US 4,687,934 AGrant
Infrared Gas Analyzer With Automatic Zero Adjustment
Issue Date:1987-08-18
•22 Claims
•18 Drawing Sheets
Abstract
Automatic zeroing apparatus zeroes an infrared gas analyzer automatically upon the occurrence of preselected conditions to indicate zero in the absence of absorption of infrared radiation by a gas mixture being analyzed. The gas analyzer has a sample cell for containing a gas mixture to be analyzed. Infrared radiation directed through the sample cell is detected at a preselected wavelength to produce a detection signal. A signal processor outputs a signal systematically related to the difference between the detection signal and a reference signal. For zeroing the sample cell is filled with gas substantially nonabsorbent of infrared radiation at the respective characteristic wavelength. A comparator produces an error signal when the output signal differs from zero. A gain control automatically controls the signal level of the detection signal to reduce the output signal substantially to zero with the nonabsorbent gas filling the sample cell. The preselected conditions may include the passage of a predetermined time and a temperature drift beyond a predetermined limit.
Metadata
Assignee
- Andros Analyzers Incorporated
Inventors
- Robert E. Passaro
- Raymond E. Rogers
- Craig J. Griffith
Application Information
Application Number:US 8183422
Filing Date:1986-01-10
Priority Date:1986-01-10
Art Unit:256
Classifications
IPC:
G01J 100G01D 1800
Field of Search:
250252.1;338 R;339;341;343;344;345
Patent Drawings (18 sheets)
Description
Background of the Invention
This invention relates to infrared gas analyzers and more particularly to such gas analyzers with an automatic zero adjustment. Still more particularly it relates to such gas analyzers having multiple channels using a programmable microprocessor for performing the zero adjustment.
Infrared gas analyzers typically utilize an infrared source to direct infrared energy through a mixture of unknown gases contained in a sample cell. The infrared energy passing through the mixture in the sample cell at one or respective frequencies is measured as indicative of the relative concentrations of respective gases in the mixture, the respective gases exhibiting different absorption characteristics at respective frequencies. That is, the absorption by respective gases is frequency dependent.
One such analyzer is described in Passaro, et al., U.S. Pat. No. 4,346,296. As there described, an infrared source emits infrared at relatively constant intensity over a relatively broad spectrum. The infrared radiation from the source is interrupted periodically by a chopper. After passing through the sample cell, the chopped infrared is detected by respective detectors. In each case the radiation is filtered by a narrow band-pass filter so that such detector is effectively sensitive only to radiation of a particular narrow band of frequencies corresponding to a respective absorption frequency characteristic of a respective gas. The respective detection signals are thus systematically related to the relative concentration of the respective gases. Because of the chopper, these signals are AC signals at the chopper frequency. The signals are amplified, detected (rectified) and filtered to produce corresponding DC signals. Each filtered signal is applied to one input of a so-called span amplifier, which comprises a summing amplifier with an offset reference input and a controllable feedback input. The offset offsets the zero point, and the feedback controls span or the magnitude of input signal required for full scale output. The signal input has a gain control to control the point at which the signal input balances out the offset signal and hence the zero point of the span amplifier. That is the point where there is zero output from the span amplifier in the absence of absorption of the incident infrared radiation. The output of the span amplifier is proportional to the difference between the detection signal (after the gain control) and the offset signal. It may be noted that difference includes the addition of signals of opposite polarities. The amplifier output is measured by a meter, which may be a recorder, which is calibrated to indicate the relative concentration of the respective gas.
As described in the aforesaid U.S. Pat. No. 4,346,296, the zero point of the span amplifier is adjusted, that is, the instrument is zeroed, by introducing a so-called zero gas into the sample cell and adjusting the gain control of the signal input to provide a zero meter reading. The zero gas is a gas, such as nitrogen, which is substantially nonabsorptive of infrared radiation, at least at the frequencies passed by the respective filters. When a predetermined calibrating gas is introduced into the sample cell, the gain of the feedback signal is adjusted to some predetermined calibrated value. Then when the gas to be analyzed is introduced, the output meter properly records or indicates the relative concentrations of the respective constituent gas. For the purposes of checking the span calibration a push-button switch was operated to put a fixed input into the summing span amplifier and the resultant meter reading set at a predetermined level by adjusting the span control in the feedback loop.
Summary of the Invention
A problem with the gas analyzer of the aforesaid patent was that the zero point drifted from time to time, requiring intervention by an operator to return the system to a proper zero condition. Drift is caused by changes in the infrared source, the sample cell, other parts of the radiation path and the detection system. Such changes may be occasioned by temperature drift. This prior system required that the operator note the occurrence of changes or follow a schedule and perform manual zeroing on a regular basis or as needed. In any event, human presence and attention was required.
In accordance with the present invention zeroing of the system is performed automatically. Under the control of a programmable microprocessor each of any number of channels, three, for example, are zeroed periodically and/or upon temperature drift. More particularly, the system operates as a satellite to a host. The microprocessor provides for warm-up delays and then enters a zeroing mode repeated after each predetermined period or upon temperature drift above a predetermined level or upon intervention by the host.
In the zeroing mode the microprocessor provides signals calling for zero gas to fill the sample cell and then controls the gain of the signal input to a summing amplifier to provide a zero output from the span amplifier as determined by a comparator. Such gain control is effected by controlling the multiplying factor of a multiplying digital to analog converter (MDAC or DAC). The factor is determined independently for each channel and is stored in the microprocessor.
Upon completion of zeroing, the system enters its normal measuring regime. The gas to be analyzed is passed into the sample cell. Separate detection systems are provided for each channel, with appropriate multiplexing so that a single DAC can be used for all channels. For each channel the respective factor is provided by the DAC at the appropriate times during measuring. The output signals from the respective amplifiers are then provided to the host, and there applied to respective calibrated meters and/or recorders which thereupon indicate respective relative concentration of the constituents of the gas being analyzed.
Provision is made for span adjustment and electronic calibration. Provision is also made for disabling the zeroing operation under certain conditions, as upon electronic calibration or during emissions testing.
Thus, an important aspect of the present invention is the automatic zeroing of an infrared gas analyzer.
Still another aspect is the automatic zeroing of an infrared gas analyzer with hysteresis.
Another aspect is providing such automatic zeroing under the control of a programmable microprocessor.
Another aspect is the control of a single multiplying digital to analog converter for the automatic zeroing of multiple channels of an infrared gas analyzer.
Another aspect is the automatic zeroing of such gas analyzer system periodically and/or upon a predetermined ambient temperature charge.
Other aspects, objects and advantages of the the present invention will become clear from the following detailed description, particularly when taken in connection with the accompanying drawings.
Brief Description of the Drawings
FIG. 1 is a simplified illustration, partly diagrammatic, of an infrared gas analyzer system including automatic zeroing in accordance with the present invention;
FIG. 2 is an isometric drawing of the source, sample cell and detector arrangement of the gas analyzer system shown in FIG. 1, with a portion broken away;
FIG. 3 is a detailed electrical schematic diagram of preferred forms of preamplifier, rectifier and filter shown in FIG. 1;
FIGS. 4A and 4B together are a detailed electrical schematic diagram of a preferred form of multiplexing, demultiplexing, DAC, comparator and microprocessor control shown in FIG. 1;
FIG. 5 is a detailed electrical schematic diagam of a preferred form of the span amplifiers shown in FIG. 1;
FIG. 6 is a pictorial representation of the transition between operational modes of the gas analyzer system illustrated in FIG. 1;
FIG. 7 is a state transition diagram illustrating the signal levels for the ready, pump & solenoid and zero request lines for the operational modes of the gas analyzer system illustrated in FIG. 1;
FIGS. 8A-8C are a detailed flow chart of the software program controlling the processor of the gas analyzer system illustrated in FIG. 4;
FIG. 9 is a detailed flow chart of the subroutine SYS FAULT called by the system program illustrated in FIGS. 8A-8C;
FIGS. 10A, 10B are a detailed flow chart of the subroutine WARM UP called by the system program illustrated in FIGS. 8A-8C;
FIGS. 11A, 11B are a detailed flow chart of the subroutine SAR called by the system program illustrated in FIGS. 8A-8C;
FIG. 12 is a detailed flow chart of the subroutine SAR TEMP called by the system program illustrated in FIGS. 8A-8C;
FIG. 13 is a detailed flow chart of the subroutine OFFSET CO called by the subroutine WARM UP illustrated in FIGS. 10A, 10B;
FIG. 14 is a detailed flow chart of the subroutine CHECK TEMP called by the system program illustrated in FIGS. 8A-8C; and
FIGS. 15A, 15B are a detailed flow chart of the program INTERRUPT which is executed in response to a timer interrupt in conjunction with the system program illustrated in FIGS. 8A-8C.
