mic pre concepts

Hi folks...

Lets start a nice thread all about the hows and whys of mic pre design!

For starters - what is the output impedance and voltage range of popular mics? How would you go about making a de-balancing buffer?

If could get that far along, I'd be a happy happy man

I believe that a standard number used for mic output impedence is somewhere around 200 ohms. As far as output voltage, I assume you want to know because you need to know how much amplification a mic amp shoud deliver. I see mic amps having anywhere from ~40 to ~80 dB of gain.

As for unbalancing, there are many ways to do this. A transformer is one common way as is using an opamp (discrete or otherwise). Some mic amps don't unbalance at all (Millennia HV-3).

Ian

> voltage range of popular mics?

0.2 to 3,000 milliVolts.

[quote author="PRR"]0.2 to 3,000 milliVolts.[/quote]
wide range, isn't it
that decimal point moves a long way
:roll:

> wide range, isn't it

Sure. What other standard audio box needs a Gain control? Volume, yes, Relative mix level, yes. Even some trim for hi-fi or Pro line-level. But nothing else routinely has 20-60dB of Gain control that isn't really about listening level, but just getting the signal to a rough level for further adjustment.

You have mikes that, at 74dB SPL, give 0.1mV (150Ω ribbons) or 2.5mV (hot large-diaphragms).

You have room-miked harpsichord that may not beat 80dB SPL peaks, and you have too-close percussion and overpowered amp-speakers with peaks above 134dB SPL.

Now, maybe you don't, or shouldn't, grab the ribbon for far-harpsichord, nor the hot condenser for snare-kissing... yet often we do.

[quote author="PRR"] ... nothing else routinely has 20-60dB of Gain control that isn't really about listening level, but just getting the signal to a rough level for further adjustment. [/quote]
people too easily forget this or just don't appreciate it in the first place. I guess a guitar amp input is another high gain wide range sort of thing.

It is also an obvious point of difference in the demands of recording verses mastering ... on the monitoring system. During recording you must expect high and wide dynamic range and a variety of levels.
At mastering these are well sorted and finesse of the monitoring becomes the prime function and need.

Pre-amps and inputs for vibration sensors must be interesting things ... the ones monitoring earth movements etc .... I wonder how much headroom they have ?
:green:

> I guess a guitar amp input is another high gain wide range sort of thing.

Guitar inputs are easy.

yeah you are probably right guitar inputs are much easier.

Back to Mic-pres ...
so having established the wide voltage range required, what is the next most important and restrictive factor one has to deal with ??

I talked to this genius at the AES show for about an hour. He wouldn't let me leave! Smart people are like that. There used to be a PhD that worked here. Once he had you as his verbal hostage, he wouldn't shut up, so you just had to sort of slowly walk away unless you wanted to be trapped all day. The smarter they are, the tougher it is to get away. These types aren't very good bussinessmen, as pratically tell you the whole deal about their products, especially if they are really proud and excited about it.

Anyway, I heard more new ideas out of this guy than anyone at the show. I do not know how you will implement them.

http://www.proaudioreview.com/par/june03/gordon_micpre.shtml

http://www.gordonaudio.com/

Grant Carpenter:


That's all one big board, hand soldered! Available in Teflon and Teflon caps at an extra charge.

For the first stage, do you generally use a pair of followers on the hot and cold to present a huge input impedance, or do you want gain immediately to get the signal up out of the noise floor?

so why use feedback to begin with? with a fet, you can feed it whatever voltage as long as the gate is being fed enough voltage.. so why not adjust the voltage to the fets instead of adjusting the feedback to cancel signals for volume control?

[quote author="cjenrick"]IAnyway, I heard more new ideas out of this guy than anyone at the show. I do not know how you will implement them.
[/quote]
It can be useful idea to make a set of amplifiers with definite
gain and to switch between.

But why use rellays. Why not use mercury contact.
It can have better noise performance?


And why not to set on outer limits?
To construct amplifier with paralel amplifiers with different gain?
You can record everything from thermal noise of microphone
to 140 dB parts, and after recording you can choose?
Neumanns (ant the others) did very hard work on it and here it is without response?

Is there someone else, who want to record unrepeatable music
with high dynamic ?


xvlk

[quote author="PRR"]> I guess a guitar amp input is another high gain wide range sort of thing.

