Design and Fabrication of a Microheater Control System. Mike Chambers

Size: px

Start display at page:

Download "Design and Fabrication of a Microheater Control System. Mike Chambers"

Erik Logan
5 years ago
Views:

1 Design and Fabrication of a Microheater Control System Mike Chambers Senior Project Mentor: Florian Solzbacher, PhD Senior Project Advisor: Ken Stevens, PhD Correspondence to: mike.chambers@utah.edu Project Homepage: University of Utah

2 Abstract Microsensors are becoming increasingly important to society as the field of nanotechnology advances. One such microsensor is a solid state gas concentration sensor, which has been used for over a decade, especially in the automotive field to control air/fuel ratios in the combustion process. My senior project is to develop the closed loop microheater controller needed to regulate the high temperature setpoint(s). Feedback for the closed loop will be provided by an on-board thin film Resistance Temperature Detector (RTD) located near the microheater. This device will be capable of maintaining a constant set temperature for gas sensing, as well controlling step and impulse temperature profiles needed for testing, tuning and diagnostic purposes, as well as to explore novel methods of gas concentration detection. Because combustion engines operate in thousands of RPMs, the device will be able to respond very quickly, on the order of one millisecond. This device will be small and very precise (within a few degrees at C). The device will interface with a computer to assign the microheater profiles and setpoints. Resistance processing of the microheater itself will be implemented to check consistency of thin film RTD, as well as to provide information about the current gas concentrations. Introduction Solid state gas concentration sensors have been used extensively in the automotive industry for over 10 years. These sensors provide feedback to the fuel injector controls to constantly maintain the most efficient ratios of air to fuel, thereby increasing the most efficient combustion. Simply put, this type of sensor utilizes semiconductor properties of surface adsorption to detect changes in resistance as a function of varying concentrations of different gases. In order to detect these resistive changes and equate them to changing (and static) gas concentrations, the gas temperature must be held constant. A microheater or an array of microheaters near the gas sensor is controlled to maintain this temperature and to account for the convective cooling caused by gas flow. Fig. 1 below shows an example of a microheater. This picture is one of three microheaters on an array of three, which reportedly exhibits excellent temperature homogeneity [1]. Fig. 1: This figure displays an example of a microheater. The green rings are the heaters, the central array is the temperature sensor, and the central speckled bars are the gas sensors. The gas sensors are approximately 50 µm wide, or half the width of a human hair.

3 A disadvantage of many current gas sensors is that they are rated at relatively low temperatures (up to about 300ºC). Thus, they are located downstream from the combustion chambers far enough to allow sufficient cooling of the exhaust gases. However, high operation temperatures are needed for optimum performance and accuracy. Mike Sorenson, Srinivasan Kannan, and Xiaoxin Chen, under the supervision of Dr. Florian Solzbacher, are currently working on a solid state gas concentration sensor that will be capable of operating at temperatures of over 600ºC. This will allow the sensor to be placed directly outside the combustion manifold to provide superior accuracy. The design is currently in the last stages of fabrication and consists of interdigitated metal-oxide fingers serving as the gas sensor, with the heater encapsulating these fingers and a platinum RTD temperature sensor located near the heater. Platinum RTDs are one of the most stable, linear, and reproducible temperature sensors [2]. Rapid, accurate control of the microheater can be accomplished using a proportional-integralderivative (PID) controller. A basic block diagram of this controller is displayed in Fig. 2 below [3]. Fig. 2: A PID controller sums the 3 terms derived from the error signal e(t) to provide very accurate and stable control of a process. PID controllers are well documented. Please refer to [3] for a very detailed description of the PID controller. Basically, for temperature control, the temperature is compared with the desired setpoint temperature. The difference or error signal e(t) is applied to the controller, which uses e(t) to produce the control signal u(t). u(t) manipulates a physical input to the process, thereby causing a change in the regulated temperature that will reduce the error. The PID controller sums the 3 terms derived from e(t), namely: the proportional term K p applies a corrective term proportional to the error; the integral term K i seeks to hold its average input at zero; and the derivative term K d improves stability, reduces overshot caused by high K i and K p terms and improves response time by anticipating changes in error. To control the temperature of the microheater, the system must accurately measure the current temperature of the microheater and adjust the input power accordingly. This can be done strictly using discrete analog components. However, in order to obtain response times on the order of 1 ms, as well as to perform signal

