Page 63 - Analog Circuit Design Art, Science, and Personalities
P. 63

Max Wien, Mr. Hewlett, and a Rainy Sunday Afternoon




               Figure 7-1.
               One of the
          original Hewlett-
            Packard Model
           200A oscillators
           -the  good guys
           won and nobody
              lost. (Photo
              courtesy of
          Hewlett-Pa cka rd
               Company.)










                         (DeForest hadn’t even dreamed of adding a third element to Edison’s Effect in
                         1891), so he couldn’t readily get anything to oscillate. Anyway, Wien was preoccu-
                         pied with other problems and developed the network for AC bridge measurements.
                           Hewlett saw that Wien’s network, combined with suitably controlled electronic
                         gain, offered significant potential improvements over approaches then used to make
                         oscillators. These included dynamic tuning range, amplitude and frequency sta-
                         bility, low distortion, and simplicity.
                           Hewlett had something else besides electronic gain available; he also had the
                         new tools of feedback theory. Harold S. Black’s pioneering work, “Stabilized Feed-
                         back Amplifier,” appears as the fourth reference in the thesis bibliography. Simi-
                         larly, Nyquist’s “Regeneration Theory,” a classic describing necessary conditions
                         for oscillation, is reference number three.
                           Hewlett synthesized all this nicely to show that Wien’s network could be made
                         to oscillate. Then he added a single (quite literally) crucial element. The oscillator’s
                         gain must be carefully controlled to support stable sinusoidal oscillation. If gain is
                         too low, oscillation will not occur. Conversely, excessive gain forces limit cycles,
                         creating a square wave oscillator. The problem is to introduce an amplitude regu-
                         lation mechanism that does not generate output waveform distortion. Hewlett
                         describes the elegant solution:


                               The  last  requirement,  an  amplitude-limiting
                             device  that  will not  introduce distortion,  is more
                             difficult  to achieve.  It is well  known  that  the  gain
                             of  an  amplifier  with  negative  feedback  is 1/R,  providing
                             AD  is large  compared  to 1.  Thus  if  a  resistance  whose
                             value  increases  with  the  current  through  it  is used  as
                             past  of  the negative  feedback  network,  the gain  of  the
                             amplifier  may  be  made  to decrease  with  an  increase  in
                             the  input  voltage.  If  an  amplifier  of  this type  is
                             used  as part  of  the  oscillator,  it can  be  adjusted  so
                             that  oscillations will just  start.  As  oscillations


         44
   58   59   60   61   62   63   64   65   66   67   68