vkV ^ V ' ^^ >' • » 'VVT' ./V \ 0°*-:^^^% /\gi^'\ O^^.t^^'^^o /\*-^%' /^'•\o^ ""V'-7^*\-&^^ "v'^^^^o^^ %;^?^-\/^ "%''^*y %'^?^->' ;-^^'V^ "V^^V^ ^^'^^V^"^ "V^^V^ \J'^^V "V^^V ^^^/ J" "^o^'^P^y^ ^^'^^V"" "v^^V^ \J'^^V "V^^V IC 8934 Bureau of Mines information Circular/1983 A Dynamic Gas-Mixing System By C. R. Carpenter, J. E. Chilton, and G. H. Schnakenberg, Jr. UNITED STATES DEPARTMENT OF THE INTERIOR ^^tU^vUili^/. ^MjUJiM ^ yit^'jX Information Circular)8934 A Dynamic Gas-Mixing System By C. R. Carpenter, J. E. Chilton, and G. H. Schnalcenberg, Jr. UNITED STATES DEPARTMENT OF THE INTERIOR James G. Watt, Secretary BUREAU OF MINES Robert C. Norton, Director <^5 ^0' % .3H This publication has been cataloged as follows: Carpenter, C. R. (Clarence R.) A dynamic gas-mixing system. (Information circular / Bureau of Mines ; 8934) Includes bibliographical references. Supt. of Docs, no.: I 28.27:8934. 1. Gases. 2. Mixing. I. Chilton, J. E. II. Schnakenberger, George H. III. Title. IV. Series: Information circular (United States. Bureau of Mines) ; 8934. TN295.U4 [TP242] 622s [665.7] 83-600081 CONTENTS Page h Abstract 1 Introduction 2 Background 2 System design 3 Basic concepts 3 Detailed system description 8 Mass flow controller description 13 Flow rate calibration 13 System operation and performance 16 Dilution ratios 16 Flow setting reproducibility 16 Flow controller linearity 16 Binary mixture preparation 19 Conclusion 20 Future plans 20 Appendix A. — Suppliers ' addresses 21 Appendix B, — Digital timer operation and construction 22 Appendix C. — Gas-mixing system program for Texas Instruments SR52 calculator... 28 ILLUSTRATIONS 1 . Basic elements of the gas dilution system 4 2 . Flow control and mixing elements of the gas dilution system 4 3. Electrical elements of the flow system 4 4 . Linear and loop mixing manifold flow patterns 5 5. Four controllers and two manifolds create two gas blends simultaneously.. 6 6. Flow controller solenoid valves simultaneously actuated to direct flow to alternate manifold. 7 7 . System with manually operated three-way valves 8 8. Location of controls and connections for the manifold console and the control console 9 9. Parts identification and functional layout of the manifold console 9 10. Functional diagram of the dynamic gas-mixing system 10 1 1 . Electrical schematic of the gas-mixing system 12 12. Functional representation of the electronic mass flow controller sensor and valve 13 13. Calibration curve for controller 1 17 14. Calibration curve for controller P 17 15. Calibration curve for controller 2 17 16. Calibration curve for controller 3 17 17. Calibration curve for controller 4 18 18. Calibration curve for controller 5 18 19. Measured methane concentration versus desired values 19 vSi B-1. Functional block diagram of the digital timer 23 i B-2. Time base generation and selection schematic 23 B-3. Control and gating logic schematic 24 B-4 . Counters and display schematic 24 vj B-5 . Soap bubble transit detection circuitry 26 y1) B-6 . Timer power supply schematic 26 B-7. Timing diagram of logic states of control line of the timer 27 C-1 . User instructions for gas dilution mixing program 29 C-2. Program listing for gas dilution mixing program 30 Ji XI TABLES Page 1 . Flow conversion factors 15 2. Reproducibility of a flow setting 16 3. Summary of the linear regression analysis for Tylan flow controllers 18 4. Desired and obtained methane-air mixtures 19 UNIT OF MEASURE ABBREVIATIONS USED IN THIS REPORT A ampere ml milliliter ° C degree Celsius ml/min milliliter per minute cm centimeter msec millisecond ft foot yA microampere hr hour liF microfarad Hz hertz pm micrometer in inch ys microsecond in Hg inch of mercury yyF micromicrofarad k kilohm n ohm K kelvin pet percent kHz kilohertz ppm part per million 1/min liter per minute psi pound per square inch M megohm sec second mb millibar V volt MHz megahertz vol-pct volume-percent min minute A DYNAMIC GAS.MIXING SYSTEM By C. R. Carpenter, J. E. Chilton, and G. H. Schnakenberg, Jr. ABSTRACT A dynamic gas-mixing system assembled by the Bureau of Mines for the generation of precise gas mixtures from sources of concentrated or pure gases and diluent gases is described. A set of electronic mass flow controllers with maximum flows ranging from 10 to 5,000 ml/min, a so- called Pure Air Generator, and gases in cylinders are used to generate differing gas concentrations. The repeatability of the delivery rate of a flow controller, measured on different days, has a precision of 0.3 pet of the setting. This dynamic gas-mixing system reduces the cylinder inventory and the demurrage charges for special gas mixtures. Gas mixtures that cannot be shipped commercially, such as flammable mixtures of methane in air, can be conveniently prepared by this sys- tem. The maximum dilution ratio is 2.5 x 10^. Thus, gas mixtures can be made over a wide range of concentrations, from the percent region (by dilution of a pure single-component gas) to fractional parts per million (by dilution of a premixed standard, e.g., a 1,000-ppm mix- ture) . This system is especially useful in determining the response of gas detection devices over the entire range of their measurement. Be- cause the controllers are voltage controlled they lend themselves eas- ily to automated control using computer-based systems. ^Electronics technician. ^Research chemist. ^Supervisory research physicist. All authors are with the Pittsburgh Research Center, Bureau of Mines, Pittsburgh, PA. INTRODUCTION There are many applications for gas de- tector devices for personal safety and property protection in the mining indus- try. The devices are used to detect the presence and concentration of (1) toxic gases (e.g., hydrogen sulfide), (2) gases that are precursors of fire (e.g., car- bon monoxide) , (3) gases that may accumu- late to form explosive mixtures (e.g., methane), and (4) life-threatening atmos- pheres (e.g., deficiency of oxygen). These gas detectors must be sensitive over a wide range of concentrations to selected species of gas. The performance of each type of detector should be evalu- ated prior to its general use to insure its adequacy and suitability. The evaluation of gas instrumentation designed to detect low concentrations (parts per million) of CO, NO, NO2 , SO2 , and H2S and high concentrations (volume- percent) of O2 , CH4 , and CO2 for use in mining environments requires gas mixtures of high accuracy for both instrument cal- ibration and instrument performance eval- uation. The performance evaluation of gas instrumentation will use well- characterized and accurately determined gas calibration mixtures for tests including those for accuracy-over-range, drift, and precision. Calibration gases may be purchased in cylinders from commercial sources but not without several disadvantages: high ini- tial cost; uncertainty in accuracy of analysis; slow, unpredictable delivery; a finite cylinder capacity; and extensive demurrage charges incurred in the storage of a large number of reference gas mix- tures in cylinders. Certain mixtures of gases may not be available because of their flanmability. One solution to this problem of obtaining different concentra- tions of a gas is to dilute a high- concentration gas accurately to a lower concentration, A dynamic gas dilution system involves the mixing of two gas streams to produce a predetermined spe- cific concentration of a gas in a flowing gas sample. This report describes the design and fabrication of a dynamic gas dilution system that satisfies the demands of gas instrument evaluation test methods that require a flowing source of several accu- rately known gas mixtures. BACKGROUND The Bureau of Mines instrumentation group at the Pittsburgh Research Center has used various