RTD is being driven by one of the IDACs with 1mA of currrent. Again the RBIAS 1k allows for a ratiometric measurement. As for the results, the 4-wire RTD measurement allows for a force/sense connection to the RTD so that the line resistance reacts to the very small input differential current to the ADS1248. 2009 European FAE Summit, Munich

Thermocouple Measurement with 3-Wire RTD as Cold Junction Compensation
The thermocouple unlike the RTD needs no excitation source and generates a potential difference, across its terminals, which is proportional to the temperature (TJ – TREF). The thermocouple consists of a two metal junction which produces a voltage difference proportional to TJ, the junction temperature which has to be measured. Since these two metals have to be connected to the copper line two more junctions are created. These two metal junctions have to be placed at the same temperature TREF. Placing them at the same temperature TREF creates a voltage proportional to TREF and opposing that produced by the thermocouple junction. If the temperature TREF is a known temperature then the temperature TJ can be calculated by adding TREF to it. But if TREF cannot be forced to a known temperature it can be measured with an RTD. A three wire RTD method for the junction temperature compensation is shown above. When measuring the temperature using the thermocouple, the output of the thermocouple has to be biased. The ADS1247/48 provides a bias voltage generator for this purpose. The bias voltage is equal to the mid-supply voltage i.e. AVSS + (AVDD-AVSS)/2 and has to be tied to one of the terminals of the thermocouple as shown above. The internal reference is used to measure the voltage from the thermocouple. A filter may be used to suppress noise in the thermocouple voltage due to noise coupling to the lines. When measuring the temperature of the junction using the RTD, the external reference has to be selected to obtain the best noise performance using the ratio-metric approach. The three wire RTD section explains this in detail. The three wire RTD measurement with hardware compensation may be used instead of the normal method. 2009 European FAE Summit, Munich

Advantages: Thermocouple needs no excitation source RTD used for cold junction compensation. Disadvantages: Complex Requires multiple resources of the ADS1248 Internal reference used in measuring thermocouple The advantages to this setup are: The thermocouple itself needs no excitation. However, the VBIAS connection to the input is required to establish the DC operating point. A second setup, in this case an RTD in a three-wire configuration is needed to determine the cold junction temperature. Disadvantages include: It’s complex and requires calculation of two temperature systems. It requires the use of many of the ADS1248 resources on the device. The measurement for the thermocouple, requires the internal reference because of the absolute value. After that the temperature is either calculated or compared on a lookup table. 2009 European FAE Summit, Munich

Thermocouple Measurement with 3-Wire RTD as Cold Junction Compensation Setup
Two measurements Thermocouple uses VBIAS, but no IDAC current. Three-wire RTD setup as before Board: Thermocouple: AIN0 to AIN1 RTD Black, Green: AIN2 RTD White: AIN3 RTD Red: AIN5 Reference Resistor: AIN5 to GND, 499 Jumper: GND to REF- Wire: AIN5 to REF+ Plug-in: Thermocouple PGA Gain = 1, Data Rate = 20 Block Size = 128 AINN = AIN0 < VBIAS, AINP = AIN1 Reference Select = Internal, VREF = 2.5V Three-wire RTD AINP = AIN2 < IDAC0, AINN = AIN2 < IDAC0 Reference Select = VREF0 Internal Reference = On IDAC mag = 1000uA, IDAC0, IDAC1 = AIN VREF = 1V ≈ (2000uA x 499) Here are some notes on the setup of the Thermocouple measurement demo: The three wire RTD setup is the same as before only using different inputs in the mux. The thermocouple is sent to AIN0 and AIN1 as the inputs. The polarity does matter and in this example I have AIN0 as the negative input and AIN1 as the positive input. The VBIAS is set on the AIN0 to give the voltage a DC common mode bias. This is important because the input PGA is similar to an INA input and requires the DC bias. For the thermocouple, the measurement is not ratiometric and requires a known voltage. For this example you can turn on the internal reference and have that selected as the ADC’s reference. 2009 European FAE Summit, Munich

Example: Thermocouple: K-type RTD: PT100 with 3-wire measurement We get: The thermocouple is DC biased with VBIAS Measured using internal reference. The cold junction uses an 3-wire RTD In this example, we start by measuring the temperature of the cold junction with the RTD. As shown previously, this is a 3-wire RTD measurement. Next we measure a k-type thermocouple using the internal reference. Based on this voltage, and the RTD cold junction temperature measurement, we can determine the temperature if the thermocouple measurement either by calculating from voltage coefficients of an equation, or from a lookup table. 2009 European FAE Summit, Munich

Thermistor with Shunt Resistor Measurement
In this example, we use an NTC Thermistor with a 10kohm resistor in parallel. The IDAC current is set to 100uA. The 10kOhm resistor is used to help linearize the temperature response. In this example, I have taken out the line resistances. The thermistor and resistor combination is 5kohms, which is often much larger than Thermistor has a nominal 10k response at 25oC 2009 European FAE Summit, Munich

Advantages: Inexpensive temperature element Disadvantages: Shunt resistor needed to linearize the response Requires reference voltage Less accuracy, temperature determined by comparison to graph or lookup table One of the main benefits of using a thermistor is the cost. However, the disadvantages to using a thermistor are: A shunt resistor is required to linearize the voltage response. A known voltage reference is required There is less accuracy and the temperature is determined by comparison to a graph or a lookup table. 2009 European FAE Summit, Munich

