Mic preamplifer.. another way

[Sallivan]

Hi all , I found this transformless mic preamp scheme.
The input stages are designed whit a pair of LM394 (two transistor perfectly matched), and the others opamp are ne5532 (but it is not specified).
The specifications that I have read seem to be really good.
Low noise
cmrr  120

The gain can vary from 4 to 280.

I drew the circuit of which I spoke.
What do you think about this circuit?
Someone knows the LM394? I did a search on the focus site, and I discovered the two transistor are in one single Ic. Right?

 

[PRR]

It will work.

There are better plans.

There's a $5 chip which does all this.

 

[Bill Wilson]

There are simpler designs using the 394. There is one problem; national made the last run 5-6 mos. ago.
I have 85 pieces I was able to purchase for the right price, would be glad to send you one or two.

I also have a simpler mike pre that will only use one lm394. The thing to remember about LM394 is that the device is actually 25 or so transistors paralelled on each side of the device, that is why it is so quiet.
Send me your email address & I will send you the diagram for a preamp I manufactured & installed in MCI 416 & 428 consoles. It used 2ea. 2n4403's in paralell on either input. very quiet as long as it was terminated w a low impedance device. The one drawback is that it is very noisy withoiut a termination on the input; you just had to keep the fader closed on the unused preamps.

Bill Wilson

 

[JohnRoberts]

What PRR said...

LM394 are reasonably quiet, but that is not the quietest topology.

JR

 

[abbey road d enfer]

In addition, the 120dB CMRR is dependant on the 1ppm matching of several components...

 

[analag]

A transformerless mic pre is not a desirable thing around here. It will always be the second best way to design a mic pre. Took me a while to wrap my mind around it but I did.

 

[JohnRoberts]

opinions vary...

I have favored transformer-less, since well before they were as common as today. 

The threshold event for me was finding some .4nV/rt Hz bipolar devices (late '70s).

JR

 

[abbey road d enfer]

A transformerless mic pre is not a desirable thing around here. It will always be the second best way to design a mic pre. Took me a while to wrap my mind around it but I did.

Such a bold statement needs to be substantiated; until then, it will remain a belief.

 

[PRR]

Good slide-show on mike-amp basics with focus on transformerless (favoring THAT Corp parts, but the ideas are universal):
http://www.thatcorp.com/datashts/AES129_Designing_Mic_Preamps.pdf

 

[Samuel Groner]

The thing to remember about LM394 is that the device is actually 25 or so transistors paralelled on each side of the device, that is why it is so quiet.

The paralleling is done mainly to ensure good matching/insensitivity to process gradients; low noise transistors are easily done otherwise as many discrete specimen proof.

There is one problem; National made the last run 5-6 mos. ago.

MAT12 is the way to go (or the cheaper SMD ones from ADI).

The one drawback is that it is very noisy without a termination on the input.

That's mainly a function of its input impedance. Lower input impedance means less noise unterminated. But lower current noise (read low Ic and high hFE for the input transistors) will help too, although possibly at the cost of voltage noise.

In addition, the 120 dB CMRR is dependant on the 1 ppm matching of several components...

At 40 dB gain this topology only needs ~100 ppm matching for R16/R19 = R17/R18; the first stage amplifies differential signals only, so gives about 40 dB free CMRR without any matching need.

Samuel