methods in its attempts to design and build a universal dynamic gas-mixing system. One system employed precision needle valves to control the flows of the test gas and the diluent gas. The flow rates were monitored with high-quality Labcresf*'^ rotameters. Although this method was exclusively used for several years, it had several deficiencies. The main deficiency was its imprecision or lack of resetabillty; ^Use of trade names is for identifica- tion only and does not imply endorsement by the Bureau of Mines. ^Suppliers' addresses are listed in appendix A. each and every flow setting required ver- ification using a soap bubble meter to obtain desired mixing accuracies (±1 pet or less error) . Another deficiency was its lack of long-term stability. One source of this instability was the vari- ation in the regulation of pressure from the cylinder gases. At high dilu- tion ratios, a small change in input gas pressure of the minor component gas caused a large change in the diluted gas concentration. In an effort to improve our testing facilities, we experimented with restric- tors. We found the usable range of flow rates of each resistor to be limited owing to their nonlinearity . Additional- ly, the instability problem was not significantly improved even though additional pressure regulators were placed in the input lines. Our next major effort was centered on the use of electronic mass flow con- trollers (Tylan model 261). Preliminary tests on the devices indicated that they would be very suitable in our applica- tion. We had several reasons for select- ing electronic mass flow controllers (MFC): 1. As flow controllers, the mass flow rate would be independent of the pressure upstream of the controller. 2. The mass flow could be set by ap- plying a voltage between 0.1 and 5 V, and voltages of sufficient stability could be easily produced. Furthermore, the flow rate would be approximately proportional to the control voltage over the range of the controller. 3. The controller provided an output signal of 0.1 to 5 V, also approximately proportional to the mass flow of the gas. This signal could be accurately read by a digital panel meter, and it is also equal to the flow rate control signal when con- trol of the flow was attained by the controller, 4. The manufacturer claimed a preci- sion (resetability) of 0.2 pet of full scale, a precision which we thought nec- essary for our gas detection instrument investigations and evaluations. 5. Once a calibration curve was estab- lished for a particular controller for one gas, a simple application of the gas laws and well-known physical properties of gases could be applied to this curve to produce a calibration curve for other gases or for other ambient pressures and temperatures. 6. Since flows are voltage controlled, these controllers could be easily used in an automated testing facility should this be needed. The following sections de- scribe first the design concepts and rationale and then the full system design and performance of our dynamic gas dilu- tion system using mass flow controllers. SYSTEM DESIGN BASIC CONCEPTS The purpose of this gas-mixing system is to dilute a high concentration of test gas species to a more appropriate lower concentration for use in gas instrument testing, thus conserving gas and permit- ting a variety of concentrations from one gas cylinder. The inputs to the system are the one or more test gas species to be diluted (source gas) and the diluting gas (diluent) . The output is the diluted gas mixture. Because the output concentration is determined by flow rates, they become the critical elements in the system design. Figure 1 depicts the basic elements of the gas dilution system involving two components (source and diluent). The flow controllers are adjusted by the operator to produce the desired flow ra- tio and, therefore, the concentration of the blended gases. After mixing, the gas is conducted by suitable connectors to the system output. Figure 2 emphasizes the flow control and mixing elements of the system. The flow controllers control the flow rates of the gas in response to a command (set point) voltage. In our systems, this voltage, between 0.1 and 5 V, is ob- tained from a front panel control that is set by the operator (fig. 3). A voltage proportional to the flow rate is devel- oped by the flow controller. This volt- age, which, as previously mentioned, is equal to the command voltage when the flow is stabilized, is displayed by a digital panel meter (DPM) . As a refine- ment, since each controller has a DPM dedicated to it and since each covers different flow ranges, we have adjusted the calibration of each meter to provide a display that is approximately numeri- cally equal to the flow in milliliters per minute. Since it is necessary for Flow (fi) Flow 2 (fg) Mixing Total flow (ft) chamber Diluent gas Diluted concentration out (Cq) FIGURE 1, - Basic elements of the gas dilution system: source of a minor component at concentration C; flowing at rate f ^; a diluent gas, e.g., air flowing at rate f2; and a mixing element. The final concentration of the mix is C. x f i/ff flowing at a rate of f^ = f ^ + fj. Gas to be diluted Diluting gas Diluted *" gas output Mixer-manifold FIGURE 2. - Flow control and mixing elements of the gas dilution system. Signal out Moss flow controller Connnnand voltage in Press for input Scale factor Adjust Digital panel meter Set point C"? vdc (command) 0.1 Vdc FIGURE 3. - Electrical elements of the flow system. Each mass flow controller has associated with it a power supply (not shown) and a command signal to be set to the desired value by the user. A digital panel meter is appropriately scaled to read flow rate directly in milliliter per minute and is used to display eith- er the command signal to the controller or (normally) the flow rate signal output from the controller. the operator to see the flow setting com- mand voltage to set the particular con- troller, this voltage can be displayed on the same digital display, using a momen- tary contact pushbutton switch. The output flows of the controllers should be mixed thoroughly and be avail- able with a minimum response (lag) time. To accomplish thorough mixing, the "o o o •a ^ Stagnant ^{ ^ {^ region ^ ^' o Q. E o o <1> c o — . C 0) 8.-2 o "^ Low-flow region "High-flow 1 region ZOul Linear manifold Major component (diluent) ^r^ Minor component Unused controllers -Out Closed-loop manifold FIGURE 4. - Linear and loop mixing manifold patterns. manifold into which the controllers de- liver the gases should have no stagnant areas that could contain unmixed gases that would slowly diffuse into the mix- ture. To minimize response time, the to- tal volume must be kept small. The stag- nant volumes are substantially eliminated by having all controllers feed into a closed-loop manifold. When a number of controllers are connected together in a linear deadend manifold, the length of tube beyond the entrance point of the ac- tive controller farthest from the mani- fold outlet would contain stagnant vol- umes of gas of unknown composition (fig. 4). Furthermore, if this control- ler was delivering the minor component at a low flow, it would require a signifi- cant time to flush the section of the manifold between it and the entrance point of the major component. Creating a loop manifold by connecting the ends of a linear manifold to each other and exiting the manifold through a tee, as shown in figure 2, causes the flow from the high- est flowrate controller to split and sweep the entire manifold at relatively high flows, picking up and mixing with the minor component in the process. An internal volume of about 2 ml in the manifold is realized by employing stan- dard 1/4-in tubing for its fabrication; this keeps