The transducer that ultimately delivers
the signal to the ear is an extremely important part of any crystal set.
Really significant performance gains can be made by simply plugging in
the right headset.
The standard headset for crystal sets
since the dawn of time has been the 2000 ohm (DC resistance) 'phones with
steel diaphragms directly actuated by electromagnets. These are hard
to purchase new, but there's a lot of them around. Crystal earplug
units are still available, they offer an excellent bang for the buck. (Actually
two bucks.) Hi-fi headsets can be used with matching transformers,
but most are pretty inefficient. There are some better, more exotic,
solutions to the headset problem. Generally, higher impedance is
advantageous. Lo-z phones mandate matching transformers. The
following report should give you some ideas:
A report from the Stanton Signal Laboratory:
A recent discussion of headphones for crystal
sets on the Boatanchors mailing list prompted me to go to the lab and make
some, at least half-way, scientific comparisons. I had previously
attempted to make measurements with a signal generator and a microphone,
but peaks and dips in the responses make this a questionable practice.
This time I decided the criterion should be
intelligibility of speech at low signal levels. I tuned my shop radio,
a Collins 651S-1, to the local talk radio station, and adjusted the output
for -20dBm into a 600 ohm load. This station is almost all talk and
the audio levels are quite consistent. The receiver output was then
applied to a stepped attenuator. At the output of the attenuator,
three different audio transformers were available to provide impedances
of approximately 50, 125, 300, 500, 1K, 2K, 4K, 8K, 12K, 50K, and 100K
ohms.
The test procedure was to set the attenuator
for an output of approximately -50dBm and choose the transformer tap that
gave the loudest signal from the headset under test. The the signal
level was then reduced to a point were I could still follow the conversation,
and, presumably, ID the station. This signal value appears in the
sensitivity column below. I also measured the DC resistance of each
headset.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
TRIMM K: Trimm type K, from 1950's. A better quality headset using steel diaphragms.
TRIMM FEATHER WEIGHT: Light-weight version of above.
MOUSER: Contemporary crystal (possibly ceramic?) ear plug from Mouser Electronics. An excellent buy at about $2.00. In actual use you'll want a 100K resistor across it to provide a DC return for the detector.
BRUSH: The typical Brush crystal headset from the 1930's-60's.
BALDWIN TYPE C: The old standby "Baldys" with the balanced armature driver and mica diaphragms. They bear 1910 and 1915 patent dates, but were manufactured until, at least, 1941. These are typically very sensitive for CW, but roll off steeply above 1 KC. Even so they're a good bet for crystal set use.
SOUND POWERED #1: Baldwin cases with elements from post WWII TA-1 P/T sound powered field phones.
SOUND POWERED #2: Elements form one of the clunky (Navy?) sound powered handsets cobbled into an H-161 VRC-12 headset.
One could probably do a lot of nit picking with my test methodology, but the results are consistent with my observations in actual crystal set use. The Baldwins are better than any of the standard headsets I've encountered. The sound powered elements, which share the same basic construction as the Baldys, are considerably more sensitive because they were designed to be peaky in the 2-3 KC range to optimize voice intelligibility.
By the way, my current double-tuned loose-coupled killer crystal set, using the SP #2 headset, and a 100 ft flat-top antenna, hears 24 day-time stations in the NYC-Philadelphia-Scranton region from my western NJ QTH. The minimum input carrier lever for a intelligible AM signal is about -66dBm.
For some interesting information on headphone testing see Ben Tongue's Webpage