question/design discussion - what are the functions of a mic preamp?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

jcharles00

Well-known member
Joined
Dec 28, 2005
Messages
51
Location
Indiana
I'm kind of afraid to ask this question, because it sounds pretty dumb, but after digging around here and elsewhere on the net, I've only come to a vague conclusion and I think hearing some different opinions will help make more sense of it

what functions does a microphone preamp perform?
I ask this in the context of most basic, but usable design. (say, something that will accept the XLR input of a dynamic mic and output line blanaced line level)

The obvious is that it increases the voltage of it's input signal, but it seems like impedance matching on the input and output are vital to the process as well. Conversion from differential signal to single ended too. ..and then filtering?



I know this is a nebulous question because we could get down the road of secondary features like phantom and pads, etc, etc.. but I'm talking about a simple device to get usable, "clean" audio from an SM57 into an mbox line-in without damaging anything. (as an example)

The reason I ask, is because I'm trying to wrap my head around mic pre design, and thought the best way would be to compartmentalize the device's operations to try and create very basic solutions to each part from datasheets, then compare what I came up with to existing designs. I could just wire an input jack's pin 1 and 2 to one side of an opamp with a non-inverting feedback loop and it would do _something_, but i would be ignoring issues like the impedance, and introducing more noise than necessary by not converting the balanced signal, etc.

Surely not a black and white question, but I'm not sure how else to approach this without just lifting someone else's design, and it doesn't seem like I'd get as much from that educationally.


Any input? Is this approach good/not? How would you start, designing something from scratch?
 
The task is pretty much to boost the low voltage coming from the microphone up to a usable line level signal.

Input impedance concerns are that the mic is generally looking for a load that is 10x it's nominal source impedance, or roughly 1.5-2k Ohm.

Input noise should be low wrt microphones source impedance (150-200 Ohms) thermal noise.

These general design requirements suggest low noise bipolar devices direct, or using step up transformers with higher noise/impedance devices (like opamps or tubes, or whatever...).

There are many ways to meet those general requirements, and other ergonomic considerations like features and gain law, etc.

JR
 
Boiled down to absolute basics, the list of what a mic preamp should do is threefold:

1. It shoudl accept signal from a microphone. (This includes issues like proper impedance termination, which these days isn't matching, and comprises a whole bunch of interesting stuff. It also covers issues like balancing.)

2. It should amplify the signal by an amount which the user chooses.

3. It should output the signal to whatever load (resistive, capacitative, balanced or not, etc.) it needs to drive.

Accept, amplify, output. That's the story in three words. The devil, of course, is in the details.

The list of things a mic preamp shouldn't do is rather longer, and worth investigating as well.

By the way, I commend your line of inquiry. The first step in designing something should always be an analysis of what the job is.

Peace,
Paul
 
> what functions does a microphone preamp perform?

Extend to the complete audio chain.

From our point of view, there is the Telephone Line, about a part-Volt at about a part-K impedance or about a milliWatt, and things which can be stuck in Lines (loss-recovery, EQ, recorders).

If you have a carbon mike against a mouth, it outputs Line Level.

If you have a ringy receiver against an ear, it accepts Line Level.

Most other gizmos need boost. A SM57 2 feet out from an acoustic guitar needs gain of like 100 to reach Line Level. A loudspeaker to fill a room needs a Whole Watt (or more) and needs a boost up from Line Level.

Lines may be short or long. Long lines pick up hash from distant lightning, plus nowdays all the electric crap of the modern world. There are several ways to reduce this crap. A very effective one is to run two wires "balanced". When everything was carbon-mike and ear-phone, this was trivial. Until the common battery was invented, then every line needs transformers or other way to balance circuits derived from a common battery. However short lines have often worked fine unbalanced.


> one side of an opamp with a non-inverting feedback loop and it would do _something_

That works and I've made money with less.

If the opamp is not AT the microphone, the weak mike signal is especially prone to pick up electric crap noise. Dynamic mikes are easily made balanced and the SM57 is made that way. I done a lot of work un-balanced, but in general you should be prepared to extract the '57's signal balanced.

But in specific: use a good op-amp, let the non-inverting input impedance be 47K, let the gain be adjustable from 10 to 100 with as low a feeback network impedance as the chip can drive, and feed an unbalanced line-in. It will work, and may work wonderfully well.

The self-noise of a 200 ohm resistor (or SM57 in vacuum) is 0.2uV over the audio band. The acoustic noise from a '57 in a quiet studio is 0.2uV. The noise of a simple transistor in a cheap circuit is 0.5uV; a NE5532 about 0.5uV; a TL072 about 2uV; a 12AX7 about 1uV-2uV; a TL072 in a 1-opamp 4-resistor diff-amp 3uV. When I used cheap transistors in PA, I could hear the system was "on" by the hiss, but not with a crowd; when I was forced to use a CRATE head with TL072 inputs the hiss was pretty audible.

NE5532 is now under a buck. It will be "noiseless" in all but the quietest rooms. For feedback, use 1K fixed plus 10K variable from out to "-", and 100R from "-" to ground. Power it with two 9V batteries, the simplest and cleanest +/- supply possible. You can do a whole album with a 10-pack of batteries which is cheaper than a clean +/- wall-supply.

You may find that your room has too much electrical noise to run a dynamic mike line unbalanced any distance. You can put the small battery preamp AT the mike, and that will generally overcome induced crap; I run hundreds of feed unbalanced and even unshielded. If not, then you have wasted $10 and a couple hours, need to move to a much more elaborate balanced input (or a different room).
 
thanks for the input guys! these posts are exactly what I was looking for and should give me a great starting point.
 
alright, time for my next question..

i've breadboarded a little amp that accepts a balanced signal, unbalances, and then has a variable gain stage. running it on 9V batteries per PRRs suggestion and to help eliminate potential causes of noise.
it's really noisy and seems to pick up the same ESPN radio station (apparently AM1420) that my green pre gets, but only when there is a mic cable attached to it.

so I've been looking at filtering a lot. i still don't really understand the techniques. I note that most mic pres seem to use 47uF caps, and in the case of simple amps i've been looking at, they are usually used in conjunction with 10K resistors. I plugged this in to a filter calculator, and came up with a frequency of 0.3386255319148936Hz. My curiosity is, what exactly are we filtering out that exists at that frequency? I know it's a dumb question, but I'm having a hard time finding answers dumbed down enough for me to understand. From what I do understand, AM radio lives around 1MHz.


I know these posts are probably pretty annoying to most of you who can build this stuff in your sleep, and I surely could just take the 47uF cap and 10k resistor and use them the same way as the other designs i've seen, but it's fairly important to me to figure out what exactly is going on before throwing parts at it. hopefully i'll figure out how to fix the scope i bought soon, and maybe being able to observe this stuff will help me answer my own questions. ;)
 
See page 15 below:

http://www.national.com/ds/LM/LME49740.pdf

It is quite common to have a filtering network on the input. Here is my very poor representation of one:

+
  |
  |
  __
  __ 470p
  |
  |_________
  |            |
  |            __
  __          __ 47p
  __ 470p    |
  |            \/
  |     
-

Also, here is an interesting paper about ferrites:

http://www.audiosystemsgroup.com/SAC0305Ferrites.pdf

Don't forget the "openness". A mic preamp must contain not only "magic smoke" in order for it to work, it must also contain that mystical, magickal, elusive property called "openness".
 
Back
Top