the response time small. Tur- bulence of the gas caused by 90" bends and the fittings in the closed loop in- sures complete mixing of the gases. The addition of a second mixer- manifold, the two three-way valves, and another output connector increases the operating flexibility of the system (fig. 5) . The two manifolds are designated as A and B. The need for this added flexi- bility is described in the following example. Suppose the operator wanted to chal- lenge a gas detection instrument with a test gas that rapidly changes concentration — a step change. To do this the operator would set up a gas mixture Inputs output B 5 and 4 detector) Dual mixer-nnanifold FIGURE 5. - Four controllers and two manifolds create two gas blends simultaneously. Solenoid valves for controllers 1 and 2 direct the gases into manifold A, with controller 1 having the larger of the two flows. Similarly controllers 3 and 4 feed manifold B, with controller 3 having the great- er flow. Compare output blends with those in figure 6. in manifold B using the flow controllers 3 and 4 in figure 5. The solenoid valves, 3 and 4, associated with these controllers would be set to deliver the flows to manifold B, The gas detection instrument would be connected to mani- fold B. Next, using flow controllers 1 and 2 with their solenoid valves set to deliver the flow to manifold A, the oper- ator would establish a different mixture in manifold A and route the output to a vent. To deliver a step change in gas concentration to the gas detection instrument, the operator must simultane- ously switch all four solenoid valves, whereupon controllers 3 and 4 would de- liver flows not to B but to A, and con- trollers 1 and 2 would deliver flows not to A but to B, as shown in figure 6, This, in effect, would challenge the de- vice under test with a "step function" of the test gas. This "step function," among other things , would be useful in determining the response time of the gas instrument being tested. Rapid switching of all of the flow controllers between manifolds A and B is assured by using three-way electrical solenoid valves at the output of each flow controller. Si- multaneous switching of all valves is ac- complished using a single switch. A Swaglok manually operated three-way valve located at the output of each con- troller provides additional flexibility to the system, as shown in figure 7. One position of this valve directs the gas to the solenoid manifolds selection valve described above. The other position di- rects the gas from the controller to an associated auxiliary output port. These auxiliary outputs provide the user with an undiluted gas at a controlled flow rate. The valve may also be used to cut off the gas flow by setting them to a midposition. The source gases and the nitrogen dilu- ent gas are usually obtained from cylin- ders. These gases are delivered from a cylinder storage area to the gas-mixing system by separate stainless steel tubes Inputs output B I and 2 detector) Dual mixer-manifold FIGURE 6. - Flow controller solenoid valves simultaneously actuated (energized or deenergized depending on original state) to direct flow to alternate manifold. Thus, at manifold A the gas mix rapidly changed from a blend of 1 and 2 to a blend of 3 and 4; similarly the output from B changed from a blend of 3 and 4 to a blend of 1 and 2. Compare with figure 5. ■,—r. Gas HID output B (~ii Gas output A Manual selector valve I Dual mixer-manifolds FIGURE 7. - System with manually operated three-v^/ay valves. from each cylinder. The pressure is reg- ulated to the operating input pressure range (10 to 40 psi differential) of the controllers . Some gases , such as NO2 , H2S, and SO2 , are conveniently obtained from a permeation tube system. These gases are normally diluted in a fume hood with the diluent gas controlled by the gas-mixing system; this is an ideal use of the auxiliary outputs. Air, when it is used as the diluent gas, is supplied from a commercially available generator of pure (zero) air. This air is made available to the gas-mixing system through a separate line. DETAILED SYSTEM DESCRIPTION The mixing system consists of two units , a manifold console and a control console (fig. 8), Three auxiliary sys- tems are routinely used with the dynamic gas-mixing system: a cylinder manifold, a commercially available Pure Air Gener- ator, and a permeation system. The manifold console (fig. 9) consists of two closed-loop manifolds (CLM-A and CLM-B) , six mass flow controllers (MFC) , six manual selector valves (MSV) , and six three-way solenoid valves (SV). Six stainless steel filters, 7-ym pore size, are located outside the console for easy replacement and are placed upstream of the MFC to prevent aerosols from entering the system. The gases to be mixed are connected to the system by compression seal bulkhead fittings on the top of the console, with one corresponding to each controller. Each gas flows through the filter and then directly to a mass flow controller. Immediately downstream of each controller is a manual selector valve that is used to route the gas to either an auxiliary output port or an associated solenoid valve which, in turn, routes the gas flow to one of two output manifolds (labeled CLM-A or CLM-B). With the manual selec- tor valve knob in the center position (horizontal) , the gas flow through that channel is cut off. The auxiliary outputs, when selected, allow the precisely controlled flow to be accessed directly when mixing is not re- quired or when it is to be used as a source for associated equipment, i.e., as the dilution gas for a permeation tube system. When gas mixing is desired, the manual selector valves associated with the se- lected controllers are positioned to di- rect the flow of the gases to the associ- ated solenoid valve. With the solenoid valves deenergized, the gases flow through CLM-B. In the manifold the gases are mixed and presented to the front pan- el output port labeled B. When a sole- noid valve is energized, the controlled flow from its associated controller p O O O O o. Gas inputs Top panel (detail) Auxiliary outputs O O O O O ^ o Front 12 3 4 Manual selector valves Hood vent output ^ g Manifold Q Manifold g output DPM-P DPM-I DPM-2 DPM-3 DPM-4 DPM-5 Po lo 2o 3o 4o 5o vPB-l /?=xPB-2 ^^=nPB-3 /?=^PB-4 /^PB-5 /?=^PB- OOOOOOOOOOOO S-l S-7 S-2 S-8 S-3 S-9 S-4 S-IO S-5 S-ll S-6 S-12 Manifold control panel o 9 S-14 I- 1 O S-13 Manifold console Control console FIGURE 8. - Location of controls and connections for the manifold console and the control console. Top panel r - — D— Out I — D— In Out 2 Out 3 — D— Out 4 -D— Out 5 — D-H In MSV positions Auxiliary Monifold Front panel LEGEND * 7-^m replaceable filters FIGURE 9. - Parts identification and functional layout of the manifold console. 10 Auxiliary gas outputs Gas inputs MSV< CLM-A [ H XLM-B '_ MFC-4 MFC -3 MFC-2 MFC-I MFC-P r~r Fume hood I I I I j lPPMl fo PB-2 iDPMl T-O ,PB-3 IDPMI *-o I P PB-4 iDPMl 1 — 1 I |PB-5 PPM I [— fO fpB-6 PPM ADJ rVsAn nAAn r\AAn nAAn iVsAn iWV> ■ "-^^ I i_ADJ_i. I ADJ__i |J\Dm I J^DJ_i. |.ADJ_i ,_2____^_3____^.i _^ 5 ^ 5 Vdc _ Flow control _ Manifold control , _6^ I ^ ^ ^ T Normal 6^ ^ , LReyers_e_Ps.|3 So^^ °II7 Vac Control console FIGURE 10. - Functional diagram of the dynamic gas-mixing system. The control console is con- tained within the dashed-line box; the manifold console contains the remaining parts of the diagram. 