[AMRadio] Microphone recomendation

D. Chester k4kyv at charter.net
Fri Sep 25 12:42:17 EDT 2009

I currently use two microphones mixed together in proper phase, with a 
homebrew tube type 2-channel audio mixer.  One mic is a vintage hi-Z Astatic 
D-104 and the other is an Electrovoice model 670 dynamic that otherwise 
might have been tossed out because it somehow developed very poor high 
frequency response.  I mix in some audio from the dynamic to give the D-104 
a little more bass response and mellow out the characteristic shrillness.

Each channel of the mixer pre-amp has two hi-mu triode stages of 
amplification using a 12AX7 tube.  Since the audiophools have driven up the 
price of n.o.s. 12AX7's to ridiculous extreme, a good alternative is a pair 
of older tubes, the 6F5 or 6SF5, which are electrically identical and have a 
mu of 100, just like the 12AX7.  The D-104 has a built-in acoustical 
pre-emphasis curve, achieved by a response peak, that gives it the unique 
D-104 sound.  But I added some additional pre-emphasis to the D-104 by 
careful choice of cathode resistor by-pass in the 12AX7 stages, so that the 
pre-amp has a rising characteristic starting about 800~ and rising up to 
about 9 dB at 2000~ and then levelling off.

The combination of electrical pre-emphasis and peaked response curve of the 
D-104 gives it good punch by emphasising the sibilance frequencies of the 
voice. The bass response of the dynamic balances out the shrillness of the 
D-104 to give the overall audio a pleasant sound, at least with my voice.

I use a low-pass filter in the audio line between the pre-amp/mixer and the 
compressor/peak limiter.  It gives me the switchable choice of two passive 
L-C filters.  One, which I use most of the time, has a gradual roll-off, 
beginning at about 5000~ with complete attenuation at about 7500~.  For 
congested band conditions I switch to a sharp cut-off brick-wall filter that 
is flat out to about 3300~, but near complete attenuation at 3400~.  I can 
switch out the filter entirely, but rarely run that way because I know some 
of the "broadcast quality" audio transformers in the chain have measurable 
phase shift distortion beginning around 10K, and this could cause some of 
the push-pull stages to generate splatter and distortion around 10 kHz from 
my carrier frequency.  Besides, very few receivers used by amateurs would 
respond to 10 kHz audio because of the necessary selectivity for listening 
on the ham bands.

Even with the 3400~ filter, the rising characteristic and peak in the 2-3.5 
kHz region compensates for the lack of highs above the cut-off frequency 
while balancing out the low frequency response of the dynamic, and I 
routinely get reports of "broadcast quality" even when everything is cut off 
above 3400~.  Most signal reports say there is only a subtle difference 
between the 3400~ and 7500~ cut-off, although with wide-open selectivity at 
the receiver the difference is said to be  readily noticeable.

The filter is followed up with a UREI BL-40 Modulimiter, which is a 
processing device built for AM broadcast use.  It keeps the modulation level 
near 100% without overmodulating, and I don't have to ride the gain and 
closely watch the scope at all times to maintain a good modulation 
percentage without pinching off the carrier.  Reports say the limiter is 
very transparent and has  little effect on the sound of the audio.  I do not 
try to compress or limit excessively in hopes of increasing the average 
power level in the sidebands; that produces distortion with little benefit. 
I just follow the manufacturer's recommendation for broadcast operation. 
The Modulimiter has two limiter stages, the first one called "RMS limiting", 
which is basically the same thing as compression, followed by "peak 
limiting" which is pretty much the normal meaning of the term.

All units in my audio chain: pre-amp/mixer, low-pass filter unit, 
Modulimiter, line amplifier and transmitter are linked together with 
balanced 500/600 audio lines.

Click on the link below to see where I first got the idea for my audio 
response curve when I built the mic pre-amp some 30+ years ago.  I didn't 
follow his instructions exactly, and I don't guarantee that George's theory 
is 100% correct, but using it as a starting  point for trial-and-error 
experimentation I got the best sounding audio possible with the microphones 
I had on hand. Since then I have tried some expensive broadcast type 
microphones running the pre-amp at flat response, and the audio was always 
inferior to what I get with my two-mic combination.

Note: results may vary with different voices.  I have monitored my audio 
while visitors were at the mic, and while some sounded great, others sounded 
like crap. I firmly believe you need to tailor your audio to sound best with 
YOUR voice over the air.



This message was typed using the DVORAK keyboard layout.


More information about the AMRadio mailing list

This page last updated 16 Dec 2017.