Detailed Description
Generally, as shown in FIG. 1, the infrared gas analyzer system of the present invention includes a sample cell 11 for containing a gas mixture to be analyzed. A source 13 directs infrared radiation through the sample cell 11, and the infrared radiation is periodically interrupted by a chopper 14 at a predetermined frequency. A detection signal is produced by each of detectors 15-1, 15-2, and 15-3 corresponding to the infrared radiation received at a respective preselected wavelength, such signal being an AC signal having an amplitude monotonically related to the infrared radiation passing through the sample cell at that wavelength and having a frequency corresponding to the predetermined frequency of the chopper 14. The preselected wavelength corresponds to the characteristic absorption wavelength of the respective gas it is desired to detect in the sample mixture. Such selection is effected by respective narrow band pass filters. The AC signal outputs of the detectors are processed to produce respective DC signals having amplitudes systematically related to the concentration of the respective gas to be detected. Each AC signal is amplified by a respective amplifier 17-1, 17-2, 17-3, full-wave rectified by a respective full-wave rectifier 18-1, 18-2, 18-3 and filtered by a respective filter 19-1, 19-2, 19-3 to produce the corresponding DC signals. The DC signals are applied through a demultiplexer 20 to a gain control in the form of a multiplying DAC 21. The DAC 21 adjusts the gains of the respective channels, and the adjusted signals are applied through a multiplexer 22 to respective span amplifiers 23-1, 23-2, 23-3. The outputs of the span amplifiers are transmitted to a host 92 where they are applied to respective meters and/or recorders calibrated to read relative concentrations of the respective gases.
In FIG. 2 is illustrated a sample cell 11 showing the physical disposition of the source 13, chopper 14, and detectors 15-1, 15-2 and 15-3 and filters relative thereto. This part of the system may be substantially the same as the corresponding part of the system shown in the aforesaid patent, except that there are three channels rather than two.
The preferred embodiment of the present invention is for detecting the relative presence of three gases in the exhaust gas of an automobile engine. The three gases may be hydrocarbons, carbon monoxide and carbon dioxide, but the invention is not limited to use in connection with such specific gases, nor is it limited to use in connection with the exhaust gas of an automobile engine. There will be other uses, apparent to those skilled in the art, for the analyzer of the invention where a low cost unit of the type described will be acceptable.
The sample cell unit is mounted on a base plate 25 having a generally rectangular shape and having rigidizing walls 27 at its periphery connected by rigidizing cross webs 29. At one end of the unit, the infrared source 13 is shown mounted in a source housing 31 secured to the base plate by mounting brackets 33. The source 13 includes a plate 35 having a reflecting recess 37 of parabolic, conical or other suitable shape to reflect the infrared energy into a beam. The source element 39, which may be of any suitable type, is mounted on a holder 41. Suitable electrical connection is made to energize the source element by means not shown. The source housing 31 includes a window 43 for transmitting infrared radiation from the source element out of the housing. The window may be of any material substantially transparent to infrared radiation at the wavelengths of interest.
The sample cell 11 is mounted on a pair of roughly U-shaped mounting clips 47, 49 attached to the base plate and is in alignment with the infrared energy eminating from the infrared source 13. Each of the clips 47, 49 comprises a pair of upright tines 51 which are resilient and which have projections 53 extending inwardly therefrom. The resilient tines enable the sample cell 11 to be easily removed from the device for replacement or cleaning. Each end of the sample cell 11 is provided with an end sleeve 55, 57, respectively, each of which retains in place a window 59, 61, respectively, transparent to infrared radiation. A suitable material for the windows is, for example, mica. A coupling 65, 67, respectively, is mounted in each end sleeve 55, 57, for permitting exhaust gas from the automobile engine exhaust being analyzed to pass into and out of the sample cell 11.
Each of the sleeves 55, 57 is provided with a semi-annular collar 73 which wraps approximately 180.degree. around the underside of each sleeve and extends axially about one half the length of the sleeve. The collars 73, with the sample cell 11 in the operative position, engage the underside of the projections 53, to retain the sample cell in place. However, since the sample cell 11 will, during use, become dirty from the exhaust gas being analyzed, it is desirable that it be easily cleaned or even replaced. By merely separating the tines 51 so the projections 53 clear the collars 73, the sample cell 11 may be easily removed or installed.
The end of the sample cell assembly adjacent the source housing 31 is spaced a short distance therefrom so as to enable the interposition of the chopper wheel 14. The chopper wheel 14 is rotatably mounted on the spindle 77 of a drive motor 79 which, in turn, is mounted to the base plate 25 by a motor mount bracket 81. The chopper wheel 14 comprises a circular opaque disk which may be any suitable construction to interrupt periodically the infrared energy emanating from the source. However, it is preferred that the disk have a cut-out region 87 of generally arcuate shape extending through 180.degree. of the wheel rotation. In this way, the beam is blocked 50% of the time during each rotation. The chopper thus provides an AC signal, which has the effect of cancelling out any background DC radiation. Of course, other types of chopper wheels may be used, such as ones having two 90.degree. cut-out regions, to provide the same duty cycle.
The end of the sample cell 11 opposite the source 13 is abutted flush against a filter and detector block 89. The block 89 is secured by suitable means to the base plate. The block 89 is generally cylindrical in shape and is axially aligned with the sample cell 11. The block 89 contains the filters 16-1, 16-2, 16-3 disposed side by side in alignment with radiation passing through the sample cell 11. The filters pass infrared energy of the characteristic absorption wavelengths of the gases of interest, for example, hydrocarbons, carbon monoxide and carbon dioxide.
Immediately behind the filters are mounted the detectors 15-1, 15-2, 15-3, which may be of any suitable construction. The detectors provide an output which is systematically related (e.g., proportional) to the magnitude of the infrared energy passing through the corresponding filter. Thus, the output of each of the detectors comprises an AC signal, the amplitude of which is systematically related to the energy at the respective frequency passing through the sample cell 11.
Referring now more particularly to FIG. 1, a form of the signal processor used in the apparatus of the invention is shown. The output of each detector 15-1, 15-2, 15-3, which is a substantially square wave signal having the frequency of the rotation of the chopper wheel 14, varies between nearly zero and a negative (or in certain embodiments positive) voltage. This signal is applied to the respective preamplifier 17-1, 17-2, 17-3 by which it is suitable amplified.
In the illustrated embodiment, each preamplifier may be in the form illustrated in FIG. 3. FIG. 3 illustrates the preamplifier 17-1; however, the preamplifiers 17-2 and 17-3 may be identical thereto. The AC signal from the respective detector 15-1 is applied at a terminal J1-8 through a coupling capacitor C2, and the amplified output is output through coupling capacitors C4 and C3.
The amplified signals are applied to respective full-wave rectifiers (demodulators) 18-1, 18-2, 18-3 which act to demodulate the signals. In the illustrated embodiment, each rectifier may be in the form illustrated in FIG. 3. FIG. 3 illustrates the rectifier 18-1; however, the rectifiers 18-2 and 18-3 may be identical thereto. The amplified signal from the preamplifier 17-1 is applied through the coupling capacitors C4 and C3, and the rectified signal is output through a resistor R10.
The rectified signals are applied to respective filters 19-1, 19-2, 19-3 which act to filter the rectified signals to corresponding DC signals. In the illustrated embodiment, each filter may be in the form illustrated in FIG. 3. FIG. 3 illustrates the filter 19-1; however, the filters 19-2 and 19-3 may be identical thereto. The rectified signal from the rectifieer 18-1 is applied through the resistor R10 and the filtered DC signal is output on a terminal P1-15.
As shown with particularity in FIG. 4, the outputs from the filters 19-1, 19-2 and 19-3 are applied at respective terminals P1-15, P1-21 and P1-27 to the respective X.sub.0, X.sub.1 and X.sub.2 input terminals of the demultiplexer 20, which may be, as shown, of the type CD4052BE, connected as shown. The input signals are demultiplexed under the control of control signals applied at terminals A and B of the demultiplexer 20 from a microprocessor 90. The multiplexer 20 connects a respective input terminal to the output terminal X as synchronized by the control signals. Thus, output signals at the terminal X are successive different input signals in a particular order. Each successive output signal is applied through an inverting amplifier U1-1,2,3 which acts to match impedances. The output of the amplifier is applied to the voltage reference terminal (V.sub.REF) of the DAC 21, which may be of the type MP7623.
