Alice is a popular version of the Schoeps circuit by Scott Helmke; probably
because its simple and works well.

But there is an important omission. The Alice versions do not have the zeners, 
D3 & D4 in Schoeps.pdf

http://tech.groups.yahoo.com/group/micbuilders/files/Schematics/

Many people think that if they use high voltage BJTs, there is no need for
these.  But like most of Dip. Ing. Wuttke's stuff, every item does at least 2
things.

If a the mike cable is shorted, the supply capacitors in the mike discharge
through the BJTs and zener their Vbe.

The mike still works but is now very noisy.  The zeners will forward conduct &
prevent this from happening.

So even if you use high voltage BJTs, you need diodes at these positions; 1n4148
will do.  Use zeners if you have low voltage transistors.

Could I ask those of you who have Alice / Schoeps circuits in our Files to
review them and put in the diodes if missing.

Zapnspark's stuff is all correct but RichPeet has a copy of the original Alice
without diodes.

There is also another part omission from the Alice circuit.
The original Schoeps circuit has a 47 uF cap from the common collector junction
of the BJTs (PNP output transistors) to ground.
That cap reduces the output impedance of the mic, increases cable drive and
reduces some overall distortion.
That 47 uF cap omission has been discussed here in some much older posts at the
Micbuilders group.

Cheers.

Jim G.
ZAPNSPARK

--- In micbuilders@yahoogroups.com, Richard Lee <ricardo@...> wrote:
>
> Alice is a popular version of the Schoeps circuit by Scott Helmke; probably
because its simple and works well.
>
> But there is an important omission. The Alice versions do not have the zeners,
D3 & D4 in Schoeps.pdf
>
> http://tech.groups.yahoo.com/group/micbuilders/files/Schematics/
>
> Many people think that if they use high voltage BJTs, there is no need for
these.  But like most of Dip. Ing. Wuttke's stuff, every item does at least 2
things.
>
> If a the mike cable is shorted, the supply capacitors in the mike discharge
through the BJTs and zener their Vbe.
>
> The mike still works but is now very noisy.  The zeners will forward conduct &
prevent this from happening.
>
> So even if you use high voltage BJTs, you need diodes at these positions;
1n4148 will do.  Use zeners if you have low voltage transistors.
>
> Could I ask those of you who have Alice / Schoeps circuits in our Files to
review them and put in the diodes if missing.
>
> Zapnspark's stuff is all correct but RichPeet has a copy of the original Alice
without diodes.
>

Thanks for this Jim.

The 47u cap (C9 in Schoeps.pdf) does indeed improve performance but I didn't
want to re-design everyone else's circuit.  There's no end to this and you only
get a name for being a circuit Nazi.

The reason I'm making an issue of the diodes is that it can result in really bad
noise under mysterious circumstances.

I'd like everything on MicBuilders to at least "appear" good and more
importantly, be reliable.

Hence my urging people to clearly mark "Experimental" circuits.  Your directory
is a model of how it should be done.

I'm one of the guilty parties... though I don't think I actually have a
schematic posted here.

But I do have a couple comments about this.  In general I'm in favor of
these changes (or restorations, perhaps), though I mostly don't actually
use them in my own mics.  I've got the capacitor in a few mics, but
don't have the diodes anywhere.

The cap may well have been the difference between success and failure on
a gig or two - funky house splitter snake leading to the Alice mics
being unusable at one show.  Haven't seen anything like that since I
added the cap to my own show Alice/Audrey mics.  So that's on my todo
list to add to the standard circuit for new mics.  For existing models,
well, do I make a change that may affect the sound?  "Beta Alice",
maybe?  Something to consider if I go back to building mics for money
instead of as a hobby that generates roughly enough money to pay for
itself but no profit.

I've never had a problem with noise from damaged transistors, aside from
one or two possible cases when powering up new mics for the first time.
   Maybe bad transistor instead of damage?  I dunno.  But there are
something like 40-50 of my mics out there in the wild (either in my own
working collection or ones I've sold to others), and I can't recall any
of them ever coming back for repair.  Granted most of them likely belong
to studio people who are careful, but I've got at least one guy touring
with several mics and no reported problems.  At this point I'm not in a
big hurry to add those protection diodes - I'll likely save that for a
new production model, whenever that might happen.

-Scott

Richard wrote:
> Thanks for this Jim.
>
> The 47u cap (C9 in Schoeps.pdf) does indeed improve performance but I didn't
want to re-design everyone else's circuit.  There's no end to this and you only
get a name for being a circuit Nazi.
>
> The reason I'm making an issue of the diodes is that it can result in really
bad noise under mysterious circumstances.
>
> I'd like everything on MicBuilders to at least "appear" good and more
importantly, be reliable.
>
> Hence my urging people to clearly mark "Experimental" circuits.  Your
directory is a model of how it should be done.
>
>
>
>
>
>
> ------------------------------------
>
> Yahoo! Groups Links
>
>
>
>

--
---- Scott Helmke ---- scott@... ---- (734) 604-9340 ----
And you'll visualize not taking any chances
But meet them halfway with love, peace, and persuasion
And expect them to rise to the occasion
(from "Glad Tidings", by Van Morrison)

the transistors i have always used are matched bc560s they are wonderful and
cheap but have had several zener breakdowns. when i did not at first know what
the reason for the suddenly noisy mic was, i tried to improve the voltage
regulator zener, thinking it was coming from there! now i always have zeners
(5v6)

umashankar
 i have published my poems. you can read (or buy) at
http://stores.lulu.com/umashankar





________________________________
From: Scott Helmke <scott@...>
To: micbuilders@yahoogroups.com
Sent: Mon, May 2, 2011 8:26:36 AM
Subject: Re: [micbuilders] Re: Missing Diodes in Alice circuits

 
I'm one of the guilty parties... though I don't think I actually have a
schematic posted here.

