Thread: preamp design
View Single Post
  #9   Report Post  
Scott Dorsey
 
Posts: n/a
Default preamp design

David Grant wrote:
That's the first hurdle. A good understanding of differential circuitry,
feedback methods and noise calculations takes you further. Then you get

into
understanding device parameters. Does rbb mean anything to you for example

?

We've covered differential BJT amps, cascode configuration, some frequency
response, feedback (shunt/series sampling) and some stability. I remember
some stuff on calculating CMRR if that's what you mean by noise
calculations. As for the device parameters, I'm not sure. Rbb sounds like
the symbol one of my proffs used to denote the base resistance used for
biasing a BJT, but otherwise it doesn't mean anything to me.


Okay, your typical preamp is a differential BJT amp, with something after
it to give it a little bit more gain.

OR, it's a transformer with some single-ended gain stages.

We did all the derivations of gain, in/out resistance for most of the amps
we discussed in class, and usually everything simplified down to something
neat and tidy. I've tried to do small signal analysis on some of the mic-pre
schematics I've looked at, but I always end up with huge long expressions
that turn very messy. I also never end up with the quoted specs in the end!
It seems like there must be some simplification that I'm missing. Common
mode gain control for example... what on earth is that (apart from a gain
control)?


Basically you throw out all the small terms.... the input stage noise floor
is so high that it swamps everything else. There was a paper by Marshall
Leach in the JAES a decade ago which went through all the noise floor math
for a dozen different preamp configurations, in a clear and concise way.
You need to read it. Everybody needs to read it.

Maybe I just need more experience with this stuff, but I'm not sure where to
find that.


Go to your school library, get old copies of the JAES, and read through
them. Also get the RCA Radiotron Handbook, which is thoroughly obsolete
and talks about tubes, but the math is the same and the circuits can generally
be stolen outright and implemented with FETs.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."