12 l_. enters CLM-A, The system is designed so that the solenoid valves are individually energized and deenergized. This feature adds flexibility to the system (fig. 10). The manifolds are constructed from standard compression fittings and lengths of 1/4-in stainless steel tubing. The closed-loop manifold design eliminates dead zones and thus minimizes the volume of pockets of unmixed gas that would slowly diffuse into the main stream. The internal volume of each manifold is kept to a minimum (approximately 2 ml) . A stainless steel tube from an adjacent fume hood is brought out to the front panel of the console and is terminated with a tubing-to-hose connector fitting. This fitting is marked "H" on the panel. This line is used to carry test gases to the hood. The control console is divided, func- tionally, into two sections, the flow rate control and manifold control, as in figure 8. The sections are electrically connected using jack J-1 and plug P-1. The control section (refer to figs. 8 and 10) has six miniature toggle switches, S-7 through S-12, that are used to ener- gize (and deenergize) the individual solenoids that select the output CLM-A or CLM-B for each controller in the manifold 11 console. Associated with these switches are two control switches. One switch, S-13, is designated "NORMAL/ RE SERVE" on the panel. The purpose of this switch is to produce a rapid change of all control- ler outputs from one manifold to the oth- er by simultaneously switching the state of each solenoid valve. Having two manifolds also permits two operators to use the mixing system. That is, one operator can use manifold A while the other operator uses manifold B, pro- vided there is no conflict in the use of individual controllers. The most used controllers in our laboratory have been the 4- to 200-ml/min and the 40- to 2,000-ml/min controllers. Our system in- corporates two of each. Therefore, this needed flexibility is usually available. In addition, of course, the AUX outputs can be used simultaneously and blended external to the system. The second switch, S-14 (POWER/OFF), is a master on-off switch for the solenoids, A neon panel lamp (I-l) illuminates when the solenoid power is on. The remainder of this console is de- voted to the control of the gas flow rates. Each flow controller or gas com- ponent "channel" is laid out vertically on the panel. Each channel includes a digital panel meter, a potentiometer for setting the flow-controlling (command) voltage, the switch that effects a momen- tary display of the control voltage on the panel meter, and the manifold se- lection switch described previously (fig. 8), In our system the flow channels labeled "P" and "1" through "4" use 3-1/2-digit, digital panel meters (DPM) to indicate the gas flow rates, Channel 5 employs a 4-1/2-digit meter since this is a 100- to 5,000-ml/min controller and requires a display resolution beyond that pro- vided by a 3-1/2-digit meter. Potentio- meters at the inputs (pin 1) of the panel meters (fig, 11) form voltage dividers which allow calibration of the read- outs to be equal to the flow rate in milliliters per minute. The range of the 3-1/2-digit meters is 0.000 to 1.999 V. The most significant digit of the 4-1/2-digit DPM (that is, the "1" of the 19999 maximum display) is not used. This gives a usable display of the digits up to 9999, as opposed to 1999 of a 3-1/2-digit meter. As mentioned earlier, the full-scale signal output for maximum flow of the controllers is 5 V regardless of the full-scale range of the control- ler. The potentiometers at the panel me- ter inputs allow adjustment for an exact panel meter indication, directly in mil- liliters per minute, of the rate of gas flow through the channel at one flow set- ting. In our system, we set the meter to agree with the flow near the full-flow rating of the controller in each channel. Subsequent checks at various flow rates within the range of each controller were performed and used to develop a set of calibration graphs (figs, 13-18) that re- late the controller output voltage to ac- tual measured gas flow. All values were corrected to our reference standard tem- perature and pressure (25° C and 28,92 in Hg), Using these graphs and tempera- ture and atmospheric pressure correction factors , the operator can routinely ob- tain mixing accuracies with less than 5 pet error. When greater accuracy is required, the flow rates are verified by the use of a soap-bubble meter. The operator sets the flow rates by ad- justing the set-point potentiometer (R-1 through R-6) in the desired channel (figs, 8 and 11), To facilitate this ad- justment, the controller input signals (set points) may be observed on the panel meters (in milliliters per minute) by de- pressing the PRESS FOR INPUT pushbutton switches (PB-1 through PB-6) . Each controller is energized or deener- gized independently by operating its POWER-ON/OFF switch (S-1 through S-6) . The power switches supply 115-Vac power to the operational power supplies, P.S.-2 and P.S,-1, (±15-Vdc and 5-Vdc reference supplies) and the panel meters. The me- ter associated with channel 5 uses an 12 LEGEND » Press for input Manifold console FIGURE 11. - Electrical schematic of the gas-mixing system. 13 external 5-Vdc (P.S.-3) supply; the other meters are powered directly from the switched 115-Vac line. The output from the 5-V reference supply is permanently connected to the clockwise end of all six set-point (command) potentiometers. The ±15 Vdc is connected to the individ- ual controllers through separate poles of the power switches . Thus only the controllers in use are powered, saving wear and tear on any unused controllers. MASS FLOW CONTROLLER DESCRIPTION resistance temperature sensors and heat- ers. This is shown in figure 12. A more complete explanation can be found in the Tylan manual. FLOW RATE CALIBRATION Calibration of the system was performed using zero air produced by a Pure Air Generator, a soap-bubble flowmeter set, an electronic timer, an electronic digital thermometer, and an aneroid barometer. The electronic mass flow controllers used in this system accurately and reli- ably measure and control the mass flow of gases. The principles of heat transfer along a capillary tube are used to devel- op a linear output signal of 0. 1 to 5 V over a selected flow range. The control- lers incorporate a valve and appropriate electronics to automatically regulate flow rates in response to an external command signal (0.1 to 5 V). The valve (fig. 12) , made of 316 stain- less steel, is a unique thermal expansion design that eliminates friction and moving seals . It is a small thin-walled tube with a ball welded to the end. The seat is a cone. Inside the tube is a heater wire that causes the tube to ex- pand relative to the outer shell, moving the ball relative to its seat and thereby varying the flow. The sensor section consists of a small, stainless steel capillary with external The Pure Air Generator, according to the manufacturer (AADCO) , produces air with less than 0.005 ppm hydrocarbons, carbon monoxide, carbon dioxide, methane, ozone, sulfur, hydrogen sulfide, and ox- ides of nitrogen but with 22.5 pet O2 in nitrogen. A methane reactor accessory, mounted within the cabinet, is a low- temperature catalytic oxidizer which re- moves all combustible hydrocarbons, in- cluding methane. The air compressor for this unit is capable of supplying up to 10 1/min of zero-air and is mounted in a housing to reduce noise to an acceptable level. The Teledyne soap-bubble flowmeter set contains a set of three glass tubes of a calibrated volume and traceable to the National Bureau of Standards (NBS). Each tube has a different volume (10, 100, and 1,000 ml) and fits in a glass base which contains a soap solution such as those available for detecting gas leaks. (We use SNOOP.) The gas is directed into the Sensor assembly ,lr/jj 'm ^^^^ Bypass assembly Gas out Valve assembly FIGURE 12. - Functional representation of the electronic mass flow controller sensor and valve. 