The DAC 21 operates to multiply the voltage on the voltage reference terminal by a factor determined by the digital inputs to input terminals BIT 0 to BIT 11 from the microprocessor 90. How these inputs are determined will be explained below. Suffice to state here that the product appears as current outputs at terminals I out 1 and I out 2. These outputs are applied to a differential current amplifier U1-5,6,7, the output of which is a voltage proportional to the product of the signal from the multiplexer 20 and the factor from the microprocessor 90. This output is fed back to the DAC 21 at its terminal R.sub.FEEDBACK and is also applied to the X terminal of the multiplexer 22, which is also preferably of type CD4052BE. The multiplexer 22 is also controlled by control signals applied to its terminals A and B from the microprocessor 90 in synchronism with the demultiplexer 20, with an additional signal applied to its inhibit terminal INH to inhibit the connection to the output terminal for brief periods during switching to prevent false signals. The multiplexer 22 provides signals for the respective channels at terminals X.sub.0, X.sub.1 and X.sub.2. These are applied through respective sample and hold circuits U2-1,2,3, U2-5,6,7 and U2-8,9,10 to respective span amplifiers 23-1, 23-2 and 23-3 at terminals P1-17, P1-24 and P1-33, respectively. The outputs of the respective span amplifiers are transmitted over respective lines to the host 92 from terminals P1-18, P1-26 and P1-31, respectively (FIG. 5). In the example, the three channels are 0-hydrocarbons (HC), 1-carbon monoxide (CO) and 2-carbon dioxide (CO.sub.2). At the host the signals are applied to meters, which may include recorders, which are calibrated in a conventional manner to indicate the relative concentration of the respective gases in the sample being analyzed.
The preferred embodiments of the span amplifiers 23-1, 23-2 and 23-3 are illustrated in greater detail in FIG. 5. The signal inputs from the multiplexer 22 (after the sample and hold circuits) are applied at respective input terminals P1-17, P1-24 and P1-33 through respective input resistors R31, R63 and R96 to the summing junctions at the (-) inputs (U2-6, U4-6 and U6-6) of respective summing amplifiers U2-5,6,7, U4-5,6,7 and U6-5,6,7. A reference voltage is developed by a zener diode 94 from a voltage applied thereto from a -12 v source through a dropping resistor 96. This reference voltage is applied to the respective summing junctions through respective input resistors R33, R65 and R98. The span amplifier outputs appear at respective terminals P1-18, P1-26 and P1-31. In addition to being transmitted to the host 92, these signals are fed back to the respective summing juntions through respective external span adjust resistors 98-1, 98-2 and 98-3, which may be located in or at the host 92. The span adjust resistors are in series with respective variable resistors R14, R46 and R79 which are factory set for appropriate span adjustment range. The respective (+) terminals U2-5, U4-5 and U6-5 are connected to ground through respective resistors R34, R66 and R99.
Referring particularly to the span amplifier 23-1, the summing amplifier U2-5,6,7 operates to keep its summing junction U2-6 at ground potential. The current feedback through the resistors 98-1 and R14 balances the sum of the positive current from the signal input through the resistor R31 and the respective negative current to the voltage reference through the resistor R33. When channel 0 is properly zeroed, in a manner to be discussed below, the output signal at the terminal P1-18 is zero in the presence of zero gas, as indicated at the respective meter and/or recorder at the host 92. Then the presence of hydrocarbons in the sample cell 11 reduces the signal input at the terminal P1-17 accordingly, and the resulting unbalance gives rise to a proportionately related output signal at the terminal P1-18. The magnitude of this output is dependent upon the magnitude of the span adjust resistor 98-1. This span adjust resistor 98-1 is set by introducing a predetermined gas (hydrocarbons in this example) of known concentration into the sample cell 11 and adjusting the resistor 98-1 to the point where the meter reading at the host 92 is a value previously determined to indicate such predetermined concentration of the calibration gas. The meter and/or recorder is calibrated in a known manner so that all readings of the meter and/or recorder properly indicate the relative concentration of the particular gas. The other span amplifiers 23-2 and 23-3 operate in the same fashion; however, a small portion of the signal output from the amplifier 23-2 is picked off from a voltage divider consisting of resistors R67 and R51 and applied through an input resistor R84 to the summing junction U6-6 of the span amplifier 23-3 to compensate for the effect of carbon monoxide on the carbon dioxide measurement.
As an alternative to calibrating the system using a known gas of known concentration, the system can be calibrated electronically. Taking channel 0 as an example, with channel 0 properly zeroed and zero gas present in the sample cell 11, an additional known calibration signal current is applied to the summing junction U2-6 and the span adjust resistor 98-1 is adjusted accordingly. This additional calibration signal is applied through a resistor R32, which is normally grounded at a terminal P1-20 through a normally closed switch 100-1, so that no current is input through the resistor R32. When the switch is opened, the resistor R32 is connected through a variable resistor 102-1 and a resistor R18 to the voltage reference at the zener diode 94. This applies the calibration signal to the summing junction. The switch 100-1 and the variable resistor 102-1 may be in or at the host 92. The variable resistor 102-1 may be adjusted empirically to correspond to a predetermined hydrocarbous concentration in the sample cell 11. To this end, with the calibration gas filling the sample cell 11, the span adjust resistor 98-1 is adjusted for proper calibration. Then with zero gas in the sample cell 11 and the span adjust resistor in the position so determined, the switch 100-1 is opened and the variable resistor 102-1 set to give the same meter reading as during normal calibration with the calibration gas. Thereafter, whenever the span adjust resistor 98-1 gets out of the calibration position, it may be reset by electronic calibration by filling the sample cell 11 with zero gas and opening the switch 100-1 to apply the additional known calibration signal current and adjusting the span adjust resistor accordingly. It is unnecessary to have the calibration gas at hand. Channels 1 and 2 may be similarly calibrated using switches 100-2, 100-3 and variable resistors 102-2, 102-3.
As stated above, the measurements depend upon a proper zeroing of each channel. This requires the adjustment of each respective factor provided by the DAC 21 to produce zero output when the zero gas is in the sample cell 11. Such adjustment is effected by the microprocessor 90 in response to the outputs of the respective span amplifiers 23-1, 23-2, 23-3 at respective output terminals P1-18, P1-26 and P1-31 when zero gas is in the sample cell 11. The microprocessor is put in its zeroing routine, as will be described below. The outputs from terminals P1-18, P1-26 and P1-31 are applied through a demultiplexer 104 to a comparator 106. The comparator 106 compares the respective output signals with 0 or ground and outputs a binary 0 or 1 to indicate whether or not the signal is too large. The comparator output is input to the port PA3 of the microprocessor 90, which is programmed, as described below, to output on ports PC0-PB7 a binary signal that, when applied to the input ports BIT 0-BIT 11 of the DAC 21, produces zero output of the respective span amplifier 23-1, 23-2, 23-3. After each channel is thus zeroed, the system is returned to its normal measuring routine, as will be described below.
In the illustrated embodiment, the microprocessor 90 is of the type MC68705P3. Control signals therefor are applied to its port PA4 through a demultiplexer 108 which is timed by signals from the microprocessor in synchronism with the multiplexer 20. The timing signals for the multiplexer 22 and demultiplexers 20, 104 and 108 are derived from an internal clock within the microprocessor 90 (or an external clock, if desired), which may be at 4 MHz, and output from ports PA0-PA2.
There are four inputs to the demultiplexer 108. Terminal Y.sub.0 receives a signal indicating that the system is performing an electronic calibration. As described above, when an electronic calibration is performed, one, or more usually all, of switches 100-1, 100-2, 100-3 are opened, lowering the voltages on respective terminals P1-9, P1-19 and P1-25, which are normally grounded when the switches are closed. The lowered voltage on one or more of these terminals is applied through a wired-OR gate 110 to a comparator 112 where it is combined with a reference voltage REF2 and compared to ground to develop an appropriate digital signal when the system is being electronically calibrated. Terminal Y.sub.1 receives a signal indicating that an emissions test is in progress. When an emissions test is in progress, there will likely be a large carbon dioxide content in the gas in the sample cell 11. This results in a large signal in the carbon dioxide channel, channel 2 in this example. This signal, at the terminal P1-31 is applied to a comparator 114 where it is compared with a reference voltage REF1 to produce an appropriate digital signal when the comparator 114 indicates high carbon dioxide. Terminal Y.sub.2 receives zero command signals from the host 92 and/or directly from the microprocessor 90. The microprocessor signal is developed on its port PA6 (as will be described below) and inverted by an inverter U7-3,4. This is a "ZERO REQUEST" signal that is applied directly to the terminal Y.sub.2 when a switch 116 is closed. The terminal Y.sub.3 is not used in the microprocessor program described below. It is provided in order to accommodate different numbers of channels. A three channel system is described below. A signal on the terminal Y.sub.3 would be used if an optional single channel operation were programmed.
