Improved mic amp posted on Rapidshare

[Eeyore]

By popular request I've posted my latest version of an improved microphone
preamplifier here.
http://rapidshare.com/files/21272377/mic_amp_2.jpg

As it seems some can't see it on abse I've reposted it there too as 'improved
mic amp'

Graham

 

[tempus fugit]

By popular request I've posted my latest version of an improved microphone
preamplifier here.
http://rapidshare.com/files/21272377/mic_amp_2.jpg

As it seems some can't see it on abse I've reposted it there too as 'improved
mic amp'

Graham

Thanks posting your design Graham. You obviously actually work in this
field, so I was hoping I could ask some questions regarding your design; or
design in general.

How would this design compare with one of the all-in-one ICs (like
http://www.analog.com/en/prod/0,2877,SSM2019,00.html)? Would they 'sound'
the same? I ask this for 2 reasons - one, I'm using a micpre that I made
using that IC, and two, if there performance ans sound is close, wouldn't it
be easier and cheaper just to use the IC? Is there something better about
the discrete approach?
I'm sure you're aware of much higher priced micpres like Focusrite, etc. Is
their design so radically different (and more expensive to implement)from
the type of design you have here, or is the difference in cost more due to
audiofoolery?
Along that same line, I've never had the $ to be able to use one of those
hi-end pres, so I can't make a direct comparison myself, but is there
something in the design of those really expensive micpres that makes them
more "musical" or appealing to the ears of musicians (I ask this because you
work in the field and so assume that you've heard some of these others, or
done some comparisons yourself)?

Thanks again

 

[Fred Bloggs]

By popular request I've posted my latest version of an improved microphone
preamplifier here.
http://rapidshare.com/files/21272377/mic_amp_2.jpg

As it seems some can't see it on abse I've reposted it there too as 'improved
mic amp'

Graham

Finally got it- interesting, but I don't see how the operating points of
the pnp's are well defined.

 

[Eeyore]

Finally got it- interesting, but I don't see how the operating points of
the pnp's are well defined.

Their bases are close to ground. Ic (well Ie) is predominantly set by the 4k7 tail
resistors which in turn sets the collectors' DC operating point at V- + (Ic*1k5)

Graham

 

[Fred Bloggs]

Fred Bloggs wrote:




Their bases are close to ground. Ic (well Ie) is predominantly set by
the 4k7 tail resistors which in turn sets the collectors' DC
operating point at V- + (Ic*1k5)

That much is obvious but your differential feedback only forces equality
and not any particular Ic,Q for the pnp's. So where is the magic I have
missed? Did you ever build that circuit?

 

[Eeyore]

That much is obvious but your differential feedback only forces equality
and not any particular Ic,Q for the pnp's.

Is Ie not good enough for you ? The emitters are connected to -17V by 4k7 and 0V
by 6k8. That makes the tail current set by V = -17/(6k8/6k8+4k7) = -10.05V
divided by 6k8 in parallel with 4k7 = 2k78. Ie = 10.05/2k78 = 3.62 mA.

So where is the magic I have missed? Did you ever build that circuit?

That specific one with all those values ? No. It had swiched resistor values to
set the gain in a rather perverse change from the norm. I did another one with
some differnt values long ago though and yet another in 1988 with a rather more
interesting feedback arrangement. Problem ?

Graham

 

[Fred Bloggs]

Fred Bloggs wrote:




Is Ie not good enough for you ? The emitters are connected to -17V by 4k7 and 0V
by 6k8. That makes the tail current set by V = -17/(6k8/6k8+4k7) = -10.05V
divided by 6k8 in parallel with 4k7 = 2k78. Ie = 10.05/2k78 = 3.62 mA.

Are we looking at the same circuit? You have the 4k7 going from the
emitter /node/ to V+. That only sets the current into the /node/ which
is also drained by the 6.8K ckt, it does not set Ie for the pnp.

That specific one with all those values ? No.

I thought so...

 

[Fred Bloggs]

Problem ?

I see where you're coming from now. First we contrive a circuit
configuration with indeterminate operating point, then we force it to
behave by our desires. This is a very enchanting mix of mystery,
confusion, and emotional investment.

 

[Winfield Hill]

Is Ie not good enough for you ? The emitters are connected to
-17V by 4k7 and 0V by 6k8. That makes the tail current set by
V = -17/(6k8/6k8+4k7) = -10.05V divided by 6k8 in parallel
with 4k7 = 2k78. Ie = 10.05/2k78 = 3.62 mA.