14 glass base and passes out through the graduated tube. A squeeze bulb, fitted on the base, is used to raise the level of the solution in the base to block the entrance to the graduated tube. When the bulb is released, the liquid level drops; a film of the solution remains across the flow passage and moves up the graduated tube as it is pushed by the flowing gas. The flow rate is determined by measuring the transit time for the film to pass through a given volume. That is, flow rate = volume (ml) transit time (min) Corrections for temperature, atmospheric pressure, and gas composition if other than air are then applied to obtain the mass flow rate (i.e., a volume flow rate referenced to a given pressure and tem- perature) . Because our laboratory is slightly greater than 1,000 ft above sea level, and operating equipment keeps it warm, we standardize at 28.92 in Hg (979.3 mbar) and 25° C. To improve precision and eliminate the random operator errors in measuring bub- ble transit times, we developed a digital timer (appendix B) for which the start and stop signals are obtained from two phototransistors which detect the pas- sage of the moving film (bubble) that is illuminated from above. ^ The first phototransistor is positioned at the bot- tom or zero graduation on the tube. As the edge of the film passes, it acts as a "light pipe" that momentarily increases the light transmission to the phototran- sistor. This signal starts the timer. As the film passes the second phototran- sistor, positioned at the top graduation (10, 100, 1,000 ml) of the tube in use, a stop signal is similarly generated that stops the timer. Comparison between a series of soap- bubble transit times measured manually ^No commercial bubble transient timer was available at the time of our need. Presently, we understand such a device is commercially available from at least one source (Mast) . and those simultaneously measured by the phototransistor-timer verified the accuracy of the phototransistor separa- tion and the increased precision obtain- able with the timer. For a particular measurement a tube is selected to provide a sufficient transit time to insure adequate resolution at the flow rate being measured. The timer, designed and constructed specifically for this job, incorporates other features that allow its use as a general purpose laboratory timer. A block diagram and schematics for the timer are included in this paper as ap- pendix B. For use with the soap-bubble flowmeters, the 0- to 10-min range is normally selected. This gives a resolu- tion of +0.0001 min for times up to 9.9999 min. For gas temperature measurements, we use an electronic digital thermometer from Stow Laboratories. The platinum re- sistance sensor is attached to the gradu- ated tube of the flowmeter to measure, indirectly, the ambient temperature, and hence, the temperature of the gas. The digital display on this unit provides a resolution of +0.1° C. Barometric pressure is measured on either an aneroid barometer or an elec- tronic meter; the aneroid barometer is an analog device, and resolution of 0.01 in Hg is obtainable. The Serta Systems electronic meter resolves to 0.1 mbar. A short calculator program was devel- oped to expedite the calibration of the gas-mixing system. The program (appendix C) was written for the Texas Instruments (TI) SR-52 programmable calculator and can be directly entered in the newer TI 58 and TI 59 calculators by replacing the HLT commands with R/S. For calculators that do not use the Algebraic Operating System, the equation is fc = Ts Pa fi Ta • Ps (1) 15 where fc = corrected flow rate, ml/min, Ts = standardized temperature (25° C + 273.15), K, Pa = actual barometric pressure, Hg or mbar, Ta = actual temperature (° C + 273.15), K, Ps = standardized barometric pressure (28.92 in Hg or 979.3 mbar). and fg = flow rate of other gas, ml/min. TABLE 1. - Flow conversion factors and fi = indicated flow rate, ml/min. The permeation of the gas through the soap bubble did not cause any significant error in our flow measurements with transit times of less than 2 min. The flow controllers, as received, were calibrated for air by the manufacturer. One must use a correction factor to use the controllers with gases that have dif- ferent molecular structure. In their Operation and Service Manual, Tylan Corp. provided a table of correction factors for various species of gas. We conducted experiments in the laboratory to deter- mine the corrections required for meth- ane, carbon monoxide, and carbon dioxide. The values obtained were in very close agreement with the values published in the Tylan table. Typical examples of this correlation are shown in table 1. Flow controller 2 was set during each gas measurement to an indicated flow rate of 150 ml/min (on as- sociated panel meter). The bubble meter tube volume was 100 ml. The flow rates shown are corrected to our standard tem- perature (25° C) and pressure (28.92 in Hg). The equation for the conversion factors is r - 9, where Cf = conversion factor, fa = flow rate of air, ml/min. Manufacturer's Experimental Measured Gas conversion conversion flow factor factor rate, ml/min Air 1.00 1.00 147.1 CH4 .72 .72 105.9 CO 1.00 .99 146.1 CO2 .74 .74 108.9 Most of the test gases used in our testing programs are obtained from cylin- ders that are located remotely from the dynamic gas-mixing system. We have found it undesirable to store some gases in cylinders (NO2 SO2 , H2S, etc.) because of reaction with the cylinder walls (unsta- ble concentration) or toxicity of the gas. For generating mixtures of these gases, we use a permeation tube system. Our permeation system uses a Forma Sci- entific water bath in which the tempera- ture is held constant at 20° ±0.1° C by an integral refrigeration and heating sys- tem. Two Pyrex glass permeation tube holders (U-tubes) are immersed in the wa- ter bath. This allows us to generate two species of test gas simultaneously. We use 10-cm-long permeation tubes , which gives an active tube life of more than 12 weeks at the operating temperature. Dry nitrogen is used as the carrier gas for these tubes at a flow rate of 100 ml/min for each tube holder. The two outputs from the system are brought out inside a fume hood which vents the gases when they are not in use. Under the operating con- ditions , the output concentrations are relatively high (for NO2 , ~75 ppm) and the gases are normally diluted to usable levels by mixing with air or nitrogen from the flow control system. The output concentration from this sys- tem is determined gravimetrically by mea- suring the weight loss of the permeation tube every few days. The short-term and running-average weight losses are com- puted, recorded, and graphed. 