One of the causes of zero drift is variation of ambient temperature, which may cause changes in the source 13, the detectors 15-1, 15-2, 15-3 or the transmission of radiation between them. As will be explained in greater detail below, provision is made for entering the zeroing regime whenever there has been a drift in ambient temperature of more than 6.degree. C. since the last zeroing. This requires a temperature measurement. In the illustrated embodiment, temperature is sensed by a thermistor 118, which is connected in a sort of bridge circuit including a series resistor 120, the DAC 21, a resistor 122, and a resistor 124. The thermistor 118 is connected to +15 v and the resistor 124 is connected to -15 v. The output of the bridge is taken from the junction of the resistors 122 and 124 and is compared to o or ground. The bridge is balanced when the output is o. This occurs when the DAC effectively complements the resistance of the thermistor 118. That is, if temperature changes change the resistance of the thermistor 118, the bridge is balanced when the DAC changes in complementary fashion. The state of the DAC at balance is thus a measure of temperature. The temperature measurement is effected through channel 3 of the demultiplexers 20 and 104 and the multiplexer 22. When channel 3 is energized by the control signals from the microprocessor 90, the microprocessor 90 operates as described above (and below) to place the DAC 21 in the condition that drives to 0 the voltage at the junction of the resistors 122 and 124, balancing the bridge. The microprocessor 90 includes a register that holds the digital signals from the microprocessor 90 to the DAC 21 at balance during any zeroing. The microprocessor compares subsequent digital signals at balance to the stored signals and flags a difference signifying a change of more than 6.degree. C.
As will be explained further below in connection with the operation of the microprocessor 90, there are three microprocessor output signals in the preferred embodiment, READY on output port PA5, as inverted by an inverter U7-1, 2, ZERO REQUEST on output port PA6 as inverted by the inverter U7-3, 4 and PUMP & SOLENOID on output port PA7 as inverted by an inverter U7-5, 6. These signals may be indicated by respective LED circuits, indicated as optional in FIG. 4A. These signals are transmitted to the host 92. The host 92 responds to these signals by producing appropriate control signals for filling the sample cell 11. The sample cell is filled by a pump 126 through the inlet coupling 67. The filling gas is determined by a valve 128, which may be a solenoid operated spool valve, which selects between two input gases, the zero gas and the sample gas. In the course of calibration, a known gas of predetermined concentration is furnished as the sample gas; whereas, during measurement the gas being analyzed is furnished. The host operates the solenoid actuator of the valve 128 and the motor driving the pump 126 at such times as to effect the filling of the sample cell 11 with the appropriate gas.
The infrared gas analyzer shown in FIG. 1 will now be more fully described with respect to the software which controls the automatic zeroing of the channels. The software is stored in a ROM internal to the microprocessor 90 and controls the operations of the system and communications with the host 92. The system operation is shown in flow chart form in FIGS. 8-15 and a state-transition diagram of the operation of the control program is shown in FIGS. 6 and 7.
By reference to the state transition diagram of FIG. 6, it is seen that after a power-up mode 400, the first operation of the system is a warm-up mode 402. In the warm-up mode 402, the system waits through a warm-up interval. Afterwards, the CO channel (channel 1) is zeroed and measured against an offset after a drift delay. If the CO channel zero has drifted more than the offset, the system is not entirely warmed up and should be allowed to settle, and thus, the cycle is repeated until the drift is less than the offset. After the warm-up cycle is exited by a successful passage of the drift test, the mode will be changed to those modes comprising the basic operation of the system. The mode that begins basic operation is the zero mode 406 where all channels including the CO, HC, CO.sub.2, and temperature channels are balanced or zeroed against references with a zero gas contained in the measuring or sample cell 11.
Once the zero mode 406 has been completed successfully, the system will make a transition into the normal mode 408. In the normal mode, measurements or analyses of the unknown gas mixture can be made by the host with the system maintaining its calibration. After warm-up mode 402 and zero mode 406, this is the main operational mode for the system, and most of the system time will be spent in the normal mode 408. However, for a number of predetermined conditions, the system is automatically zeroed through the zero mode 406 and passes through three other modes including a zero request mode 410, a wait mode 414, and a purge mode 412 to get there.
The zero request mode 410 is entered from either the zero mode 406 or the normal mode 408. The condition which switches the system from a zero mode 406 to a zero request mode 410 is an indication that one or more of the channels have failed to come to a zero condition. The zero request mode 410 is the first mode in a sequence which automatically attempts to rezero the channel which caused the request (actually all channels). When the system senses the zero fail condition, the zero request mode 410 is enabled and a zero request is generated on the zero request line either to the host 92 or back to the microprocessor 90 through a jumper (switch 116) to the zero command line (Y.sub.2 of the demultiplexer 108). The conditions which switch the system from a normal mode 408 to a zero request mode 410 are the expiration of a predetermined time limit or the change of the ambient temperature greater than a predetermined limit. In the present embodiment, the time limit is 15 minutes and the temperature limit is .+-.6.degree. C. different from the ambient temperature at which a particular channel was last zeroed. In response to the zero request mode, a zero request is made of the host 92, which returns a zero command, or the zero request is applied through the jumper 116 to the microprocessor 90, directly. Further, if the system is in the normal mode 408, the host can issue a zero command on its own.
In response to a zero command, either in answer to the zero request or from the host directly, the system enters a wait mode 414 where two conditions are tested. First, the system tests whether the CO.sub.2 channel reading is greater than a predetermined limit. If this limit is exceeded, it is an indication that the system is performing an emissions test, and it would be inopportune to perform the zero operation. Therefore, the system will loop at the wait mode 414 until the CO.sub.2 channel reading becomes less than the limit. Further, during an electronic calibration, the system should not be zeroed, and the wait mode 414 will be executed until an indication that a calibration is complete has been given.
If the signal in the CO.sub.2 channel is not greater than the limit and the electronic calibration is not engaged, the system will go directly to a purge mode 412. The purge mode 412 is a timed mode where the pump and solenoid are turned on to fill the sample cell 11, insuring that the zero gas enters the cell and all remaining constituent gases of the last analysis are exhausted. After the time has expired for the purge mode 412, the system will cycle back to the zero mode 406 and continue the loop from there. Upon the zero mode 406 finding any dead channel for CO, HC, or CO.sub.2, the system will enter a system fault mode 404 which causes an entire system shutdown. The system fault mode 404 is also entered upon the finding of a dead channel in the warm-up mode 402. The system fault mode 404 causes the microprocessor 90 to enter an infinite loop which can only be exited by a reset or power-down of the system.
The most common operation of the system is a cycling between the normal mode 408 and the zero mode 406 every 15 minutes because of a time limit. This is the auto-zero function of the system which uses the zero request mode 410, wait mode 414, and purge mode 412 to assist in the cycle.
FIG. 7 illustrates the states of the three communication and status lines, ready, zero request, and pump & solenoid for each of the system modes. The zero request line is only true during the zero request mode 410 when requesting the host 92 to generate a zero command. The ready status line is only true during the normal mode 408 when the host can perform calibrated measurements. During the zero request mode 410 and the wait mode 414, an indicator is set to blink to alert a subsequent ready state. The pump and solenoid line is true during the modes of warm-up 402, zero 406, and purge 412 to request the host to fill the cell 11 with a zero gas.
These functional modes are implemented by a program stored in the microprocessor 90 for which a system flow chart is shown in FIGS. 8A-8C. The program begins the operation of the system by initializing particular constants in the microprocessor 90 in block A10. Particularly, the port data directions and initial port patterns for port A, port B, and port C are initialized with the predetermined values. Next in the initialization step, a number of pointers are stored to provide address references to particular points in the program. The initialization then clears particular variables used in the operation of the program, and starts the hardware timer on the microprocessor 90. After initialization, the program enables the interrupts in block A12 to allow an interrupt routine to act as a foreground routine while the following program executes in background. The hardware timer will cause a real time interrupt and a transfer of control from this main routine to the interrupt routine every 10 milliseconds.
In the next block A14, a warm-up routine is executed which will be more fully described hereinafter by reference to FIGS. 10A and 10B. Basically, the warm-up routine delays a specified period of time and then zeros the CO channel and determines whether the channel is drifting. As soon as the system is stable, the program continues to block A16 where the start of the zero mode 406 is located. The RAM mapped images of the output lines PA5, PA7, and PA5 of the microprocessor 90 corresponding to the zero request bit, pump and solenoid bit, and ready bit, respectively, have their states set according to the mode. Because the system is in the zero mode, the zero request bit should be cleared, the pump and solenoid bit should be set, and the ready bit should be cleared. These operations are performed in blocks A16-A20. The setting of the pump and solenoid bit will request that the host 92 turn on the pump 126 and switch the solenoid valve 128 to the zero gas side such that the auto-zeroing of the system can be accomplished. Next in block A22, the CO channel is zeroed by calling a subroutine SAR which will be more fully explained hereinafter. In block A24 the CO.sub.2 channel is zeroed in the same manner by again calling subroutine SAR. Likewise in block A26, the HC channel is zeroed by calling the subroutine SAR.
