Journal of The Society of Instrument and Control Engineers
Online ISSN : 1883-8170
Print ISSN : 0453-4662
ISSN-L : 0453-4662
Volume 3, Issue 4
Displaying 1-8 of 8 articles from this issue
  • [in Japanese]
    1964 Volume 3 Issue 4 Pages 243-247
    Published: April 10, 1964
    Released on J-STAGE: October 21, 2010
    JOURNAL FREE ACCESS
    Download PDF (815K)
  • Eisuke MASADA
    1964 Volume 3 Issue 4 Pages 248-254
    Published: April 10, 1964
    Released on J-STAGE: October 21, 2010
    JOURNAL FREE ACCESS
    In digital control system driven by relay element, dead time existing in process dynamic characteristics seriously affects the stability of the system. In this paper, Equation (6) showing the relation between loop gain and dead time is obtained to make stable the secondary system shown in Fig. 3 by phase plane approach. In order to keep the system stable, however, loop gain must be considerably small and from this reason, the response will be much delayed. In digital control systems, it is desirable to compensate this dead time by logical operations only. Therefore, a simple nonlinear compensator, which is mainly composed of comparators shown in Fig. 14, is introduced by measuring the incremental pulse duration. An illustrative example shows that this compensator considerably shortens the response time to the step input.
    Download PDF (711K)
  • Masami ITO
    1964 Volume 3 Issue 4 Pages 255-261
    Published: April 10, 1964
    Released on J-STAGE: October 21, 2010
    JOURNAL FREE ACCESS
    In this paper are considered relations among the following three delay systems with gain compensation, feedback compensation or series lead compensation. They are the zero non-regular control system whose open loop transfer function has zeros in the right half S-plane (Z-system), the control system with dead time (D-system), and the control system with S-shaped lag (S-system). In these cases the zero non-regular delay, dead time or S-shaped lag is assumed to be added to the fundamental second order linear control loop. Conclusions obtained as follows:
    1) The Stable domain of S-system is wider than that of D-system and the same domain of D-system is wider than that of Z-system.
    2) The stability limits of D-system are approximate to those of Z-system in all cases of the three compensation methodes, so that in the design of control system with dead time, the stability limits of Z-system serves as approximate stability limits of D-system.
    3) If the delay time is small, dynamic properties of the three dalay systems with gain compensation or series lead compensation are tolerably similar each other. But dynamic properties of the three delay systems with feedback compensation are not always similar each other, even if the delay time is small.
    Download PDF (661K)
  • Kazuo NAKANO
    1964 Volume 3 Issue 4 Pages 262-271
    Published: April 10, 1964
    Released on J-STAGE: October 21, 2010
    JOURNAL FREE ACCESS
    The technics of adopting a dither to eliminate the dead zone or hysteresis caused by the static friction have often been used for a long time in measuring or indicating instruments. In spite of this fact, there has been no reasonable standard to select the optimal dither for electro-hydraulic servovalves up to now. In this paper, therefore, the nonlinear second order system with static and Coulomb friction was considered as a mechanical model to represent the mechanical characteristics (excluding the electromagnetic characteris tics) of the servovalves and the effects given to the system by the dither were analyzed. In consequence, several important points in selection of the dither were found by this analysis and they were proposed as a standard to give an adequate dither to the system for eliminating the dead zone or hysteresis due to the static friction. The effects of the dither exerted upon the electromagnetic and dynamic characteristics were not touched in this paper.
    Download PDF (995K)
  • Iwao YAMAMOTO
    1964 Volume 3 Issue 4 Pages 272-281
    Published: April 10, 1964
    Released on J-STAGE: October 21, 2010
    JOURNAL FREE ACCESS
    A pneumatic micrometer single circuit with negative gauge pressure is a type of pneumatic micrometer consisting of a measuring nozzle and a constant orifice and operates with the pressure between atmospheric pressure and vacuum. Although this type of pneumatic micrometer has such an advantage that it does not need pressure regulator and accordingly circuit mechanism becomes simple, it has many nonlinear parts in the transfer characteristics and especially dynamic characteristic is poor. Described in this paper is a method for improvement of the pneumatic single circuit characteristics. Trially manufactured apparatus is composed of two single pneumatic amplifiers, one, having a jet nozzle held by a spring-bellows negative feedback transducer, serves as a preamplifier for output circuit and the other, connected in cascade to the former through a pressure displacement transducer, serves as an output amplifier. The transfer amplification factors are given as K*/S when K* is a spring constant of the spring-bellows transducer for feedback and S is a bellows' effective area of pressure, and they are both constant regardless of change of atmospheric pressure. Experiments conducted by using spring assemblies, each of which is composed of four compression springs, proved that more than 95% range of full scale is linearized with an accuracy of less than 1%. Response time has also been improved compared with the single pneumatic micrometer which has a same output circuit as that of this cascade micrometer. Also discussed are dynamic characte ristic and the effects upon changing the circuit component.
    Download PDF (2928K)
  • Iwao MORISHITA
    1964 Volume 3 Issue 4 Pages 282-288
    Published: April 10, 1964
    Released on J-STAGE: October 21, 2010
    JOURNAL FREE ACCESS
    This paper describes a new method for calculating correlation functions and its simulation studies on a digital computer. In this method the exact autocorrelation function of a zero-mean gaussian random signal x (t) is obtained by calculating only the mean value of sampled data x (ti+τ), i=1, 2, …, where ti is given as the time instants when the signal x (t) crosses over an arbitrarily preset non-zero levelξ. Therefore a correlator can be made up of an amplitude descriminator, a pulse signal delay, a sampler and an integrator. No multiplier is necessary. This method is also applied to hand-calculation from a data recorded on a chart. Discussion on an accuracy of the results calculated from a finite time length data proves that the optimum value of ξ is √2σ, where σ2 is variance of x (t). This is verified by some simulation studies on a digital computer.
    Download PDF (801K)
  • [in Japanese]
    1964 Volume 3 Issue 4 Pages 289-293
    Published: April 10, 1964
    Released on J-STAGE: October 21, 2010
    JOURNAL FREE ACCESS
    Download PDF (580K)
  • [in Japanese]
    1964 Volume 3 Issue 4 Pages 294-300
    Published: April 10, 1964
    Released on J-STAGE: October 21, 2010
    JOURNAL FREE ACCESS
    Download PDF (809K)
feedback
Top