🚀 go-pugleaf

My last question was "What is the truth about caps", now I have
questions about resistors in a low noise preamp design.

Metal film has low noise but but SMD are thick film types and
has poor noise performance. So where do I have to use low noise
resistors and where can I use SMD resistors?

Imagine a fairly standard two transistor preamplifer. Witch ones should be
of low noise type:

	R1C	39K collector resistor for the first stage

	R1E 	2.2K emitter resistor that are in parallel with a
		electrolytic cap and in series with...

	RS	a 6.6 ohm resistor from R1E to ground.
		There are two emitter resistors in series, R1E and RS.

	R2C	1K collector resistor for the second stage

	R2E	180 ohm emitter resistor for the second stage that 
		has a electrolytic cap in parallel

	RD	1K DC feedback resistor from first stage base to 
		last stage emitter

	RA	2.2K a AC feedback resistor from the connection of RS and
		RE1 and to the collector of stage two, i.e. the output

This is what I think. R1C is important because it is in the first stage
and current variations will be amplified of stage two. RE1 is not
important because it has a cap in parallel that short circuit most 
AC including the noise. The same is for R2E. R2C Is in the last stage
and will not give significant noise. RD will for AC be a resistor
from the input pin to ground and could give a lot of noise. RA and
RS is harder. The value of RS is just 6.6 ohm, will it give significant noise?
If RA gives much noise I don't want to replace it with a trim potentiometer
but use a swither with different metal film resistors for gain adjustment.
(I assume there will be noise when I switch ;-(

The source will be a 200 ohm mic.

Do you have any thoughts about this?

Now some more cap questions. The caps used in this design is rather
large, 470uF for the input, 100uF bypass in the first stage and 220uF
bypass in the second stage. 

I can't do much about the input cap, I suppose it has to be that high
for a 200 ohm source and a 200 ohm inout.

My first thought was that if I reduce the bypass caps values I will 
lose gain in the low frequencies. I think this is wrong, it will give 
higher distortion in the low frequencies?

The gain is set with RS and RA to 333 (=2200/6.6). As long as the total
gain without feedback at the lowest frequency I want it to handle is
over 333 there will be no gain loss with feedback. If the gain without
feedback is nearly as low as 333 the distortion will be very high.

Am I right? 

How bad is it to accept less gain without feedback?
Does transistors give more or less distortion at frequencies
below, lets say 200 Hz. How about the human ear. Is it more or less
sensitive for distortion at these frequencies?

Is there any general rule how much higher the gain without feedback
should be in relation to the gain with feedback? Maybe I can remove
the bypass caps totally. If I remember correctly removing them is
also feedback to each stage and distorsion goes down?

Why do I bother, caps are very cheap? I want to keep the values as low
as possible to save space on the PCB but keep the good performance.....

The input transistor is a selected BC 337. It is not stated HOW it is
selected. It was chooses because it is high current and has low input
impedance. Do you understand how it is selected? The book it is taken 
from is rather old, do you know a better "low noise" transistor to use?

/kgb



--
Kent Boortz
boortz@sics.se