The subroutine SAR basically determines the MDAC gain value which will cause the comparison of the channel output and the reference voltage to be zero. If the subroutine SAR cannot zero a particular channel, it will set a flag which is termed the zero fail flag. After all three channels have been zeroed, in block A28 the system will measure the temperature with a subroutine SAR TEMP which determines the ambient temperature at the time of the auto-zero operation. Next in block A30, the zero fail flag is tested to determine whether or not any of the channels have not been properly zeroed. If the zero fail flag is not set, then the program advances to the normal mode 408 which starts at block A32. Otherwise, the program will branch to block A62 (FIG. 8B), which is the start of the zero request mode 410.
In the normal mode 408, the program clears the temperature fail flag in block A32 before resetting the three output bits for output lines PA5, PA6, and PA7. The zero request bit is cleared in block A34, the pump & solenoid bit is cleared in block A36, and the ready bit is set in block A38. These operations alert the host with the ready line that a measurement may now be made and that the auto-zeroing has been concluded. The ready line indicates that all channels have been successfully zeroed and that the system is prepared to accurately measure the constituent gases within the cell 11. The pump and solenoid bit is cleared to move the solenoid valve 128 to permit the unknown gas to enter the cell 11 and to allow the host 92 to control the pump 126. Next in block A40, the hardware timer for the normal mode 408 is reset. This is accomplished by clearing the hardware timer location, the seconds location, and the minutes location. These locations are used in a timing loop illustrated by block A42 and blocks A50-A60. The timing loop counts the time interval between the auto-zero operations in 10 millisecond increments.
In block A42 a test is performed to determine if enough 10 millisecond increments have been accumulated to equal one second. If not, the program drops through to a number of tests in blocks A44, A46, and A48. In the first test, block A44, the program determines whether or not the host has generated a zero command to the microprocessor 90 via input Y.sub.2 of the demultiplexer 108 and pin PA4 of the microprocessor. If the test in block A44 is affirmative, then the program will branch to block A74 (FIG. 8C) where the wait mode 414 is entered. However, if there is no zero command, then the program will measure the temperature change at that particular point in time in block A46 by calling the subroutine CHECK TEMP. The temperature change is measured in this routine by comparing the gain corresponding to the current temperature with the stored gain corresponding to the temperature at the last auto-zero operation .+-. an offset. The comparator input on pin PA3 will then indicate a change in temperature. If the change in temperature is greater than .+-.6.degree. C., then the program will set the temperature fail flag. This is an indication that the temperature has changed more than the range allowed within the last 15 minutes, the time limit interval.
Therefore, in block A48, if the temperature fail flag has been set, the program will branch to block A62 which is the start of the zero request mode 410. However, if the test in block A48 is negative indicating that there was no zero command and the temperature is still within range, then the program loops back to block A42 to continue executing the timer loop. At the end of every second, the test in block A42 will be affirmatively passed and a loop consisting of blocks A50, A52, and A54 executed. In block A50 the timer location is cleared and in block A52 the location labelled seconds is incremented. Thereafter, the program tests in block A54 whether one minute has expired by comparing seconds with a constant 60. Thus, this loop counts in increments of one second to one minute and then clears the location labelled seconds in block A56 and increments the location labelled minutes in block A58. The test in block A60 is executed to determine if the number of minutes counted by the loop is equal to the time interval, which in our example is 15 minutes. If not, the loop repeats by executing the blocks A44, A46, and A48 before incrementing the timer counts. After the time interval has expired, the program will affirmatively branch from block A60 to block A62, which is the beginning of the zero request mode 410.
Thus, there are three ways to exit from the normal mode loop, namely a zero command, a temperature fail flag, and the time out of the auto-zero interval. On either the temperature drift or the time limit conditions, the program will sequence from the normal mode 408 to block A62 which begins a zero request mode.
In block A62, the zero request bit is set while in block A64 the pump & solenoid bit is cleared, and in block A66 the ready bit is set. Setting the zero request bit requests the host to reply with a zero command for auto-zeroing. In this manner the system signals the host 92 that it is ready to do auto-zeroing. With the jumper 116 in place, the zero request can directly generate the zero command such that this path will automatically put the microprocessor 90 into the zero mode 406. In blocks A68 and A70, the blink flag is set true, and the blink timer location is cleared. This will allow the interrupt routine to determine whether the ready line should be toggled at half second intervals, with such operation taking place in that routine. After these operations have been concluded, the program loops at block A72 until a zero command is recognized at pin PA4 of the microprocessor 90.
When the zero command is sensed, it transfers the program to block A74 and the beginning of the wait mode 414. In this mode, the three reference lines are set to the particular states of the mode such that the zero request bit is cleared in block A74, the pump & solenoid bit is cleared in block A76, and the ready bit is set in block A78. In block A80 and block A82 the blink flag is set to true, and the blink timer is cleared. These are the same states as for the zero request mode except that the zero request line has been cleared to indicate that the request has been answered.
Next in blocks A84 and A86, the conditions for the wait mode 414 are tested. In block A84 the multiplexer 108 input Yl is tested to determine whether the output of the comparator 114 is at a high logic level. This is done by sampling pin PA4 of the microprocessor 90 during the switching of the multiplexer 108 to that channel. The input to the comparator 114 is from the output of the CO.sub.2 channel and is compared against a reference REF1 developed at the node of two resistors R23 and R24. If CO.sub.2 is at a higher concentration than a certain percentage, in the preferred embodiment 1.5%, then the output of the comparator 114 is at a high logic level, and the test is affirmatively passed.
A CO.sub.2 concentration of this amount or higher indicates that the system is performing an emissions test and that the auto-zero should wait until the analysis is over. This wait occurs even though there has been a zero command either from the host or in reply to the zero request from the microprocessor 90. When the voltage indicating the CO.sub.2 concentration drops below the reference REF1, the loop is exited to block A86 where the condition of an electronic calibration is tested for.
This operation is tested by sampling pin PA4 of the microprocessor 90 when the multiplexer 108 is selecting input Y.sub.0. The input to the multiplexer channel Y.sub.0 at that time is the output of comparator 112 which includes inputs through diodes 110 which are connected to the electronic calibration resistors 102-1, 102-2, and 102-3 (FIG. 5). When any one of these calibration resistors is not grounded during electronic calibration, a negative signal is applied through the wired-OR diode array 110 to the inverting input of the comparator-amplifier 112. This causes the comparator-amplifier to produce a high level output which is transmitted to pin PA4 of the microprocessor 90 and which is read during the interrupt routine for the front panel. As long as the calibration operation continues on any channel, the program will loop in block A86 such that the auto-zero operation will not occur.
After both of the conditions that maintain the system in the wait mode 414 are cleared, the negative paths of blocks A84 and A86 transfer the system into the purge mode 412 beginning at block A88. The purge mode is a timing loop to insure that the pump 126 and solenoid value 28 operate at least a predetermined amount of time, called the purge time, before the auto-zero operation is started. The operation purges the cell 11 of residual analysis gas and insures a sufficient quantity of the zero gas has been pumped into the cell 11. In blocks A88-A92 the control bits for the zero request, the pump and solenoid, and the ready lines are set according to the state table. The zero request bit is cleared in block A88, the pump and solenoid bit is set in block A90, and the ready bit is cleared in block A92. To indicate that the system is not in a ready state, the blink flag is set to false in block A94.
Thereafter, the purge timer loop, blocks A96-A104, is entered. In block A96 the timer is cleared and the seconds location cleared. The program then loops at block A98 until one second has elapsed on the timer. The program then increments the seconds location in block A100, clears the timer, and compares the seconds location with the purge time in block A104. If the purge time has not elapsed, the negative branch from block A104 begins the loop once again at block A98. Once the purge time, 30 seconds, has elapsed, the affirmative branch of block A104 transfers the program back to the zero mode 404 at A16.
The warm-up mode 402 will now be more particularly described with reference to FIGS. 10A and 10B. The warm-up mode allows a set time after the power is turned on where the cell warms up such that drift and transient readings are not taken in the auto-zeroing. In blocks A310, A312, and A314 the system status bits for the request line, pump and solenoid line, and ready line are determined. These three blocks set the pump and solenoid bit, and clear the zero request bit and ready bit. Thus, the pump and solenoid bit is set to produce a condition where zero gas is pumped into the cell and the ready signal to the host is cleared indicating a busy condition. A warm-up delay loop is then entered where for 15 minutes the system idles waiting for the timer to complete the interval. In block A316 the location seconds, minutes, and timer are cleared and the propagation wait subroutine executed in block A318. Thereafter, the timer is compared against one second in block A320. Once one second has elapsed, the program continues by incrementing the location seconds in block A322 and by clearing the location timer in block A324. Next, the location seconds is tested against one minute (60 seconds) in block A326. If one minute has not elapsed, then the program loops back to block A320 to continue to count seconds. Otherwise, in block A328 the location seconds is cleared and in block A330 the location minutes is incremented. Thereafter, each minute the program will advance through the test in block A326 and go through the test in block A332 to determine if the warm-up interval (15 minutes) has elapsed. If not, the beginning of the loop is branched to by a transfer of control to block A320.
