1、 Proposal for a gas flow meter at LHCGas Working GroupThe place where gas systems forhigh energy experiments become realityF.Hahn, S.Haider , M. v.d. Klis , C. ZinoniAbstractA gas channel-flow-meter for all LHC detectors is presented. The technology described resembles a hot-wire meter measuring mas
2、s flow. This report discusses the electronic circuitry, the measurement procedure, the best set of parameters and a technical solution with industrial available components. CERN Gas Working Group, EP/TA110.2.2001IndexIntroduction.0Flow cell.0Measurement principle .0Results.0Measurement .0Final set o
3、f parameters.0First prototype .0Electronics layout.0Electronic components .0References .02Gaseous detectors at LHC will need adequate flow measurements at inlet and outlet of chambers. These devices should be precise, small in dimensions, have a low pressure drop, be able to operate in magnetic fiel
4、ds up to 0.1 Tesla and most of all, they should be cheap because of the large numbers needed. Industrial flow measurement devices do violate at least one of these requirements. It was therefore necessary to develop a meter suitable for all detectors, scalable to all flow ranges and all gas types. Th
5、is report describes the development of a flow meter that resembles a Hot-Wire Mass Flow meter1: a Pt100 temperature sensor is used for both heating and measuring temperature. A relatively high current (30mA) is passed through the sensor for some seconds: the resulting rise of temperature is a good m
6、easure of the gas flow that cools the Pt100. We shall discuss the parameters that were optimised on repeatability and measurement accuracy as well as the results of the gas flow measurement capability for different gasses. In the last part we shall describe a possible technical realisation with indu
7、strial, modular electronics component controlled by a PLC. The flow cell consists of an aluminium bar with a central bore parallel to its long side. One end is permanently closed with a plug. Perpendicular to this bore, two holes complete a gas passage with in- and outlet. Gas tightness is achieved
8、with o-rings in grooves in the supporting distribution manifold. The flow cell is held in place by four screws. Figure 1 shows a cross section of the flow cell. The Pt100 sensor is mounted in the middle of the centre bore with the sensor head either parallel or perpendicular to the gas flow. In the
9、first prototypes, the sensor was adjustable in depth, which was observed to have a large influence in the readings. For the final flow cells, the sensors will be glued in position with the help of an assembly jig. The sensor will be equipped with a cable and connector. 3 Figure 1 Flow cell mounted o
10、n gas distribution manifoldA hot object surrounded by gas will heat up the gas. When the gas is in motion, heat will be transported away: the object cooled. Since this process depends on the number of molecules in contact with the hot surface, the cooling effect will be proportional to the mass of t
11、he gas that passed by. A larger gas flow would cool the object more, or stated differently: a constantly heated object would take longer to reach a certain temperature had it been subjected to a large gas flow. As a result of a series of tests 2 it was found, that the best quantity to observe was th
12、e Rate of temperature change after the sensor had heated up by T (i.e. 30 oC ) from ambient. Figure 2 shows the measurement principle: ambient temperature needs to be measured first since the final measurement depends on it. A small measurement current ( 1mA ) is passed through the pt100. With the c
13、orresponding measured voltage one determines the resistance and thus the temperature.1 Next, the heating current is switched on, i.e. 30 mA, which immediately leads to a voltage increase over the sensor. The voltage at which the heating current is turned off, calculates as follows:Tend = Tamb + T an
14、d therefore Uend = Rend * Iheating = ( Ramb + 0.385 * T ) * Iheatingwith Ramb = Uamb / Imeasure and Imeasure = 1 mAIn order to obtain the rate of temperature change just before switch-off, all voltage readings within the last 5 mV are taken for a line fit: the resulting slope is the measurement resu
15、lt. 1 T amb = ( R-100 ) / 0.385 /KoC4 Figure 2 Measurement principle: 1st; measure ambient temperature, 2nd; switch on heating current, 3rd; wait until temperature has risen by T and switch off heating current, 4th; take all points within the last 5 mV and fit line to them, 5th; wait until cooled of
