Sec. 1-6] MECHANICAL INPUT TRANSDUCERS 131 



fluctuation and cancels out temperature-fluctuation effects. Direc- 

 tional-sensitive hot-wire probes can also be made by the use of two 

 wires parallel to each other and spaced at a distance of about 0.5 

 mm. 1 



When measuring at very low air speeds, the hot-wire probe must be 

 used in the position in which it was calibrated, since the self-induced 

 air current varies with the probe position. 



The transducer is apparently not (or very little) influenced by 

 humidity and, at least at steady flow and low velocities, by the pres- 

 sure of the medium. But the temperature of the conduit wall can be 

 of disturbing influence if the hot wire is in close proximity to the wall 

 (order of 2 mm). This influence is reduced at high gas velocities. 

 Errors are also caused by the deposit of dirt or dust upon the wire, by 

 mechanical stress in the wire, and by vibration of the suspension 

 wires under the influence of the gas-stream impact. Further errors 

 can be caused by a variation of the composition of the gas 

 mixture. 



The finer the wire, the faster it will follow a fast variation in flow 

 velocity, but the easier it will break or will be altered by mechanical 

 stress. Platinum Wollaston wires of a thickness of the order of 1 ju 

 respond in general to gas- velocity fluctuations in the frequency range 

 of 1,000 to 10,000 cps (in special cases up to 100,000 cps). The 

 response of a hot-wire probe to alternating air currents has been 

 studied by Richards 2 and Maxwell 3 by mounting the hot-wire trans- 

 ducers on a tuning fork. The time response of a hot-wire system can 

 also be measured, according to Ziegler, 4 by connecting the hot-wire 

 transducer in a bridge and applying a direct current and a super- 

 imposed alternating current or square wave to the bridge input. The 

 instantaneous output of the bridge is proportional to the resistance 

 fluctuation of the wire and the resultant signal can be vised for the 

 determination of the frequency response. 



Hot-wire probes are usually made from platinum, platinum 80- 

 irridium 20 alloy, tungsten, or nickel. The length of the wires varies 

 from several centimeters to a fraction of a millimeter; the diameter 

 varies from 0.1 mm down to the micron range (0.00005 in.). 



Heated thermoelements can be used instead of heated wires, but 

 their time constant is considerably larger than that of a single hot 

 wire, and their output depends strongly upon heating-current 



1 Thomas, loc. cit. 



2 R. C. Richards, Phil. Mag., (6) 45, 926 (1923). 



3 R. S. Maxwell, Phil. Mag., (7) 6, 945 (1928). 



4 M. Ziegler, Proc. Koninkl. Ned. Akad. Wetenschap., 34, 663 (1931). 