Once the warm-up interval has elapsed, the program advances to location WM1 where the routine SAR is called to zero the CO channel in block A334. If the CO channel cannot be zeroed or fails the zero offset test, then either the zero fail flag will be set or the system will enter the system fault routine. However, if a valid zeroing of the CO channel is accomplished, then the program will progress to the next block A336 where the subroutine OFFSET CO is executed. This subroutine takes the MDAC gain determined in block A334 and subtracts an offset from it to determine an offset gain used for the drift test. If the CO channel is stable, then the original CO channel gain minus the offset will not cause the comparator 106 to make a transition. This offset test gain is saved in block A338 and the locations in which the MDAC gain zeroed is stored is cleared in block A340 during a drift interval.
A timing loop consisting of blocks A342-A360 is then entered to delay the program for the drift interval time. The timing loop initiates by clearing the timer and clearing the location seconds in blocks A342 and A344, respectively. Seconds are tested at block A346, and each one second interval causes a clearing of the timer and an incrementing of the location seconds in block A348 and A350, respectively. Minutes are tested for by comparing the seconds location to 60 in block A352 and then clearing the seconds location and incrementing the minutes location in blocks A354 and A356, respectively, upon the passage of each minute. The minutes location is then compared to a drift time constant in block A360 to finally exit the loop when the drift time has elapsed.
In block A360 the offset gain is output to the MDAC and the propagation wait routine called in block A362. In block A364 the comparator 106 is tested to determine whether the output has changed more than a particular amount during the drift time interval. If the output of the comparator 106 as tested for in block A364 is high, this condition indicates that the system has drifted. An affirmative response to the test causes the program to cycle back to the address WM1 where the CO channel is again zeroed starting at block A334. The loop continues until the channel is stable enough such that it does not drift more than the offset over the drift time interval. At that point the warm-up mode 402 is exited by the negative branch from block A366 to the address zero (FIG. 8A) to begin the zero mode 406.
The subroutine OFFSET CO is illustrated in FIG. 13. The purpose of the OFFSET CO routine is to subtract an offset from the MDAC gain used to zero the CO channel to determine if the system is still drifting during the warm-up mode 402. In block A176 the low byte of the CO offset is read and compared in block A178 against the least significant nibble of the MDAC gain. Depending upon which is larger, the nibble is subtracted from the offset in block A180, or the offset is subtracted from the nibble in block A182. If the nibble is subtracted from the offset, then a borrow from the most significant byte of the MDAC gain must be accomplished in block A184 and the most significant byte decremented. The result of the subtraction in block A180 is then subtracted from a full least significant byte increment (OF) in block A186, the difference being equivalent to the subtraction of the offset from the nibble after borrowing. Next the high byte of the CO offset is read in block A188 and then subtracted from the most significant byte of the MDAC value in block A190. The routine then returns to the warm-up mode program from which it was called.
The subroutine SAR, more fully detailed in FIGS. 11A and 11B, is the routine used to zero a channel by successively approximating the digital gain for the MDAC 21 which will cause a zero condition at the comparator 106. The program begins in block A106 by clearing the zero fail flag which will be later used as an indication that the channel failed the zeroing operation if set. Next in block A108, the gain of the MDAC 21 is set to 100% by outputting a 12-bit word of digital ones (FFF) on port pins PC0-PC3, PB0-PB7 of the microprocessor 90. The program in block A110 enters a propagation wait subroutine so that the change in gain has time to propagate through the multiplexer 22, sample and hold circuits, and span amplifiers 23-1, 23-2, 23-3 to the comparator 106. In block A112 the program determines after the wait if the comparator is low. When the MDAC 21 is programmed with the highest gain value (FFF) this should bring the output of the comparator 106 low. If not, the channel cannot be zeroed and the absence of low comparator signal will be taken as a dead channel with the program branching to a system fault subroutine SYS FAULT at block A114.
The system fault subroutine SYS FAULT performs certain system shut down functions and then enters an infinite loop in the system fault mode 404. Once a dead channel is detected in this manner, the only methods of exiting the system fault mode infinite loop are either to reset the microprocessor 90 or power-down and power-up the system. The system fault subroutine SYS FAULT is illustrated as a detailed flow chart in FIG. 9. The system fault subroutine SYS FAULT is a infinite loop which is entered when the channel output of any of the channels cannot be made to produce a low output state on the comparator 106 even with the highest gain. This indicates that the channel is dead, i.e., there is some broken connection or inactive sensor in the input circuitry such that the system cannot measure the input from that particular channel. In block A304 the program clears port A essentially zeroing pins PA5, PA6, and PA7 to indicate the system fault mode 404. In block A306 the blink flag is set to false to disable any other setting of the ready line, and finally, in block A308 the interrupts are disabled. The microprocessor 90 then enters an infinite loop where a branch is made back to block A304 and continues in this loop until the microprocessor is powered down or externally reset by a switch.
Returning to the subroutine SAR (FIG. 11A), if the comparator 106 is determined in block A112 to be low, the system enters a toggle operation where the most significant bit (bit 11) to the MDAC 21 is turned off in block A116. The propagation wait subroutine is then entered in block A118 to allow the change in gain to propagate to the comparator 106. The comparator 106 is then read in block A120 by sampling pin PA3 and determining whether the comparator has remained low. If the comparator 106 has not remained low, the condition indicates that not enough gain is being applied to the MDAC 21 to zero the channel, and, therefore, the most significant bit is again turned on in block A122. If the comparator 106 does not remain low, the gain that needs to be applied to the MDAC 21 to zero the channel is greater than the most significant bit. However, if the comparator 106 remains low, this indicates that there is still too much gain but the gain necessary to zero the channel is less than the most significant bit (800H) and it is left off.
Thereafter, a successive approximation of the MDAC gain needed to zero the comparator 106 is determined by toggling each bit after the most significant bit in succession. Blocks A124-A144 toggle bits 10-0 OFF in succession and, similar to the previous case, either leave them off if the comparator 106 remains low or turn them back on if the comparator goes to a high state. The MDAC gain is found in this manner to zero the channel by successive approximations to a resolution of 12-bits and is thereafter saved in block A146.
Next, the MDAC gain which has been determined is used to calculate a noise margin in block A148. The noise margin is determined as 1/256th of the original MDAC gain which zeros the channel. This is the amount by which the signal may change when the channel is measured but still be considered to be zeroed. To determine if there is noise on the channel greater than this margin, the offset calculated is subtracted from the derived MDAC gain in block A150. The new gain (MDAC (ZERO)-OFFSET) is then applied to the MDAC 21 and allowed to propagate to the comparator 106 while the system waits in block A152. The comparator 106 is then read in block A154 to determine if comparator output is at a high value. If the comparator 106 does not make a transition to a high value when the offset is subtracted, then the zero gain was not calculated sufficiently accurately and the negative branch of block A154 transfers the program to block A156 where the zero fail flag is set.
If the system passes the first test, then in block A158 the gain derived from zeroing the channel is added to the offset. That new gain (MDAC (ZERO)+OFFSET) is then applied to the MDAC 21, and the system waits until the changes of the signal propagates through to the comparator 106 by calling the subroutine propagation wait in block A160. The comparator 106 is then read to determine if it is at a high logic level in block A162. If the comparator has not toggled to a low value with the addition of the offset, the MDAC gain which was derived will not zero the channel and, therefore, in block A164 the zero fail flag is again set.
After exiting from block A156, A162 or A164 the program continues at block A166 where the original MDAC gain is reset for the channel and the propagation wait routine is entered in block A168. If the tests are all passed, then that MDAC gain derived from the successive approximations will be used in the normal mode 408 for subsequent analysis readings of that channel. However, if the zero fail flag is set, then when the program returns from the SAR subroutine, the program in the zero mode 406 will branch to the zero request mode 410 to again request zeroing and a determination if the noisy channel can be zeroed.
The successive approximation routine for measuring the temperature will now be more fully explained with respect to FIG. 12. The temperature measurement subroutine labeled SAR TEMP begins in block A170 by setting the MDAC gain to its highest value (FFF). The system then waits for the propagation delay in block A172 so that the change in gain is allowed to change the output of comparator 106. Next in block A174, similar to the successive approximation routine for zeroing each channel, bits 11-0 are toggled to successively approximate the gain which balances the two legs of the bridge comprising the fixed reference voltages +15 V, and reference resistors 120, 122, 124 and the thermistor 118. After this value has been found, it is stored for further use as a reference temperature when the normal mode 408 is entered.