Without linearization With linearization Here I show what happens without the linearization. Without the 10kohm resistor, the usable range is limited by two things. First, the non-linear response of the thermistor is extreme sensitive as it gets colder. With a 100uA driving the device, the sensitivity is very low at hot. However, the sensitivity is very high at as temperatures get cooler. This leave a small usable range. Second, the IDAC is limited to AVDD-0.7V so as the thermistor voltage rises, the IDAC loses current regulation. With the linearizing resistor, we can extend the usable range by giving up a little of the sensitivity. 2009 European FAE Summit, Munich

Thermistor with Shunt Resistor Measurement Setup
Similar to 2-Wire measurement sensitive to series resistance Resistor in parallel with thermistor for linearization Thermistor nominal value 1kW with negative temperature coefficient (NTC) Plug-in: PGA Gain = 1, Data Rate = 20 Block Size = 128 AINP = AIN0 < IDAC0 AINN = AIN1 Reference Select = VREF0 Internal Reference = On IDAC mag = 1000uA IDAC0 = AIN, IDAC1 = Off VREF = 1V ≈ (1000uA x 1kW) Board: Thermistor||Resistor: AIN0 to AIN1 Reference Resistor: AIN1 to GND, 1k Jumper: GND to REF- Wire: AIN1 to REF+ Note: For the demo, I could only find a 1kW NTC thermistor. The parallel resistor is 1kW as is RBIAS. Here are some notes on the setup of the Thermistor demo: This setup is very similar to the two wire RTD setup Again, on the EVM R4 and R5 are removed Set the PGA Gain = 1 Set the Data Rate to 20 and the Block Size to 128 (you don't want to wait too long to get a result) AINP = AIN0, set IDAC0 to come to this input, AINN = AIN1 Select the reference as REF0 The internal reference should be on for IDAC configuration The IDAC magnitude should be 1000uA, IDAC0=AIN, IDAC1=off. The parallel combination of the RTD and the resistor is nominally about 500 ohms at 25C, while the RBIAS is 1k ohms. The ADC should read about 500mV as a measurement. During this demo, I had a 4 wire RTD used to look like a two wire RTD. The RTD goes across AIN0 and AIN You will need a jumper to short REF- to ground and a wire to tie REF+ to AIN1 2009 European FAE Summit, Munich

Improved linearity with shunt resistance Non-linearity is under 3% when Rshunt equal to the thermistor at the circuits median temperature Heavy shunting reduces output NTC Thermistor has a nominal 10k response at 25oC In this example, we use a 10k resistor in parallel with the thermistor. The thermistor is an NTC with a nominal response of 10kohms at 25C. (Again note from the previous slide, the demo was done with a 1k thermistor and all other devices scaled). Much improved linearity with shunt resistance added (limited temp range) Non-linearity is under 3% for example when R-shunt equal to the thermistor at the circuits median temperature Heavy shunting reduces output Since the thermistor response alone is very non-linear, the 10kohm resistor acts as a ballast to linearize the response at its nominal value. This may give smaller temperature range for a more accurate measurement. The graph in this slide shows the response. At 25C, the parallel combination of the thermistor and resistor becomes 5kohms. With a 100uA IDAC as the source, this will yield a measurement of 0.5V. Note that RBIAS is needed to raise the measurement into the input common mode range. A minimum of 1kohms is necessary. 2009 European FAE Summit, Munich

Evaluation with the ADS1248EVM is easy with ADCPro
Conclusions We’ve covered three temperature elements: The RTD, thermocouple, and the thermistor Evaluation with the ADS1248EVM is easy with ADCPro There are many ways to connect the ADS1248 up to get a temperature measurement Conclusions So what have we come to today? We’ve covered three temperature elements: The RTD, thermocouple, and the thermistor Evaluation with the ADS1248EVM is easy with ADCPro There are many ways to connect the ADS1248 up to get a temperature measurement 2009 European FAE Summit, Munich

Questions? Comments? 2009 European FAE Summit, Munich

References ADS1248 Datasheet
ADS1148/ADS1248EVM and ADS1148/ADS1248EVM- PDK User's Guide Agilent Application Note 290 — Practical Temperature Measurements, pub. no EN "Sensors and the Analog Interface", Tom Kuehl, Tech Day Presentation “Developing a Precise PT100 RTD Simulator for SPICE", Thomas Kuehl, Analog ZONE.com, May 2007 "Example Applications For Temperature Measurement Using the ADS1247 & ADS1248 DS ADC", Application Note, (to be published) " Wire RDT (PT100 to PT1000) Temperature Measurement", Olaf Escher, Presentation 2009 European FAE Summit, Munich

Precision Temperature Measurement with the ADS1248

Similar presentations

Presentation on theme: "Precision Temperature Measurement with the ADS1248"— Presentation transcript:

Similar presentations

About project

Feedback

Log in

Auth with social network:

Precision Temperature Measurement with the ADS1248

Similar presentations

Presentation on theme: "Precision Temperature Measurement with the ADS1248"— Presentation transcript:

Similar presentations

About project

Feedback