[AMRadio] D-104 modifications

D. Chester k4kyv at charter.net
Fri Aug 27 16:09:22 EDT 2010

>The continuing
> erosion of sales combined with problems securing the Rochelle salts
> required for making the crystal elements was apparently what brought
> them down.
> ~ Todd,  KA1KAQ/4

Why is there a problem securing Rochelle salt?  I have heard this elsewhere 
on the net too, but Rochelle salt is not exactly unobtanium. There are 
instructions available for making the stuff at home using every-day 
household chemicals, for example 
http://rimstar.org/materials/piezo/rochelle1.htm  I think that story of 
scarcity of Rochelle salt is bogus.

I bought a couple of crystal elements directly from Astatic in the mid 70's 
and they seem to work very well. I have an older one that uses a ceramic 
crystal instead of Rochelle salt, but those have weaker output and  don't 
seem to sound as good.

For the past few months I have been using a D-104 connected for balanced 
output.  One of the nice things about the classic stock D-104 elements is 
that they are mounted in a bakelite case with two ungrounded output 
terminals.  Per the factory data sheet, the element can be used to drive a 
push-pull preamp stage.  All that is required is a two-conductor shielded 
microphone cord.  The two conductors are each connected to one of  the 
output terminals of the cartridge, and the shield is grounded to the case at 
the interior of the head.  With a plug-in head, this is done through the 
three-prong plug; two to the mic element and one to ground.

The balanced push-pull audio is established through the grid resistors of 
the pair of pre-amp tubes.  The leads from the mic element each go directly 
to one of the tube grids.  No blocking cap is needed.  The grid resistors 
must be matched with an ohmmeter to have very close to the same resistance. 
The common ground connection between the two grid resistors establishes the 
midtap for the push-pull audio.

With this circuit, the mic can run with a load as high as 20 megohms, since 
as a load on the mic element, the grid resistors appear in series.  A 10 meg 
grid resistor for each tube gives a 20 megohm load, or a 5 meg resistor for 
each tube gives a 10 meg load.

A crystal mic element can be thought of as an ideal a.c. generator with 
about a 500 pf capacitor in series.  That's why such a high load resistance 
is required for good low frequency response.  For a low-frequency equivalent 
time constant, a very high resistance is required when the series 
capacitance is low.

In addition to improved low  frequency response, I have found mine to be 
immune to rf pickup and a.c. hum.  My old single-ended preamp required an rf 
filter network at the mic input connector to avoid feedback. I had to run DC 
on the filament of the input stage to reduce hum to a satisfactory level. 
My pushpull preamp  has no rf suppressor circuit at the input, and the  hum 
level is satisfactorily low even though I run a.c. on the  filaments.

I use a pair of 6F5 tubes, a metal octal tube that is equivalent to one 
section of a 12AX7.  I use the separate triodes instead of the dual-triode 
12AX7, because with  the 10 megohm grid resistor, the plate current is a 
little flaky and may vary from tube to tube, and I select a pair that has 
closely balanced plate currents.  The easiest way to compare the plate 
currents is simply to measure the voltage across each of the cathode 
resistors, which should be very close to the same for each tube, since the 
full plate current flows through the cathode resistor. If you can't find a 
6F5, exact equivalents include the 6SF5 and the triode section of a 6SQ7. 
Or if you have a bunch of spare 12AX7s, try to find one with both sections 

Don k4kyv


This message was typed using the DVORAK keyboard layout.


More information about the AMRadio mailing list

This page last updated 14 Dec 2017.