The subroutine CHECK TEMP will now be more fully explained with respect to FIG. 14. In block A192 the MDAC gain which caused a zero for the temperature channel is read from memory. The most significant 4-bits of that MDAC gain word are used to determine one of sixteen pointer values in block A194. These pointer values are applied to high and low temperature tables from which offset values are derived. The offset value is representative of a change of .+-.6.degree. C. from the original temperature value and indicated by the MDAC gain. Because the thermistor 118 and temperature versus voltage characteristics of the loop gain are nonlinear, this table look up allows the correct offset value from any temperature over the range of operation of the system to be determined for an accurate .+-.6.degree. C. The gain offset is then subtracted from the original MDAC gain in block A198 and that new gain (MDAC (TEMP ZERO)-(TEMP-6.degree. OFFSET)) applied to the MDAC 21. The propagation wait subroutine is entered in block A200 and thereafter the comparator 106 tested in block A202.
If the comparator 106 is at a high logic level, it indicates that the present temperature is higher than a temperature -6.degree. C. from the original temperature. However, if the comparator is not high, then the temperature has changed more than 6.degree. C. in the negative direction, and thus, the temperature fail flag will be set in block A204. If the first temperature offset test is passed, then the affirmative branch from block A202 continues the program at block A206 where the +6.degree. C. offset is added to the original MDAC temperature gain. This new gain (MDAC (TEMP ZERO)+(TEMP+6.degree. OFFSET)) is applied to the MDAC 21 and the system enters the propagation wait routine in block A208. The comparator 106 again is tested to determine the state level of its output. A low output indicates that the temperature has not changed in the positive direction by more than 6.degree. C. since the temperature channel was measured during the zero mode. Therefore, the original MDAC temperature gain is saved in block A214, and the subroutine returns from where it was called. However, if the comparator 106 is high, then in block A212 the temperature fail flag is set indicating that the present temperature is outside of the offset range in the positive direction. If the temperature fail flag is set in either block A204 or A212, then the system will loop back from block A48 to the zero request mode 410 to request an auto-zero of all channels.
The interrupt routine which is entered once every 10 milliseconds upon the hardware timer interrupt is illustrated in a detailed flow chart in FIGS. 15A and 15B. The first two operations of the interrupt routine are to determine the status of the front panel bit (output of MUX 108), which is read through the multiplexer 108 at pin PA4, and the comparator bit, which is read at pin PA3 of the microprocessor 90. Initially, in block A216 the routine loads the accumulator of the microprocessor 90 with the contents of a location labelled FRONT PANEL storing the status of the front panel bit. Next, the pin PA4 is tested in block A218 to determine its state. If the bit is on, then the channel bit is ORed with the accumulator in block A224. If the bit is off, then the channel bit is complemented in block A220 and ANDed with the accumulator in block A222. These operations update the contents of the location FRONT PANEL to the current state of the signal read from pin PA4. The results from logical operations of block A222 or block A224 are stored in the location labeled FRONT PANEL in block A226. The front panel bit at pin PA4 is, therefore, set in the particular channel location being addressed by the demultiplexer 108 such that the location labelled FRONT PANEL will store the results for all four channels to indicate whether or not the electronic calibration is in operation at Y.sub.0, whether the CO.sub.2 channel is greater than a predetermined offset at Y.sub.1, and whether there is a zero command at Y.sub.2, and whether Y.sub.3 is high or low. In the embodiment shown in FIG. 4B, Y.sub.3 is grounded and should always be low. However, this signal line can be used as a configuration indication or other optional signal as needed.
A similar routine reads the comparator bit at pin PA3 which starts in block A228 by loading the accumulator with the contents of the location labeled COMPARATOR. In block A230 the system reads pin PA3 to determine the state of the comparator 106. If the comparator state pin PA3 is high, then the channel bit is ORed with the accumulator in block A136 and, conversely, if the comparator state is low, the channel bit is complemented and then ANDed with the accumulator in blocks A232 and A234. Thereafter, the results are stored in the location labeled comparator in block A238. These operations update the contents of the location COMPARATOR to the current state of the signal read from pin PA3. For each one of the analyzing channels (0-2) and the temperature channel, this provides an indication once every 10 milliseconds of the status of the comparator 106 for that particular channel. The location COMPARATOR may then be read in the other portions of the program and is used as a RAM image of the comparator output for each of the four channels. The image is updated on every loop through the interrupt routine.
After updating the locations FRONT PANEL and COMPARATOR (images); the system disables the inputs to the sample and hold circuits in block A240. This operation is accomplished by bringing pin PA2 to a high logic level which is inverted in an inverter U7-8,9 and applied to the inhibit input INH of the multiplexer 22. The inhibit signal tristates the output of the multiplexer 22 such that the voltage on the capacitors of the sample and hold circuits are maintained at their last value. This operation is to prepare for changing channels and allows the system to come to a steady state before the input voltages multiplied by the new gain and channel are applied to the sample and hold circuits, thereby reducing any noise or transient signals which may be propagated by the MDAC 21 upon switching.
In the next blocks A242-A264 the timers are updated and the blink timer changed. In block A242 the hardware timer is incremented, while in block A244 the blink timer is incremented. Thereafter, in block A246 the propagation timer is incremented and the blink flag tested in block A254. If the blink flag is not true, then there is no need to determine whether it is appropriate to change the state of the ready line. If the blink flag is true, the blink operation is accomplished by toggling the ready line during times when the blink flag is true at one half second intervals. Therefore, in block A256 the blink timer is compared to a constant (50) to determine if a half second has elapsed. If the time has elapsed, then the program continues to block A260, otherwise it returns to the main path at block A266. If a half second has passed, the blink timer is cleared in block A260 and the ready line tested in block A262. The ready bit is inverted in blocks A258 and A264 by setting the bit if it is cleared, or by clearing the bit if it is set, thus being on for a half second and then being off for a half second. The program then returns to the main path at block A266 to await the passage of another half second.
Next, the program will determine the next channel which should be selected by the multiplexer 22 and the demultiplexers 20, 104, and 108. The channels which are selected by the multiplexers are channels 0-3 corresponding to the HC channel, CO channel, CO.sub.2 channel, and temperature channel, respectively. In block A266 the location channel bit is cleared, and in block A268 the accumulator is loaded with the contents of the location channel. The accumulator is then incremented in block A270 to choose the next channel and ANDed with a mask (11) in block A272. The result is stored back into the location channel in block A274. This operation will cause an incrementation of the channel select number from 0-3 and then a repeat of the sequence.
A series of tests are then executed to determine the presently addressed channel and to set up a pointer to the calculated MDAC gain which is to be used for that channel. In block A276 the location channel is tested to determine if it is zero. If the answer is affirmative, then the program branches to block A278 where the index register X is loaded with a pointer directed to the MDAC gain for the HC channel. Next, in block A280 bit 0 of the location channel bit is set. If the contents of the location channel are equal to a one instead of zero, then in block A282 the program continues at block A284 where the index register X is loaded with a pointer indicating the location of the MDAC gain for the CO channel. Thereafter, the location channel bit is modified by setting bit 1 of that word in block A286. However, if the location channel contains a two, then negative branches from blocks A276 and A282 will cause the test in block A248 to be executed. An affirmative channel 2 branch causes the index register X to be loaded with the CO.sub.2 pointer in block A250 and bit 2 to be set in the location channel bit in block A252. If the channel location is none of channels 0, 1 and 2, then the channel location must contain an indication of the temperature channel (3) and, therefore, in block A288 the temperature pointer is loaded into the index register X. Next in block A290, bit 3 of the location channel bit is set to indicate that the input to the temperature channel is to be measured during this time interval.
After setting the index pointer and the channel bit depending upon the value of the location channel, the program in block A292 loads the MDAC 21 with the gain from the location determined by the pointer. Thereafter, the timer is reinitialized in block A294. The last part of the interrupt routine includes setting the new multiplexer channel selection on bits PA0, PA1 of the output of the microprocessor 90. In block A296 the multiplexer address bits are first cleared and then the channel pattern from the location channel bit are added to the location storing the multiplexer address bits in block A298. The new pattern is then output to the multiplexers in block A300, and the sample and hold circuits are enabled by bringing pin PA2 of the microprocessor 90 high. The interrupt routine thereafter, reverts to the location from which the routine was called.