4 processing on the resistance of the RTD and metal-oxide layer, it was decided to implement the PID control algorithm on a microcontroller. A basic block diagram for this project showing the present control setup is shown in Fig. 3 below. RTD- High precision current forced through resistor to obtain voltage drop related to temp. PC- RS232 serial comm. link with microcontroller. Setpoint temperatures and profiles assigned. 16-Bit ADC- High speed digital voltages (heater temp) provide feedback to microcontroller. 32-Bit Microcontroller- Processes the resistance signals from the RTD and metal-oxide heater and controls the power delivered to the heater. 16-Bit DAC- High speed analog current (scaled power) output to control heater. Microheater- Heater maintained at desired temperatures. Resistance signal fed back into controller Fig. 3: Block diagram of project. The voltage drop across the RTD corresponds to the temperature of the microheater. This voltage is converted to a 16-bit word and fed to the microcontroller to perform the PID algorithm based on the difference between the desired setpoint and the actual temperature. The microcontroller communicates with a computer to relay the desired temperature profile, as well as to relay resistance data. Based upon the setpoint temperature and the error signal, the microcontroller exports a digital 16-bit current to control the microheater. This signal is converted to an analog signal and fed to the microheater. The resistance of the heater is digitized and sent to the microcontroller for processing. This diagram displays the basic structure of the project. Using the microcontroller, the heater temperature can be controlled very rapidly and accurately using feedback provided by the RTD and the heater itself. The microcontroller will process the resistance signals, compute the PID control algorithm accordingly, and control the microheater. The high speed analog-to-digital converter (ADC) is needed to convert the voltage drop across the RTD into a digital word representing the temperature of the RTD. This will enable processing in the microcontroller itself. The high speed digital-to-analog converter (DAC) is needed to ensure long life of the microheater, as a digital signal could wear it out faster. Using an on board UART serial interface on the microprocessor, a computer will be interfaced for bidirectional communication. The computer will relay the desired temperature setpoints and the microcontroller will export

5 resistance data. If time permits, the device will also utilize on-board memory to log heater resistance and temperature. It was decided that using high speed data converters and a slower microcontroller would result in a more cost effective and simple design. Deliverables and Stretch Goals The end device will be a small box with a RS232 connector interfacing a PC, with wire leads for the RTD resistor and the heater. A computer will manage the desired setpoints of the microheater, and the microcontroller will deliver power to the microheater based on the feedback it receives from the RTD temperature sensor. The resistance signal of the heater will be fed back to the microcontroller as well, currently to add precision to the temperature calculation but future improvements could be made. If time permits, special processing of heater resistance signal will be implemented. This will allow novel approaches to gas concentration detection. To display this functional device, an LCD interface will be added to display the current heater temperature. An infrared (IR) camera will be used to visibly show the temperature of the heater. Oscilloscope plots will show the response times achieved. Should there be problems with the microheater currently being fabricated, a lower temperature microheater will be purchased to display the working controller. Project Tasks Stay on track o Attend weekly meetings with group/mentor, as well as advisor o ET: ~40 hours Familiarize myself with project, goals, etc. o Research micro-heaters, PID control, DSP, and microcontrollers Sources: web scholarly papers other publications o ET: ~25 hours Plan control setup o Work with Mike Sorenson, PhD student who is on team designing the microheater o Specs: Input: signal from PC which tells the device the desired temp. This could include a temperature transient or pulse Output: signal to PC relaying resistance of heater itself as well as the resistance of the RTD Interface with PC using RS232 o On-board microcontroller UART serial converted to RS232 Response time of 1 ms (room temp to 600ºC in 1 ms) Scalable analog control of the micro-heater On/off (digital) control won t work, decreases life of micro-heaters