Guitar inputs are easy.[/quote]

Hmmm..I´m not sure I would agree with this...seen from a basic guitar input circuit...and a ELA M251 circuit..not that much more or less difficult to me..but thern again... this is from someone who for years has been saying that guitar amp circuits has been designed all wrong from the start..!!! :green:

Kind regards

Peter

One problem with the parallel amplifiers at different gain is once an amplifier clips, it takes a significant time to recover. Sometimes upwards of 100 microseconds to a millisecond, so during transients the switching from one amplifier stage to another would be tricky at best, not to mention using lots of channels for one signal if you want to do it later.

The mic head amp isn't really that tricky except for getting the grid current down to where the high resistances (30M - 1G) can deal with it. Of course, we have the opportunity to look at previous designs, I imagine it was a bit difficult to design the first condensor mic head amps when people haven't done that before. The mic preamp is a bit trickier, and there's something to be said for Neve's approach of two stages in series where one can be bypassed for low gains.

A problem with many op-amps is that they are unity-gain stable, and that limits the gain * bandwidth, so the frequency response isn't high enough to properly amplify high frequency components at high gain levels - the frequency (and phase) response of the amplifier shifts with gain. Add stages and noise gremlins come in to bite. If you're not running the amp at unity gain then you don't need unity gain stable. Really, for a preamp you would ideally be able to adjust the op-amp's internal compensation against gain. A designer here is wise not to use the Av = R1/R2 formula for gain - that's the simplified version that doesn't really work if the gain is high. A nice op-amp for this is the AD797, it allows the designer to decompensate the amplifier.

I think low-noise circuit theory indicates the first stage should have the highest gain - this minimizes noise in the overall system.

:roll:
no-one answered my question
did we get sidetracked
if so that might have been my fault by including the guitar pre-amp comment.

Mic-pres ... what do they do ?
Unless we are to invent a new Micrphone then they are used to increase the signal level from existing Micrphones and provide some drive to the next stage. The amount of drive can be debated but first the interface with the mic.

We have a voltage range of popular mics at 0.2 to 3,000 milliVolts.
Now what is the next most important or restricting factor one has to deal with ??

Having established the issues first, you can then begin to formulate a solution.

I'll vote for:
1. Phantom power and level shifting to get rid of it
2. Noise
3. Frequency response
4. Phase response

Others?
One thing I'd like to see in a preamp (in these days of digital recording) is a decent sounding peak limiter/compressor to avoid digital overs when you set your levels, hit record, and even the margin you gave wasn't quite high enough.

[quote author="dale116dot7"]
Really, for a preamp you would ideally be able to adjust the op-amp's internal compensation against gain. A designer here is wise not to use the Av = R1/R2 formula for gain - that's the simplified version that doesn't really work if the gain is high. A nice op-amp for this is the AD797, it allows the designer to decompensate the amplifier.
[/quote]

I have been entertaining this idea for some time. Using a double pole rotary switch for gain control one circuit could adjust Ccomp and the other Av. This way at higher gains the amp would be gradually decompensated as needed.

I like a mic pre not to tamper with or change the signals original or accoustic qualities as much as posible. I know it is impossible to reproduce the exact same sound but this is the quest. I loved getting to the end of my first DIY project just to discover that all it was, was an input stage.

I still learn about what a mic pre really is and I'm keeping simple cause I want to hear what kev has to say next.

Kev

input loading

One thing that might be interesting with a pre is an input Z that drops with gain setting. I have built a boost petal for guitar thats input Z drops with gain. This seems to keep the high freg distortion under control. Now this would be for a pre were you wanted distortion and were using a dynamic microphone.

ok so Gus has gone with input loading ...
we had a hint of this in Ian's post (2nd post of the thread and was a direct answer to the original question)
" ... somewhere around 200 ohms. "

0.2 to 3,000 milliVolts.
around 200 ohms (perhaps this should be a range as above - lots of mics out there)

what's next ?

noise perhaps ... and this could be expressed in a number of ways. Our aim may be to be higher spec'd than most microphones ... noise is lower and signal to noise is higher than the mics.

think about it

we like it BUT if it were noisy would we not use it ?
(if your unit can handle 0.2 to 3,000 milliVolts perhaps noise all falls into place anyway ?)

perhaps noise is not a deal breaker ...
:roll:

then what is ?