While a preferred embodiment of the invention has been illustrated, it will be obvious to those skilled in the art that various modifications and changes may be made thereto without departing from the spirit and scope of the invention. Specifically, the thermistor 118 can be placed on either side of the MDAC 21, as these elements are connected in the same leg of a bridge which to be balanced. Thus, the resistor 122 connected to the output of the multiplexer 22 can be the temperature sensitive element. Also, as stated above, other gases can be analyzed. It is also possible to provide other time delays. For example, 20 minutes between zeroing appears is practical. It is also possible to operate without a host in certain embodiments. In this case the pump and solenoid signal may operate a relay or SCR to actuate drivers for the pump 126 and solenoid 128.
The following appendix is a machine language listing of the system program and subroutines described for FIGS. 8-15. ##SPC1##
Claims
What is claimed is:
1. Automatic zeroing apparatus for zeroing an infrared gas analyzer automatically upon the occurrence of preselected conditions to indicate zero in the absence of absorption of infrared radiation by a gas mixture being analyzed, the gas analyzer having a sample cell for containing a gas mixture to be analyzed, source means for directing infrared radiation through the sample cell over a range of wavelengths, direction means responsive to infrared radiation at at least one respective preselected wavelength to produce at least one repective detection signal systematically related to the infrared radiation at the respective preselected wavelength passing through the sample cell from the source means which respective wavelength corresponds to the absorption wavelength characteristic of a respective predetermined gas, and signal processing means responsive to said at least one direction signal for indicating the relative concentration of the respective predetermined gas in the mixture being analyzed, said signal processing means including reference means for producing a reference signal, and output means responsive to said at least one detection signal and said reference signal for outputting an output signal systematically related to the difference between said at least one detection signal and said reference signal, said automatic zeroing apparatus comprising: means for filling the sample cell with gas substantially nonabsorbent of infrared radiation at the respective characteristic wavelength, comparator means responsive to said output signal for producing an error signal when said output signal differs from zero, and gain control means responsive to said error signal for automatically controlling the signal level of said at least one detection signal to reduce said output signal substantially to zero with said substantially nonabsorbent gas filling the sample cell.
2. Automatic zeroing apparatus for zeroing an infrared gas analyzer automatically upon the occurrence of preselected conditions to indicate zero in the absence of absorption of infrared radiation by a gas mixture being analyzed, the gas analyzer having a sample cell for containing a gas mixture to be analyzed, source means for directing infrared radiation through the sample cell over a range of wavelengths, detection means responsive to infrared radiation at at least one respective preselected wavelength to produce at least one respective detection signal systematically related to the infrared radiation at the respective preselected wavelength passing through the sample cell from the source means which respective wavelength corresponds to the absorption wavelength characteristic of a respective predetermined gas, and signal processing means responsive to said at least one detection signal for indicating the relative concentration of the respective predetermined gas in the mixture being analyzed, said signal processing means including reference means for producing a reference signal, and output means responsive to said at least one detection signal and said reference signal for outputting an output signal systematically related to the difference between said at least one detection signal and said reference signal, said automatic zeroing apparatus comprising: means for filling the sample cell with gas substantially nonabsorbent of infrared radiation at the respective characteristic wavelength; comparator means responsive to said output signal for producing an error signal when said output signal differs from zero; and gain control means responsive to said error signal for automatically controlling the signal level of said at least one detection signal to reduce said output signal substantially to zero with said substantially nonabsorbent gas filling the sample cell, said gain control means having a zero mode in which the analyzer is zeroed and a normal mode in which analysis is made and comprising mode control means for placing said gain control means in said zero mode, means for operating said mode control means periodically to place said gain control means in said zero mode periodically, means associated with said zero mode for operating said means for filling to fill said sample cell with said substantially nonabsorbent gas, means in said zero mode responsive to said error signal for adjusting the signal level of said detection signal by a factor to balance said reference signal to reduce said output signal substantially to zero, means for storing a representation of said factor, and means in the normal mode for adjusting the signal level by said factor according to said representation stored from the last preceding zero mode, such that after the zero mode said analyzer remains calibrated in the normal mode in the presence of an absorbing gas being analyzed.
3. Automatic zeroing apparatus according to claim 2 further including means responsive to said output signal for generating a high concentration signal when said output signal exceeds a predetermined limit, and means responsive to said high concentration signal for inhibiting said mode control means from entering said zero mode while the concentration of the respective gas is greater than that corresponding to said predetermined limit.
4. Automatic zeroing apparatus according to claim 3 including electronic calibration means for adding a calibration offset signal to said reference signal during calibration, and means responsive to operation of said electronic calibration means for inhibiting the operation of said mode control means from entering said zero mode during electronic calibration.
5. Automatic zeroing apparatus according to claim 3 wherein said respective gas is carbon dioxide.
6. Automatic zeroing apparatus according to claim 2 further including means under the control of an operator for generating a zero command signal, said mode control means being responsive to said zero command signal to place said gain control means in said zero mode.
7. Automatic zeroing apparatus according to claim 2 wherein said gain control means includes a warm-up mode entered upon start-up of the apparatus and further including means in association with said warm-up mode for operating said means for filling to fill said sample cell with said substantially nonabsorbent gas, means in said warm-up mode responsive to said error signal for periodically adjusting the signal level of said detection signal to drive said output signal substantially to zero, and means responsive to successive adjustments of said signal level for operating said mode control means to place said gain control means in said zero mode when successive adjustments are within a predetermined difference.
8. Automatic zeroing apparatus according to claim 7 wherein said predetermined gas is carbon monoxide.
9. Automatic zeroing apparatus according to claim 2 including electronic calibration means for adding a calibration offset signal to said reference signal during calibration, and means responsive to operation of said electronic calibration means for inhibiting the operation of said mode control means from entering said zero mode during electronic calibration.
10. Automatic zeroing apparatus according to claim 2 further including temperature indicating means responsive to temperature for producing a current temperature signal indicative of ambient temperature, means for storing said temperature signal produced during a zero mode, means responsive to said stored temperature signal for providing first and second temperature offset signals corresponding to the offset of temperature above and below the ambient temperature corresponding to said stored signal by at least one predetermined temperature difference, means operative in the normal mode for comparing said stored temperature offset signal with the current temperature signal to produce a temperature drift signal when said current temperature signal differs from said stored temperature signal stored during the last zero mode by more than said first and second temperature offset signals, and means responsive to said temperature drift signal for operating said mode control means to place said gain control means in said zero mode.
11. Automatic zeroing apparatus according to claim 10 further including means responsive to said output signal for generating a high concentration signal when said output signal exceeds a predetermined limit, and means responsive to said high concentration signal for inhibiting said mode conrol means from entering said zero mode while the concentration of the respective gas is greater than that corresponding to said predetermined limit.
12. Automatic zeroing apparatus according to claim 11 including electronic calibration means for adding a calibration offset signal to said reference signal during calibration, and means responsive to operation of said electronic calibration means for inhibiting the operation of said mode control means from entering said zero mode during electronic calibration.
13. Automatic zeroing apparatus according to claim 11 wherein said respective gas is carbon dioxide.
14. Automatic zeroing apparatus according to claim 10 further including means under the control of an operator for generating a zero command signal, said mode control means being responsive to said zero command signal to place said gain control means in said zero mode.
15. Automatic zeroing apparatus according to claim 10 wherein said gain control means includes a warm-up mode entered upon start-up of the apparatus and further including means in association with said warm-up mode for operating said means for filling to fill said sample cell with said substantially nonabsorbent gas, means in said warm-up mode responsive to said error signal for periodically adjusting the signal level of said detection signal to drive said output signal substantially to zero, and means responsive to successive adjustments of said signal level for operating said mode control means to place said gain control means in said zero mode when successive adjustments are within a predetermined difference.
16. Automatic zeroing apparatus according to claim 15 wherein said predetermined gas is carbon monoxide.
17. Automatic zeroing apparatus according to claim 10 including electronic calibration means for adding a calibration offset signal to said reference signal during calibration, and means responsive to operation of said electronic calibration means for inhibiting the operation of said mode control means from entering said zero mode during electronic calibration.
18. Automatic zeroing apparatus according to any one of claim 2 to 17 wherein there is a plurality of predetermined gases, and said means for adjusting signal level includes means for independently adjusting the signal level in respect to each of said plurality of gases by a respective said factor.
19. Automatic zeroing apparatus according to claim 18 wherein said means for adjusting said signal level comprises a multiplying digital to analog converter for adjusting each said signal level by said respective said factor.
20. Automatic zeroing apparatus according to claim 19 wherein said gain control means includes a digital processor for determining said factors for said multiplying digital to analog converter and for storing said representations of said factors upon completion of zeroing.
21. Automatic zeroing apparatus according to any one of claim 2 to 17 wherein said means for adjusting said signal level comprises a multiplying digital to analog converter for adjusting said signal level by said factor.
22. Automatic zeroing apparatus according to claim 21 wherein said gain control means includes a digital processor for determining said factor for said multiplying digital to analog converter and for storing said representation of said factor upon completion of zeroing.