[ARC5] Crystal Headphones

[Richard Knoppow]

    One must be careful not to confuse sensitivity with the 
impedance. Sensitivity is a power measurement, that is electrical 
power in to acoustical power out. It is possible to have high 
impedance but low power sensitivity although I can't give  you an 
example of a commercial headphone. High impedance is desirable 
for sources that are high impedance such as crystal sets and AC 
impedance bridges. However, where the source is low or medium 
impedance headphones with matching impedance will be louder than 
the high impedance phones. Testing for sensitivity requires a 
calibrated artificial ear if the results are to have any 
accuracy. These are rare. However, one can test with just  your 
ears or a sound level meter with the microphone held against the 
headphone. An interesting experiment is to take an oscillator and 
feed various headphones from it through a very high value 
resistor, preferably at least five times the expected headphone 
impedance. 100K is enough although more is better. Just connect 
the phones and see which is the loudest. Usually, it will be the 
ones with the highest impedance but some with the same impedance 
will differ. In some cases a lower impedance phone will be louder 
than a higher impedance one. The difference is probably in the 
magnet strength but can also be due to differences in the 
diaphragm spacing from the pole pieces.
    The loudest phones I have are: Western Electric 509W, WE 
Signal Corps P-11, Baldwin Type C.  However, if you make the same 
test using a low impedance source, say 500 ohms, the results will 
be different. For instance WE type 528 (600 ohm) will be louder 
than the WE 509W. WE and Trimm made magnetic phones down to about 
50 ohms per pair. For the same reason you may find modern 
high-fidelity moving coil phones too loud when connected to a 
receiver where they are across the loudspeaker output. Typically, 
this will be 4 to 8 ohms and most of these phones are are around 
50 ohms impedance. The receivers will often specify 500 ohms 
although 20K phones usually work just fine and are of reasonable 
loudness.
    Measuring efficiency is a matter of measuring power out to 
power in. That requires an impedance match on the electrical side 
and, for best power efficiency, also on the acoustical side.
     Most of the magnetic type phones have a strong resonance, 
typically in the vicinity of 1Khz. While some attempt was made to 
broaden out the resonance (such as using a salt shaker type cap) 
the difference is not great. It is possible to make quite non 
resonant magnetic phones but they become complex. The best 
example is the Western Electric type HA-2 designed for the series 
500 telephone. These are described in great detail in the Bell 
System Technical Journal and elsewhere by Mott and Miner of Bell 
Labs who designed it. There are other types that fall broadly in 
the "magnetic" category, such as the balanced armature type as 
typified by the Baldwin phones with mica diaphragms and also used 
in some sound-powered phones.
     While one can find oodles of information about moving coil 
speakers and microphones in the technical literature of the last 
century there is not much about the plain magnetic phones 
although they were made by many different companies. Frustrating.
    FWIW, the highest impedance phones magnetic phones I've 
measured are the WE SC type P-11 (close to 30K), others were the 
WE 509W (25K), Trimm Featherweight (24K) but the Featherweights 
are not as loud on a very high impedance source as the WE phones 
or as the Trimm Commercial phones which measure only 17K 
(advertised value).
     Measurements were made on: General Radio 650A bridge with 
external null detector, GR 1650A, GR Z-Y bridge, all at 1K
    Note also that magnetic phones are mostly inductive so the 
impedance varies with the frequency. I measure at 1K. I found the 
advertised values of Trimm phones are high at 1K but about right 
at 1200 Hz. However, the resonance is at about 1K. I tried to 
measure the motional impedance of a couple of phones but gave it 
up because I can't separate it from the overall impedance. i.e., 
you can't see a resonant rise in the impedance.
     Enough, I am glad at least one or two people find this of 
some interest.
On 9/18/2018 1:35 PM, Dennis Monticelli wrote:
> More modern magnetic phones (still of the old style) can be quite 
> good.
> 
> I have a NIB set of H-43B/U headphones (datecode 1990) that were 
> made for Geiger counters.  I measured an impedance of 21K at 1KHz 
> so I presume the transducers are 10K wired in series.  These cans 
> produce plainly audible tones with only 1mV rms.  That equates to 
> 50 picowatts!  Impressive sensitivity and high enough Z for 
> crystal radios.
> 
> Dennis AE6C

-- 
Richard Knoppow
1oldlens1 at ix.netcom.com
WB6KBL