16 SYSTEM OPERATION AND PERFORMANCE DILUTION RATIOS The concentrations of test gas required by the instrument evaluation laboratory range from high percentage to low parts per million levels, a range of 1,000,000 to 1. As a compromise the design goal of a dilution ratio of 1:1,000 was chosen. This will produce a test gas concentra- tion of 0.1 pet (1,000 ppm) from a source of pure gas or, if a cylinder of accu- rately known premixed gas (e.g., a 1,000-ppm concentration) is used as the source, low-part s-per-million gas concen- tration levels can be achieved (i.e., 1,000 ppm/ 1,000 = 1 ppm). Of course, in- termediate concentrations can be obtained by adjusting the dilution ratio; there- fore, any test gas concentration between 1 ppm and 100 pet is available for our gas detection device evaluation efforts. This satisfies our flexibility require- ments and reduces the cylinder gas inven- tory to a maximum of two cylinders for each species of gas required for testing. In our experience gas detection devices require flows of up to approximately 1 1/min, To test several devices from a common source, flows up to 5 1/min might be required. Therefore, for 1:1,000 di- lution ratio, we selected four flow rate ranges for our system controllers to cov- er this dilution range at the desired output flow rates: 0.2 to 10, 4 to 200, 40 to 2,000, and 100 to 5,000 ml/min. These ranges allow the total output flow rates to be adjusted, as required for testing, between 200 ml/min and 5 1/min at the maximum dilution ratio. The maxi- mum dilution ratio is thus 0.2 to 5,000, or 1 to 25,000. FLOW SETTING REPRODUCIBILITY The reproducibility of setting the gas flow of the Tylan mass flow controller was calculated by setting a fixed flow value on a controller and periodically measuring the resulting gas flow on dif- ferent days over a period of several weeks . Since the measurements were made on different days and thus at different ambient temperatures and barometric pressures , the daily average flow read- ings were corrected from measured temper- ature and pressure to our chosen refer- ence temperature and pressure of 25° C and 28.92 in Hg (979.3 mbar) . This value of the pressure corresponds to normal atmospheric pressure 29.92 in Hg or (1,013.2 mbar) corrected for a 1,000-ft altitude above mean sea level (laboratory elevation) . A soap-bubble meter was used to measure the flows. The mean of 10 values of bubble transit time was taken for each flow determination. The grand average flow for a setting of 604 on the digital panel meter corresponded to 661.19 ml/min when corrected to the ref- erence temperature and pressure. The standard deviation of the flows was 1.95 ml/min. The reproducibility of the flow measurement and setting (the coefficient of variation) is 0.3 pet of reading; the measured flow values are summarized in table 2. TABLE 2. - Reproducibility of a flow setting, ml/min Measured Standard Flow corrected to flow deviation reference temper- ature and pressure 661.08 2.29 660.65 660.99 1.64 659.12 658.72 1.30 660.57 657.07 .32 660.22 671.42 1.86 664.77 667.33 2.82 661.82 FLOW CONTROLLER LINEARITY We performed our own check of the lin- earity of each mass flow controller. The control voltage of each mass flow con- troller was set at selected values, and the corresponding gas flows were measured using 10-, 100- , or l,000-ml soap-bubble tubes as appropriate, as described ear- lier. The flow values and the calibrated digital panel meter readings are sum- marized in figures 13 through 18. The flows were corrected to a flow at our reference temperature of 25° C and pres- sure of 28.92 in Hg (979.3 mbar). A least squares analysis of the data was 17 4 6 DIGITAL PANEL METER READING FIGURE 13. - The calibration curve for controller 1 is the measured controller flow rate corrected to 25° C, 28.89 in Hg versus scaled digital panel me- ter reading of controller output voltage (command signal). 60 80 100 120 140 160 180 2C DIGITAL PANEL METER READING FIGURE 14. - The calibration curve for controller P is the measured controller flow rate corrected to 25° C, 28.89 in Hg versus scaled digital panel me- ter reading of controller output voltage (command signal). 1 1 1 r 1 r 60 80 100 120 140 DIGITAL PANEL METER READING 180 200 FIGURE 15. - The calibration curve for controller 2 is the measured controller flow rate corrected to 25° C, 28.89 in Hg versus scaled digital panel me- ter reading of controller output voltage (command signal). 1,800 - 1 1 T- 1 ' 1 1 1 Air-.^^ 1 1,600 - y/ - 1,400 - 1,200 - / y^-^ y 1,000 / / z"*^^/^ ^Carbon / yo dioxide BOO / / 600 _ , / yy _ y/ .^^ / y'y^ 400 " / ^ ;^^- " 200 /'^ 1 1 1 , 1 1 800 1,200 DIGITAL PANEL METER READING 1,600 2,000 FIGURE 16. - The calibration curve for controller 3 is the measured controller flow rate corrected to 25° C, 28.89 in Hg versus scaled digital panel me- ter reading of controller output voltage (command signal). 18 5,000 800 1,200 DIGITAL PANEL METER READING 1,600 2,000 FIGURE 17. - The calibration curve for controller 4 is the measured controller flow rate corrected to 25° C, 28.89 in Hg versus scaled digital panel me- ter reading of controller output voltage (command signal). 3poo - a 2,000 - 1,000 1,000 2,000 3,000 DIGITAL PANEL METER READING 5,000 FIGURE 18. - The calibration curve for controller 5 is the measured controller flow rate corrected to 25° C, 28.89 in Hg versus scaled digital panel me- ter reading of controller output voltage (command signal). conducted, and the data are summarized in table 3 for the six controllers for air. Since the mass flow to volume flow rela- tionship depends on the physical (thermo- dynamic) properties of the gas , we per- formed the linearity characterization of controller 2 for pure carbon diox- ide, methane, and carbon monoxide. All of the data are well characterized as straight lines since the square of the coefficient of regression approaches unity; values obtained for r^ ranged TABLE 3. - Summary of the linear regression^ analysis for Tylan flow controllers Gas Controller ao ai r2 Syx Maximum flow, ml/min cv, pet Air Air Air Air Air Air CO2 CH4 CO P 1 2 3 4 5 2 2 2 5.37 .82 5.44 78.50 40.10 243.60 3.51 4.89 5.83 0.9824 1.0410 .9968 1.0100 .9382 .9505 .7277 .7315 .9973 0.9999 .9997 .9992 .9999 .9991 .9999 .9986 .9986 .9997 0.739 .069 1.840 7.120 21.120 17.350 1.770 1.690 .890 200 10 200 2,000 2,000 5,000 200 200 200 0.74 1.40 1.84 .71 2.10 .69 1.77 1.69 .89 Regression equation: measured flow = ao "•" ^^ (panel meter read- ing) and ao ai = intercept, slope. r^ = coefficient of regression. Syx = standard deviation of y on X, ^y X X CV = coefficient of variation = • 100 pet. 19 from 0.9986 to 0.9999. The values of Hq, the intercept of the regression line, and a^ , the slope of the lines, are given for the general equation: measured gas flow (in ml/min) = ao + a^ (digital panel meter reading) . Each point is the average of 10 flow mea- surements. The average deviation of the points from the expected line (Sy^), di- vided by the average flow values (one- half of maximum flow) , yields coefficient of variation (CV) data; the CV values ranged from 0.7 to 2 pet for the six controllers. BINARY MIXTURE PREPARATION We used the dynamic gas-mixing system to make different concentrations of meth- ane gas in air from a standard tank of 2.58 pet CH4 , and the results are pre- sented in figure 19. Two identical flow controllers, Nos. 3 and 4, each with a range of 40 to 2,000 ml/min, were used to generate mixtures flowing at 2,000 ml/ min. The methane concentration of each mixture was measured by an ANDROS, Inc., model 209 methane analyzer, which is claimed to have an accuracy and linearity of 2 pet of full scale (5 pet methane) , i.e., ±0.1 pet methane. The output from the analyzer was also measured by a digi- tal multimeter (Tektronix DM-501) which I 2 3 METHANE CONCENTRATION MAKEUP, pet FIGURE 19. - Measured methane concentration versus desired values. Flow controller settings were selected using figures 16 and 17 and the de- sired flow rates. has an accuracy of ±0.1 pet ±2 counts and a linearity of ±0.1 pet. To prepare the mixtures , we determined the dilution ratios and thus the flow rates of each component, air and methane. Flow rates, read from the calibration curves (figs. 16 and 17), were used to obtain the appropriate command voltage settings. The gas mixture measurements are summarized in table 4. TABLE 4. - Desired and obtained methane-air mixtures Desired Desired flow. Digital panel Methane methane ml/min meter readings , ml/min analyzer concentration, Standard Diluent Standard Diluent readings , pet CH4 gas gas gas gas pet CH4 2.58 2,000 1,830 2.58 1.29 1,000 1,000 1,035 1,015 1.30 .645 500 1,500 515 1,570 .64 .322 250 1,750 265 1,850 .34 .161 125 1,875 144 1,998 .17 2.064 1,600 400 1,695 385 2.09 20 A least squares analysis was performed (measured methane concentration, pet) on the data from figure 19. The regres- sion coefficient was 0.998, and the stan- = 1.01 (makeup methane concentration, dard deviation for the data was 0.0143 for the equation pet) . CONCLUSION The entire dynamic gas-mixing system confidence in the preparation and use of and peripheral components have proved to test gases. It is in continual use and be stable, and the dilution of gases has contributes significantly to reducing the been very predictable and reproducible. costs of operating the instrument evalu- This system has increased our level of ation laboratory. FUTURE PLANS The dynamic gas-mixing system will be temperature, humidity, gas flow (gas con- interfaced to a micro or mini computer as centration) , and time for conducting part of an overall automated testing fa- tests of gas monitors for the determina- cility for the improved evaluation of gas tion of linearity, accuracy, response detection instruments and devices. Soft- time, precision, and stability, ware programs will be used to specify 21 APPENDIX A.