6 If time permits, data-logging of resistance of metal-oxide layer Sensitivity and selectivity very critical in this application Heater is ~150 ohms Max of 1 watt power is to be delivered to heater On-board thin film resistor (RTD) will provide feedback in the PID loop, ~150 ohms o Further research on micro-controllers, data converters, and DSP~20 hours o Create block diagram~2 hours o Contact manufacturers to get free academic samples of development tools, evaluation boards, and electronics~10 hours o Order parts which are not available for academic samples~4 hours o Develop working (alterable) schematic~5 hours Construct controller (evaluation board will be used until final design perfected) o Solder components to evaluation board~6 hours o Program the microcontroller~ hours C coding of PID loop Must be integer math to use slower clock speed C coding of microcontroller loops o Read a lot in this step! Get help from Mike Sorenson and CE friend Test controller o Hook up controller to oscilloscope to test I/O o Hook up controller to a micro-heater in test chamber o Examine with IR camera o Design tests for the controller~20 hours Ensure stability Ensure accuracy Interfaces Schedule ADC w/ microcontroller o 16-bit word relaying voltage drop across RTD o 16-bit word relying resistance of microheater DAC w/ microcontroller o 16-bit word relaying current to power heater (P=IV) Microcontroller w/ on-board FLASH o Store digital words, as well as logging variables Computer w/ microcontroller o RS232 transceiver chip Convert UART serial interface on microcontroller to RS232 Milestones Block Diagram: finished by Oct. 10 Basic Schematic: finished by Oct. 17 Vendors contacted to negotiate price: Oct. 24 Parts ordered: Oct. 31 Main assemble completed (assuming parts arrive in time): Nov. 14 Program/coding: Rest of November/December

7 Testing: start of next semester o On evaluation board setup Packaging o PCB layout Testing Completed microheater control device: March 21 Risk Assessment Block Diagram: Green Working Schematic: Green Parts ordered/decided on: Green Device assembled: Green Programming: o PID control: Yellow Need compact integer math code to ensure I meet the response time o Microcontroller program: Yellow Not very experienced in this, YET. o HW interfacing (control with SW): Red I will need the most help on this I am not experienced in microcontrollers, ADCs, and DACs It was recommended that I go with a slower processor speed and high speed data converters to avoid board complexity. The tradeoff is code complexity. On-board flash should not be as difficult to work with BOM (unless otherwise noted, all prices for quantities greater than 1K) I will try to get as many parts as I can for free. 16/32-bit 60 MHz Microcontroller w/ >256 kb on-board flash o Texas Instraments TMS470 series Up to 60 MHz clock <$10 o Philips ARM 2100 series Up to 60 MHz On-board flash, RAM On-board UART serial interface LPC2194 o 512 KB Flash, 32 K RAM o 47 I/O pins o 8 timers o <$10 16-bit high speed ADC o Analog Devices AD MSPS 4 channels

8 $10.45 o Texas Instraments 16-bit high speed DAC o Analog Devices AD GSPS 1 channel $28.95 o Texas Instraments Precision current source o Drive current through RTD to get sample voltage Build this Could use Wheatstone Bridge Or another method Transistors, resistors, etc Precision Op-amp o Amplify voltage to deliver to heater Precision Resistors o To set gain on op-amp o To convert current out from DAC to a voltage RS232 transceiver chip o Convert serial to RS232 MAX232 Free samples available PC w/ LabView running (I am not doing the LabView part) Development Tools o Keil Software Free lite version o Evaluation boards Get academic version for Philips ARM microcontrollers References Florian Solzbacher, PhD Ken Stevens, PhD Mike Sorenson, PhD candidate [1] M. Graf, D. Barrettino, P. Kaeser, J. Cerda, A. Heirlemann, and H. Baltes. Smart Single- Chip CMOS Microhotplate Array for Metal-Oxide-Based Gas Sensors, Transducers (2003), 123. [2] Application Note. [3] E. Neary. Mixed-Signal Control Circuits Use Microcontroller for Flexibility in Implementing PID Algorithms. Analog Dialogue 38-01, January (2004)

Design and Fabrication of a Microheater Control System

Design and Fabrication of a Microheater Control System University of Utah Senior Project Michael William Chambers Electrical and Computer Engineering University of Utah Salt Lake City, USA www.ece.utah.edu/~mchamber