[AMRadio] Xtal mike


Bob Bruhns bbruhns at erols.com
Tue Mar 15 13:21:31 EST 2005


Generally, precise impedance matching is only necessary where tuning
and loading are critical because of large power transfers, or where
absolute minimum noise is important such as weak signal work.

Crystal microphones are equivalent to a resistive source impedance
of a few K ohms in series with about 800 pF of capacitance.  800 pF
is a pretty small coupling capacitor for audio frequencies, and this
is why such a large resistive load is needed to obtain decent low
frequency response.  Assuming exactly 800 pF, a 1 megohm grid leak
resistor will produce a -3dB rolloff at about 200 Hz.  A larger
resistor will produce more extended low frequency response, and
unless the tube is gassy, the grid bias will be fine.  I used 15
megohms with a D-104 with good results.

The voltage output of a crystal mike is around 100 to 300 millivolts
peak to peak, plenty high enough for good low-noise amplification by
a tube.  Cable capacitance will reduce the output of a crystal
microphone by introducing a capacitive voltage divider.  If the
cable is 33 pF per foot, then 30 feet of it will reduce the mike
output by 50%.

Here is an interesting article on crystal microphones and preamps:
http://www.circuitcellar.com/library/print/1001/nisley135/2.asp

And some interesting reading on microphones:
http://www.du.edu/~jcalvert/tech/microph.htm

Dynamic mikes, with or without matching transformers, are expected
to drive a very high impedance amplifier.  The amplifier is expected
to have certain voltage noise and current noise characteristics,
which determine the optimum source impedance that should be
presented by the microphone and its transformer, if any.  This
optimum impedance is not by any means the optimum impedance for
maximum power transfer.  However, it will produce the best signal to
noise ratio, because it minimizes the noise.  Generally with a tube
amplifier, you want a very high impedance driving the tube grid
circuit, to produce the maximum voltage possible.  However, stray
capacitance becomes a problem when winding impedances have to be
very high.  With tubes, the optimum grid impedance is so high that
limitations of transformer or winding response are more important
than actual impedance match.

In some cases, particularly in the case of a high-impedance dynamic
mike/transformer combination, a small amount of stray capacitance
from the winding itself and from its connecting cable can cause high
frequency peaking, because of leakage inductance and any inductive
reactance in the microphone source impedance.  More stray
capacitance, from a longer connecting cable, or perhaps from RF
bypassing,  will lower the amplitude and frequency of this peak, and
cause high frequency rolloff above the peak.  Electric guitarists
know about the effect of cables on the sound of their guitars, with
their magnetic pickups.  The peaking effect can be compensated by
loading the circuit with a certain amount of resistance.  The exact
value of this resistance will have to be determined experimentally.
This resistor will reduce the microphone output slightly, and
flatten the peak.  Rolloff above the peak will be less severe than
it was without the resistor, but it will still exist.  So you want
the peak to be as high in frequency as possible, which means short
runs of low capacitance cable if mic impedance is high.  (This
effect is the reason we use low-impedance dynamic microphones when
long cables are required.)   If you add RF bypassing, you will need
to change any loading resistor value as well.

You can experiment with this, if you have some audio test gear.  I
got good results once with a 68K resistor in series with a capacitor
(I forget the value, something like 220 pF), across the transformer
secondary (high-Z side).  The input transformer had a 10 dB peak at
20 KHz!  The R-C snubber improved it tremendously.

  Bacon, WA3WDR





More information about the AMRadio mailing list

This page last updated 24 Oct 2017.