Sensitivity, selectivity and audio fidelity data for some TRF receivers produced in 1928 and 1929

By Ben H. Tongue

This Article presents two sets of data on the performance of some AM-band TRF radios manufactured in 1928 and 1929.  The first data set (A) is in a report of measurements made on forty receivers by Radio Frequency Labs (RFL).  The second data set (B) comprises 57 graphs of sensitivity, selectivity and audio fidelity of various receivers.  I picked these data up at a house sale years ago for 10¢.  The (B) tests were made in Camden, New Jersey and East Pittsburgh, PA, as noted in some of the title boxes in the upper right of the pages. This leads one to think that they might have been measurements made by RCA and Westinghouse of their competitor's TRF receivers.  This thought is further strengthened by the fact that no RCA or Westinghouse receivers are shown.

The two sets of data were taken using somewhat different standards:

Data set (A):  RFL's sensitivity measurements were made (see Document 1) using somewhat different dummy antenna and audio power output standards than those described in "Standard Tests of Broadcast Radio Receivers", published in the1929 IRE Yearbook (see Document 3).  RFL's sensitivity figures represent the minimum RMS value of the carrier voltage of a 30% AM modulated test signal voltage (Vt) that causes a standard audio power to be delivered to an optimum value audio load resistor connected across audio output terminals of the receiver under test (RUT).  Vt was series-connected to the input terminals of the RUT through a dummy antenna impedance made up of the series combination of a 100 pF capacitor and a 20 ohm resistor, for most of the tests. The unit-of-measure of Vt was microvolts.  The standard audio output power was taken as 100 mW.

Data set (B): The sensitivity measurements made for the graphed data uses microvolts/Meter (Hn) as a unit-of-measure of estimated field strength impinging upon a standard receiving antenna.  I assume that the measurements for the graphs comply with the standards laid out in Document 3.  Document 3 specifies the effective height of the antenna to be 4 Meters and its internal impedance is assumed to be that of a series combination of a 200 pF capacitor, a 25 ohm resistance and a 20 uH inductor.  See Section A, Part 6, page 108 and Section E, Part 2(a), page112 of document 3.  I am assuming that the normal test output power was 50mW, as specified in Document 3, page 107.

Normalization of "sensitivity" data to that which would have resulted if both documents 1 and 2 had used the standards in Document 3.

  •  Data set (A):  To convert the "Sensitivity in microvolts" figures to what would have likely been recorded if the test standards in Document 3 were used, one must adjust for the differences in audio output power.  The audio output power specified in page 108, reference 1 of Document 3 is 50 mW, one-half that used in the tests reported in Document 1. The "Sensitivity in microvolts" figures in Document 1 should be divided by sqrt2 because of the 2 times greater audio power output actually used in these tests.  The impedance of the dummy antenna used in Document 1 has a higher Q and a higher impedance than that used in Document 3. This should result in somewhat better selectivity and roughly no change in the sensitivity figures for many of the receivers.
  • Data set (B):  To convert the input signal (Sensitivity, Hn, in microvolts per Meter) to what would have likely been recorded using the test standards in Document 3, one must convert from EM field intensity impinging on an antenna of 4 Meters effective height to the internal source voltage of the antenna.  To do this, one should divide the plotted values for Hn by 4 to get the actual internal voltage of the assumed antenna. (see Section A, Part 6 of Document 6).
  • Help in understanding the graphs in data set (B):
    Top graph:  This is a plot of the overall frequency response from the audio modulation of an an AM input signal to the actual audio output of the receiver. Output voltage is normalized to 100% at 400 Hz.  Most of the graphs show audio fidelity when the RUT is tuned to 600 and 1400 kHz, showing how RF bandwidth, which changes with frequency affects fidelity.  Some graphs show fidelity at 600, 1000 and 1400 kHz.
    Bottom graph:  This graph shows measurements of two different parameters.
    1)  This shows a measurement of sensitivity. It is a plot of the input field intensity (Hn) of the carrier of a 30% modulated, 400 Hz AM signal impinging a receiving antenna having an effective height of 4 Meters vs the frequency of the carrier in kHz.  The unit-of-measure of Hn on the ordinate of the graph is in microvolts per Meter.
    2)  This shows a measurement of selectivity.  RF bandwidth, measured at points well down on the skirts of the RF bandpass is shown at three frequencies: 600, 1000 and 1400 kHz.  Three data points on lines labeled 10 x Hn, 100 x Hn and 1000 x Hn define RF bandwidth between two frequencies on either side of resonance (600, 1000 or 1400kHz).  For example, the bandwidth for the 10 x Hn point at 600 kHz is determined by detuning the frequency of the RF source up and down from resonance to find the two frequencies, on the skirts of the selectivity curve at which the standard audio output is attained.  The difference in frequency between them is the value plotted on the graph. An analogous process is used for the 100 x Hn and 1000 x Hn figures. The unit-of-measure of bandwidth, on the ordinate is kHz.

To simplify comparison of the sensitivity numbers in the two Data sets, one can convert the Hn figures in Data set (B) to be comparable to those of Data set (A) by dividing the Hn values by 4*sqrt2, or 5.7.

Table 1
Document #
Document Title
Report from Radio Frequency Labs on the Electrical Sensitivity, Selectivity and Fidelity of Forty 1929 Radio Broadcast Receivers  (330 kB)
2. Data set (B)
Graphs of Sensitivity, Selectivity and Fidelity of 57 radios manufactured in 1928 and 1929.  Each of the first fourteen 220kB links contains 4 graphs and downloads about 15 times faster than the last 57-graph link.  AK 37 to 42, AK 46 to 55, AK 55 to 60 , AK 60 to Balkite, Bosch to Colonial, Crosley Bandbox to 21, Crosley 21 to 41s, Crosley 42s to Grebe 7, Grebe Super Synchrophase to Majestic 72, Majestic 91 to 91 & 92, Peerless (Courier) 65 to Philco (Neutrodyne), Philco console to 512, Silver Lowboy to Sparton Equasonne, Stewart Warner to Temple, Zenith 39a,  All 57 graphs in one .pdf file of 5.6 MB  These are .pdf files. Use your Adobe Reader to rotate and resize the images as necessary.
Standard Tests for Broadcast Radio Receivers, from the 1929 IRE Yearbook  (1.3 MB)

Published: 06/09/2007
Return to the Front Page of the Site