AUSTIN! NOTES ON THE AUDION 487 



RADIOTELEGRAPHY.— Notes on the audion. L. W. Austin, 

 IT. S. Naval Radiotelegraphic Laboratory. 



A number of observations on the deForest Hudson filament 

 audion have been made at the U. S. Naval Radiotelegraphic Labo- 

 ratory, which may be of general interest. 



The gas pressure used in audion detectors is generally below 

 0.001 mm. of mercury. By substituting nitrogen for air, to 

 prevent the burning out of the filament, it has been found 

 possible to construct detectors at all pressures up to that of the 

 atmosphere. The action at 3 mm. is entirely normal. Local 

 oscillations are easily produced and the sensitiveness is fully as 

 great, both for continuous and damped signals, as at the usual 

 pressure. At 10 mm. the sensitiveness is about normal, but 

 local oscillations are more difficult to produce. In the neigh- 

 borhood of atmospheric pressure no local oscillations have been 

 observed, and the sensitiveness to spark signals is much less 

 than at the low pressures. The conditions in this case would 

 undoubtedly be much improved by bringing the electrodes 

 closer together. Even with the ordinary arrangement of elec- 

 trodes, the changes in the grid and plate currents due to the 

 incoming waves are similar to those observed in the usual vacuum. 

 With 200 volts, the plate current amounts to 20 or 30 micro- 

 amperes. 



Another series of observations has been made on the effect of the 

 direct current voltage between the grid and filament on grid and 

 plate signals. By grid and plate signals are meant the changes 

 in grid and plate current brought about by an impressed alter- 

 nating e.m.f. In the case of radio frequency signals a tuned 

 closed circuit is connected in the usual way to the filament and 

 through a stopping condenser to the grid. In this case the 

 stopping condenser is shunted by a circuit containing a potenti- 

 ometer for varying the D. C. grid potential and a galvanometer 

 for measuring the grid current. In the case of low frequency 

 excitation no tuned circuit is used but the filament is connected 

 to the grid through the direct current potentiometer already 

 mentioned, the galvanometer, and a slide wire alternating cur- 

 rent potentiometer from which the required alternating voltage 