You mean +17V, not -17V? Where's Vbe in your calculation?

 

[Eeyore]

Are we looking at the same circuit? You have the 4k7 going from the
emitter /node/ to V+.
Yes.


That only sets the current into the /node/ which
is also drained by the 6.8K ckt, it does not set Ie for the pnp.

The other end of the 6k8 is at 0V.

Can't you do the Norton Thevenin thing ? I've *already* posted the answer btw.

Graham

 

[Eeyore]

I see where you're coming from now.

AT LAST !!! Wonders will never cease.

First we contrive a circuit configuration with indeterminate operating point

Only in your mind.

then we force it to behave by our desires. This is a very enchanting mix of
mystery,
confusion, and emotional investment.

I believe the phrase you're looking for is "negative feedback" !

Graham

 

[Eeyore]

You mean +17V, not -17V?

Sorry. Mea Culpa. You're right of course.

Where's Vbe in your calculation?

I ignored it for simplification.

You're correct of course.

It's effectively (16.4/4.7k) + (-0.6/6k8). That should do the Norton Thevenin
thing in one hit.

= 3.40 mA - 88uA. = ~ 3.3 mA.

Graham

 

[Jim Thompson]

I see where you're coming from now. First we contrive a circuit
configuration with indeterminate operating point, then we force it to
behave by our desires. This is a very enchanting mix of mystery,
confusion, and emotional investment.

Your levitation not working today, Fred ?

...Jim Thompson

 

[Fred Bloggs]

Your levitation not working today, Fred ?

...Jim Thompson

Apparently Graham is unaware of the fact that transistors in the active
region develop a Vce. His circuit, which he admits never existed in
hardware, is indeterminate.

 

[Eeyore]

Apparently Graham is unaware of the fact that transistors in the active region
develop a Vce. His circuit, which he admits never existed in hardware,

You mean the pot was replaced by fixed resistors ?

is indeterminate.

Bwahahahahahaha Why don't you build it and come back and say that with a
straight face ?

The fact is that you've been 'out-thunk' by a mere audio practicioner.

Graham

 

[Fred Bloggs]

Fred Bloggs wrote:




You mean the pot was replaced by fixed resistors ?





Bwahahahahahaha Why don't you build it and come back and say that with a
straight face ?

The fact is that you've been 'out-thunk' by a mere audio practicioner.

Graham

Get real, Graham, those NJM's are railed or something, but you have no
clue what the circuit bias is. You did sucker me in for a few
milliseconds thinking you might have had something, but your remarks
have since lost you that status. From now on we will all be
....errr...skeptical of giving you serious consideration.

 

[Eeyore]

Get real, Graham, those NJM's are railed or something,

Nope. That circuit works a treat.

but you have no clue what the circuit bias is.

What you mean is that *you* can't do a DC analysis !

You did sucker me in for a few
milliseconds thinking you might have had something, but your remarks
have since lost you that status. From now on we will all be
...errr...skeptical of giving you serious consideration.

How about you build the circuit ?

Graham

 

[Fred Bartoli]

Fred Bloggs a écrit :

Get real, Graham, those NJM's are railed or something, but you have no
clue what the circuit bias is. You did sucker me in for a few
milliseconds thinking you might have had something, but your remarks
have since lost you that status. From now on we will all be
....errr...skeptical of giving you serious consideration.

in fact this circuit will work, but I'd like to hear Graham's
explanation

Details please, instead of 'did you heard of feedback'.

 

[Eeyore]

Fred Bloggs a écrit :

in fact this circuit will work,

It *does* work and is widely used.

but I'd like to hear Graham's explanation

Details please, instead of 'did you heard of feedback'.

Simplistically, the discrete input devices provide low-noise gain to 'front-end'
IC1. IC2 simply inverts the signal so that differential feedback can be applied
to the emitters.

How much explanation do you need ?

Graham

 

[YD]

Late at night, by candle light, Fred Bartoli

Fred Bloggs a écrit :

in fact this circuit will work, but I'd like to hear Graham's
explanation

Details please, instead of 'did you heard of feedback'.

This tip-toeing around the issue is getting boring fast. Everyone
giving "hints" to everyone else and expecting them to come up with a
complete analysis. I suppose it's because no one really is sure and is
afraid of making a fool of himself. That, or hopes the other will make
a mistake and get something to jump on. This is no way to have a
technical discussion. Too many divas in here.

- YD.