— SUPPLIERS' ADDRESSES AADCO, Inc., 2257 Lewis Ave., Rock- Serta Systems, Natick, MS 01760. ville, MD 20851. Stow Laboratories, Inc. Hudson, MS ANDROS Analyzers, Inc., 2332 Fourth 01749. St., Berkeley, CA 94710. Swaglok, Crawford Fitting Company, Forma Scientific, Inc. , Marietta, OH Solon, OH 44139. 45750. Teledyne, Hastings-Raydist , Hampton, VA Labcrest, Fischer Porter Co., Warmin- 23661 ster, PA 18974. Tylan Corp., 23301 South Wilmington Litronix, Inc., 19000 Homestead Rd., Ave., Carlson, CA 90745. Vallco Park, Cupertino, CA 95014. Mast Development Co., 2212 East 12th St., Davenport, lA 52803. 22 APPENDIX B.— DIGITAL TIMER OPERATION AND CONSTRUCTION The digital timer is used to manually or automatically time soap-bubble flow- meter transit times. It is constructed from readily available parts, mainly in- tegrated circuits and standard light- emitting diode (LED) displays. To sim- plify flow rate calculations, the display is formatted in decimal fractions of min- utes. The resolution is 100 ysec and 1 msec in the 10-min and lOO-min ranges , respectively. The 10-min range provides a maximum indication of 9.9999 min and is usually used for the bubble meter transit time measurements. Hour and second time bases are also available, with corre- sponding resolutions. TIME BASE GENERATION AND SELECTION All time bases are derived from a 2-MHz crystal oscillator (fig. B-1). The 2-MHz signal is divided by 2,120, and 7,200 to produce frequencies of 1 MHz (F^), 16. 66... kHz (F2), and 277. 7... Hz (F3), respectively. The desired time base is selected by SW-3 (fig. B-2). With SW-3 in the posi- tion shown, the control input to select gate 2 (pin 4) will float high, and the output (pin 6) will toggle at the F2 rate. The two inputs to both invert gates are pulled high, which forces a low at the outputs and inhibits select gates 1 and 3. Placing SW-3 in the "Sec" position in- hibits select gate 2 (through the cou- pling diode) and places a low (ground) on the inputs (pins 2 and 3) of the upper invert gate in figure B-1. The high out- put from this inverter enables select gate 1, and allows frequency F^ to appear at the output (pin 3) . A similar action occurs when SW-3 is in the "Hr" position. In this case, select gates 1 and 2 are inhibited; select gate 3 is enabled and passes F3 to its output. The selected time base (F^ , F2 , or F3) is buffered and passed on to the input of a decade counter and to the range switch, SW-4B. The output from the counter (F^/ 10) is connected to the opposite side of the range switch. The output from the time base generation and selection sec- tion is Fn or Fp/lO, depending on the po- sition of the range switch. Switch SW-4A sets the decimal point to display one (range = 10) or two (range = 100) inte- gers in the five-digit display. The selected time base is routed through one of two run gates in the con- trol and gating logic section (fig. B-3) to the counters and displays section (fig. B-4) . In this section the time base is divided by 100 and then fed sequentially to five display counters . The BCD outputs from the counters are de- coded by the 7447 drivers (fig. B-4) into seven drive lines to the LED readouts (RO). TIME BASE CONTROL The starting and stopping of the se- lected time base frequencies to the coun- ters is controlled in two ways: One uses a pair of phototransistors, one to start, the other to stop the counter; The other uses a single switch to start, stop, and reset the timer. The circuit for the time base control is shown in figure B-3. The double-pole switch, SW-2A and SW-2B, is used to select either of these two ways: In the X position shown, the pho- totransistors control the time base to the counters and SW-1 is used to reset the counters in preparation for another measurement; in the Y position, the switch SW-1 performs the start, stop, and reset operations in succession, thus con- trolling the timer as a manually operated stopwatch. Each of these time-base- controlling methods is described in de- tail below. Switch SW-1, connected to J-3 (fig. B-3) , is used to reset the timer when switch SW-2 is set to the X position. When momentary contact switch SW-1 is closed, a low is coupled through the diode to pin 3 of the one-shot multivi- brator OS-2. This triggers the one-shot multivibrator and produces two 500-psec 23 -J-^ ^ ^ ■J-2> ■J-l> ^ ^ Control and gating logic sw-3 xo 2 MHz - -rZ 'J Sec A Hr ? Min -^60 -*• 4-60 - SW-4a 10 ^100 jL Decimal point select SW-4b ) '^ m Counters and displays J 10 100 tIO — ' LEGEND F, I MHz Fg 16.66... kHz Fj 277.7... Hz \^ Time base generation and selection J FIGURE B-1. - Functional block diagram of the digital timer. 7400- 7400- Ik; 10 bt used .8 'Select gate I SW-3iSec i > T'° ^ i ?Hr 1 Invert 470 • i — o — 'W^ • point To v+ To select 100 10 i5iSW-4B To run — gates I and 2 ^ 2,3,6,7,10 -Invert 7-45; -/ jV* 240 XTAL CSC 2 MHz 2N706 '"^.'^ 14 -k,. *F| 7492 -r2 7492 -r6 8 14 5 r2 v+ 5 7490 -i-IO I a V+ ♦'^3 14 .JM 10 7490 10 a X 10 1^7490 -^6 FIGURE B-2. = Time base generation and selection schematic. 24 IOk< >0.068/iF >22k Y' 220 SW-2^ o — w^ ilL'Oio S/S/R or reset ' t SW OS- 1 500^5 / Reset gate y-^ 10 kJ 0.068/zF — ^h lOOuF He-- FF-fl ^ 7474 FF-B '2 7474 14 OS- 2 J 500^5 L 74121 SW-2b Reset - _ A lO/iF 22 k 300 k: Start detector o- input Stop 300 k; detector <> ' input Threshold director ^ i Start gate 7 13 /_ Threshold director <^ Start Stop FF r Reset counters to zero un gate Gated timing pulses to counters Not usedx Run gate Selected time base ^Stop gate FIGURE B-3. - Control and gating logic schematic. To FF pin 10 Fpor Fp/IO From<^ ilZi I 5 V+ 7490 in joT > Reset v+ 7490 D P 14 II lEA 7490 Reset Y- 7447 Reset oV+ £Jl 7490 7447 14 II 7447 Reset oV+ 5 4 lEIR 7490 14 h iLji 2_g Reset , 5 10112 -0V+ ^ 16 7 7490 7447 5 4 II "Reset 7490 oV+ A B C D 7447 I9I3I2III0 9 1 Reset MSD-I RO Litronix 707 MSD-2 D.P RO From<|> From<|> D.R RO RO 7 150 a 7 for each digit RO Litronix 707 3,9,14 •v,+ FIGURE B-4. - Counters and display schematic. 