16、f for next cycle. MeasurementIn Figure 3 one can see a typical temperature increase for different gas flows for uninterrupted heating of the Pt100. After about 40 to 50 seconds an asymptotic temperature is reached. Using this temperature as a parameter for the gas flow measurement gave less stabilit
17、y and an unpractical long heating time. Figure 4.a shows the obtained rate of temperature increase versus a calibrated reference flow of Argon. The latter one was varied from 0 to 30 litres per hour. The error bars represent the standard deviation of 20 consecutive measurements at a given flow. The
18、data points were fit with a quadratic polynomial. In Figure 4.b one can see the residuals of the individual measurements to this fit. The gaussian fit shows a standard deviation of 0.68 l/h, which means on a 30 l/h maximum flow, an average error of about 2.2%. Other methods have shown larger uncerta
19、inty 2.Figure 5 shows the result from the Pt100 flow measurement for three different gasses versus the calibrated mass flow of the reference device. 5 Figure 3 Temperature increase for different gas flows Figure 4 a) Slope of heating curve versus calibrated Argon flow b) residual to a second order p
20、olynomial fita b6 Figure 5 Pt100 mass flow measurements for different gassesFinal set of parametersSeveral parameters were varied in order to see their influence on stability of the flow measurement: The orientation of the sensor toward the flow (parallel or perpendicular) showed about 1 to 2 % infl
21、uence The sensor should be placed accurately in the centre of the tube; an influence up to about 3% of the reading was observed in case of misplacement. On a series of 10 Pt100s a sensor to sensor variation of about 5% of the flow reading was observed. Pressurisation of a sensor up to 200 mbar above
22、 atmosphere did not show any significant influence on the flow reading. Measurement current: 1 mA, heating current: 30mA, and T of 30 K have shown the best reproducability for Argon 30 l/h.The combination of these influences makes it necessary to calibrate each sensor individually. In order to cover
23、 the calibration of the sensor position within the flow cell, a calibration facility is foreseen where many flow cells can be connected in series and fed with a calibrated gas flow. Varying this flow and recording at the same time the readings from all Pt100s, many cells can be calibrated at once. E
24、lectronics layoutA simple electronic scheme is shown in Figure 6. A programmable current source provides the measurement current for ambient temperature (1mA) as well as the heating current (30mA). A multiplexer/selector circuit connects the current source to the Pt100 of interest. An ADC (Analog to
25、 Digital Converter) measures the voltage over the selected Pt100 (compare Figure 2 for measurement principle).Once the Pt100 is selected, a PLC (Programmable Logic Controller) performs a measurement cycle and goes on to the next sensor. At the end of one total cycle over all sensors, the PLC would s
26、elect a normal, very accurate, 100 resistor and applies the heating current of i.e. 30mA. The resulting 3V can be read with the ADC and thus verified that the current source delivers a stable current. This online check is important in view of the long time stability of electronic components in non-a
27、ccessible areas like the UX caverns.7 Figure 6 Electronic layoutElectronic componentsFor the prototype, industrially available, distributed I/O (Input/Output) modules from either firms Beckhoff or Wago were chosen. They offer large flexibility and the possibility to control everything from a PLC. Fi
28、gure 7 shows the assembly of the main components: a PLC that performs the measurement cycle and also communicates the flow results to the outside world via a ProfiBUS network. Further modules are: ADC 16 bit 0-10 V, two DAC current sources with each 0-20 mA output (put in parallel for 0-40 mA curren
29、t source), digital relays (closing contacts) for selecting the Pt100 channel. Figure 7 Industrial I/O components with PLCPLCADC DAC 0-40mARelay modules 81 Private discussions with M.Bosteels and S.Berry, March 20002 Report of M. van der Klis “PT100 channel flow meter for the LHC experiments“, report February 2001-02-07