25 pulses — a negative going pulse (pin 1), and a simultaneous positive going pulse (pin 6). The negative pulse goes to pins 10 and 11 of the stop gate (Start Stop FF) and resets the flip-flop (FF) , which insures that the run gate (GA-2) is in- hibited. The positive pulse (from pin 6 of OS-2) goes to the seven decade coun- ters in the counters and displays section and resets the displays to zero. When SW-2 is in the Y position, SW-1 operates as a start, stop, and reset (S/ S/R) control. (See fig. B-7 for the tim- ing diagram and fig. B-3 for the electri- cal schematic.) To explain this opera- tion, first assume that the timer has just been reset. The A and B outputs from flip-flop A and flip-flop B (FF-A and FF-B) will be low. The other out- puts, pin 6 and B, will be high. The low-level signals at the input (pins 1 , 2, and 13) of the reset gate produce a high signal at the output (pin 12) . This high is blocked by the diode, and the signal has no effect on the reset one shot (OS-2). The low (A) signal at pin 3 of the run gate (GA-1) inhibits this gate and does not allow the selected time base signal (pin 5) to pass through, (The other run gate, GA-2, is inhibited by SW-2B . ) The first closure of SW-1 (start) trig- gers the one shot OS-1, which causes flip-flop A (FF-A) to change states at the end of the 500-ysec pulse — the A out- put goes high and pin 6 goes low. The signal from pin 6 has no effect on FF-B, (The flip-flops are triggered by a low to high transition.) The output (high) from the reset gate is not affected because pin 2 is held low by the B signal. The run gate, however, is enabled because both the A and B signals are high. This allows the selected time base signal to pass through to the counters and displays. Pressing SW-1 again causes the one shot OS-1 to produce another pulse. Again, FF-A changes states. The A output goes low, and pin 6 goes high. The low-to-high transition at pin 11 causes FF-B to change states — B goes high, and B goes low. The output from the reset gate remains high because A (pins 1 and 13) goes low before B (pin 2) goes high. The run gate is inhibited because both A and B are low. This stops the passage of the selected time base to the counters (and display) , and the displays will display the accumulated number of se- lected time base pulses (or the elapsed time) , The third closure of SW-1 causes anoth- er pulse from OS-1 to trigger flip-flop A, which changes states again. Both in- puts (pins 1, 2, and 13) to the reset gate are high, which causes a low at the output. This low, coupled through the diode, triggers the one shot OS-2. The 500-ysec positive pulse from pin 6 resets the counters and display to zero. The negative pulse (pin 1) is coupled through the diode to pins 1 and 13 of the flip-flops and resets them to the condi- tions described at the beginning of this discussion. Run gate 2 (GA-2) in the control and gating logic section (figs. B-1 and B-3) is controlled by two gates which are cross-coupled to form a flip-flop (FF). The state of this flip-flop is controlled by two phototransistors of the soap- bubble detector assembly on the flowmeter through threshold detectors (fig, B-5) , in response to changes in the intensity of light falling on the transistors, A negative transition (pulse) appearing at pin 13 of the flip-flop sets the flip- flop producing a high level at the output — enabling the run gate (GA-2). A negative pulse at pin 9 resets the flip- flop and inhibits (disables) GA-2. This gates the time base signals through GA-2 only during the transit time of the bub- ble meter bubble. The diagram of the S/ S/R sequence is given in figure B-7; fig- ures B-5 and B-6 are diagrams of the de- tector amplifier and of the power supply, respectively. 26 +2.5V S300k lOk V\Ar 10 M VvAr M Start 2k'' 10 k LED' . ^AA/ ^ 100 + 2.5V To start gate ■ GA-1,3 + 2.5V 300 k lOk i< vAA/ ; ^ :^J-2 9V+ lOM 13 Stop I 2k lOk 14 LED' IM 10 •100 Ik rVW-WV '' To stop gate ■^ GA-I9 LEGEND +2.5V A,_4- LM324 FIGURE B-5. - Soap bubble transit detection circuitry. 4 Vdc ^ v + display V l '' -J Q- H E n LjJ O) 1 — 1 o. o 1— s: 1 LU 0. 1- 1— J — . C -,-—.2 Er- O --'£.— -—Li- eu E •^ ■M o s: ^— •.- > lO to •^ C (U u +-> (0 -Q r- •r- S- 3 ro C 1— 1— o l-H < m cj ^ "^ o> (U c -— N 4-> •f— > o^-s;"^ 3 -l-> LLI o 0) Y 3: 3 ^— 00 Q O) = o LU 5 LU 3 Ll_ Q. o 7 4J QJ 1— fo s- -a 00 Lj_ 1 S- 3 ^ 00 OO 0-C.jrO^O0) s XT) lO t^ to Oi ^ ^ en UJ . • S- T3 H S- S- -r- c 3 O O O Q OO 1— III LJL 1 III iillil 0a@0§s0s IBIiiilS IBIIIESiS iilBSiil ;i ■"! ~ ~ Z; ~ ~ Zj B00Q00QS G 3 n n a Z: D =i iiiiSiiS □ n~~~zj~~ c o ^« u 3 ^. ^* _C 0) 0) o> c o X C4 c ® ofessio ogrom 0. a^y * ''EIAS iNSTtJ I 1 «u 1 1 1 > < CM tn ro o ■v^ CM ^ o o tn Lf) CM _l cn ■— «* o 1— en en o 1 — to ^ «D n * • • • • • CO CO LO cy> o ^ cx) O O VD CD CO CO OO CTl CM LO CM CM O lO CTl CM «d- Q CM (Tl cri O C CM O CO en — ~ _ — — Ol UJ cC CQ O LU LU ea: CQ O ea: 00 Q DC CL CO 00 CO Ln OC ^ cr> IT) «;!- III • • * • 1- z UJ 1 CM CD O 1 1 O O 1 to CO 1 CM CM IT) O CM CM cn o ai O) CL. O) E N S- -^ 2 C7) cu •.- 1— 3 -P O C 1— OO cu 4-> O) +-> r— •r- xi s- OO (0 S- OO +-> • 3 J3 S +-> +-> 0) to (O 1— — +J 2: • . r— S- Q- to -•-> S- cc 3 s 1— Q. O ■— O Q. ^ OO cu M- E cu m J2 o 1 — O n3 Q Q. M- -!-> O cu S- Di •»— en cnLi- S- n3 O O cu 1— Q^ n— D U- C C +->(/) 2:01 3 (O UJ -o •1- •.- -D E fO c cms cu c/l +J +J +J f ■) 5- - 4-> +-> OJ O •!- S- 3cnajcU(T3EEc— O CC n3 ■o fores-ocuEEra OfOE-— q:cucuu_ E t 3 -«-> E > 3 +-> S- S- +-> cu ro oooQfO>cura(a CUcC+JCQ rtJQ-Q. rtJ to '-- s- (U -1- 1— -o s- s- cu fsl 'r— r^ r-~ 3 x: m N "O l_S-S-3to JDcnisi •.-S-t0S-3S-S-3 Q. o •1- c: Ol - O -r- ■OCUECUOCUCUU S- (O ^ +->+->+Ji — -E-f->i — S-1 — S_-l-> fO+Ji — 4->+-'i — S- D- fO = CECro+-> rCQ-fO fOES-EroEEfO Li_ C7) LU LU LU C_) (U U •>- -OLUh- LUOUJLUO ^ — " ■o +-> -o 0) 4-> T3 +-> E fO •1- c: . • . . 00 ro E n3 -^ fO • O a •>- ctCQOQrDS-OOC +-> d; CQ (J Q UJ u_ _l )— 1 U- ^^ _1 1— . 00 Q. UJ ■a; UJ C\J CO o _l Q 1— CQ 1— _J _j oo rc UU 1—1— CM CM * CM 1— O r- O _jxcoLnoo cMOOi— cQ_cr.r-(— 1— oocTif— 1— _i _ILl_CM 'OOOOCM -oot/i^: LU Q CM o 00 un 1 — c — CM CM CM o o o o o o o o o C\J o r— O «D CM I— =d- >^ 00 ocMcrii— 1 — oooc^c — o r— ooooooooo lOI — CMCOCMCMCMCOCMCMi — >*Lr)OC3^<*-*ocn<*<5i-oo o o _l CM ^ r^ o o o ooocMLDOOi— "* Calculate Enter m O o o) s C 1— n3 LU u. q:: S- I/) OJ E (O E C S- -r- UJ 1— 1— Flow Rate S- E O) CU 13 +-> J3 1— C 3 o llJ 1— > Temperature >- LU (_) o * O O IT) (^ CO * o o o CQ LO * o ■< O _I o 1— DO 1— —1 ^ oo nz * _l O 1— CQ 1— _I _1 00 D= _l _1 _l O 1— 03 O O 1— —1 _i Qi •!• Qc II oo n: * —1 O 1— CQ 1— _J _j (/I :r * _l CO LD O 1— CQ r^ r- 1— _j _l + CM • II 00 3Z * LU Q O O Lx^ ^o CO o o o 1^ o CO un 1 — o o o o ?o o — o i£> o •— o CX) o o o .— o o o r^ o r- o 1— r^ I— o r— o o o ^ CM 1— vT -^ 00 ^D CM 1— "^^ -^ CO <£) CO Ln CO Ln CM 1 — >v^ «:d- iri -^ CTi -j^ CO lO CM 1— >* -^ 00 lO ID CM CO IT) CM 1 — ^ CX5 O CTi CTi >* CX3 o O _j X> CO .— ^ Ln ijD o o o CM >* r-^ l£> VO VO o o o 00 O CO «^ 1^ 00 1— UD 1^ 1^ 1^ r^ r^ 00 o o o o o o o CM <;i- 1 — 00 00 CO o o o 00 O r— «* I^ 00 r- CO C3~i CTi C3~i C3~i C3^ O O O O O O O .— Indicated Flow Rate standard ize Flow Rate 2 LU (Operate) (Calib rate) 2 O o > LU CO 1 — CM CM CO «d- LiJ O 1 — 1 — o O O Q 1— o CM CM Lf) ^O O 1— O o o _I_J _J _l _l —1 Ol— CQO O C_) O O t— _l _i q: •!• q; X C3i •!• q; X cc: ii t/o n: _J CQ _J ■k _l _J o , o cc: •!• q; X -" -1 -■ H O C_J o «-> q; •:• Qi X o; •I'cr: ii O 1— 1— _1 OO zz LU D o O ,— ,— CM CM CO <;:f O O O O O O o o CM CM LO lO o o o o o LT) O r- ■— O O O r— o o o o o o ^'• r- O o o o .— o o o o o o o o LCCOLOcoLnroLnrouncoLDCMi — >:a- ^ i£) OOLOCOLnCOLnCOLDCMi — td-Ln^d-iiD'^Ln^o^oo o -J OcvjLntDCTiOco«d-r^oOi — CMun OOCDOOi — 1 — 1 — 1— 1 — CMCMCM ooooooooooooo CM o CO 1 — CM LD CM CO CO 00 o o o o *jDcnoco>^r^cOi — CMtn coco<*<^"^"^«^i-nLr)Ln oooooooooo FIGURE C-2. - Program listing for gas dilution mixing program. INT.-BU.OF MINES, PGH., PA. 2690 1 9263 'it.. •' nV o I. . . "^ A> '^'v -- «<^ ... •*i. •" 0-..--..V /\.-'i;..\ c»*.i^^;i:.% >*.. ' .^O^ r. -^^0^ ^\ **c,0*''.o»JL''. "^o i,^t. r7«' A >*...L:i:j* <^ *.Vo«* ,^'«' 4>"^^. ■0,. *•..-•* aP <»,. *'T7r«' A ** .t'»- '^J*^ n^ .o"». " .i.'*^ \Jiu ^^..^^ /i«^"o %..^*^ /jfe- %,^^ ,*i^jS«^^o V..^ Ao, r/..:.^j^%v o«*.^^^'\> /\c;z^/^'^^ 0° •:;^sw °- /\»-^'" "-^c,^" ♦ ^1^* V"^^ ,.** "**- ^U rt^ • • • , «^i. «> V':^'\/" V*"^'^'*^^'^ V''^^*^^*'^ "^ v^ 4O vT* ;^ < ^; WERT BOOKBINDING MIDOLETOWW, PA DEC. 83 "^J^^^'A j"^-^ ^y ^ LIBRARY OF CONGRESS llllillniii'iiiriii'riii<'i'if"iii'i!iMi'ii 002 959 883 6 ^m . k <* 1?-