But I do have a couple comments about this. In general I'm in favor of
these changes (or restorations, perhaps), though I mostly don't actually
use them in my own mics. I've got the capacitor in a few mics, but
don't have the diodes anywhere.

The cap may well have been the difference between success and failure on
a gig or two - funky house splitter snake leading to the Alice mics
being unusable at one show. Haven't seen anything like that since I
added the cap to my own show Alice/Audrey mics. So that's on my todo
list to add to the standard circuit for new mics. For existing models,
well, do I make a change that may affect the sound? "Beta Alice",
maybe? Something to consider if I go back to building mics for money
instead of as a hobby that generates roughly enough money to pay for
itself but no profit.

I've never had a problem with noise from damaged transistors, aside from
one or two possible cases when powering up new mics for the first time.
Maybe bad transistor instead of damage? I dunno. But there are
something like 40-50 of my mics out there in the wild (either in my own
working collection or ones I've sold to others), and I can't recall any
of them ever coming back for repair. Granted most of them likely belong
to studio people who are careful, but I've got at least one guy touring
with several mics and no reported problems. At this point I'm not in a
big hurry to add those protection diodes - I'll likely save that for a
new production model, whenever that might happen.

-Scott

Richard wrote:
> Thanks for this Jim.
>
> The 47u cap (C9 in Schoeps.pdf) does indeed improve performance but I didn't
>want to re-design everyone else's circuit. There's no end to this and you only
>get a name for being a circuit Nazi.
>
> The reason I'm making an issue of the diodes is that it can result in really
>bad noise under mysterious circumstances.
>
> I'd like everything on MicBuilders to at least "appear" good and more
>importantly, be reliable.
>
> Hence my urging people to clearly mark "Experimental" circuits. Your directory
>is a model of how it should be done.
>
>
>
>
>
>
> ------------------------------------
>
> Yahoo! Groups Links
>
>
>
>

--
---- Scott Helmke ---- scott@... ---- (734) 604-9340 ----
And you'll visualize not taking any chances
But meet them halfway with love, peace, and persuasion
And expect them to rise to the occasion
(from "Glad Tidings", by Van Morrison)



[Non-text portions of this message have been removed]

Hello Micbuilders

I use BC560  too and one of my Alice mic became noisy after a while (without
zeners) . I have to change the BJT .

  Does zeners will had distorsion ? any idea?

Francis





________________________________
From: umashankar <umashanks@...>
To: micbuilders@yahoogroups.com
Sent: Mon, May 2, 2011 5:28:44 AM
Subject: Re: [micbuilders] Re: Missing Diodes in Alice circuits


the transistors i have always used are matched bc560s they are wonderful and
cheap but have had several zener breakdowns. when i did not at first know what
the reason for the suddenly noisy mic was, i tried to improve the voltage
regulator zener, thinking it was coming from there! now i always have zeners
(5v6)

umashankar
  i have published my poems. you can read (or buy) at
http://stores.lulu.com/umashankar

________________________________
From: Scott Helmke <scott@...>
To: micbuilders@yahoogroups.com
Sent: Mon, May 2, 2011 8:26:36 AM
Subject: Re: [micbuilders] Re: Missing Diodes in Alice circuits


I'm one of the guilty parties... though I don't think I actually have a
schematic posted here.

But I do have a couple comments about this. In general I'm in favor of
these changes (or restorations, perhaps), though I mostly don't actually
use them in my own mics. I've got the capacitor in a few mics, but
don't have the diodes anywhere.

The cap may well have been the difference between success and failure on
a gig or two - funky house splitter snake leading to the Alice mics
being unusable at one show. Haven't seen anything like that since I
added the cap to my own show Alice/Audrey mics. So that's on my todo
list to add to the standard circuit for new mics. For existing models,
well, do I make a change that may affect the sound? "Beta Alice",
maybe? Something to consider if I go back to building mics for money
instead of as a hobby that generates roughly enough money to pay for
itself but no profit.

I've never had a problem with noise from damaged transistors, aside from
one or two possible cases when powering up new mics for the first time.
Maybe bad transistor instead of damage? I dunno. But there are
something like 40-50 of my mics out there in the wild (either in my own
working collection or ones I've sold to others), and I can't recall any
of them ever coming back for repair. Granted most of them likely belong
to studio people who are careful, but I've got at least one guy touring
with several mics and no reported problems. At this point I'm not in a
big hurry to add those protection diodes - I'll likely save that for a
new production model, whenever that might happen.

-Scott

Richard wrote:
> Thanks for this Jim.
>
> The 47u cap (C9 in Schoeps.pdf) does indeed improve performance but I didn't
>want to re-design everyone else's circuit. There's no end to this and you only
>get a name for being a circuit Nazi.
>
> The reason I'm making an issue of the diodes is that it can result in really
>bad noise under mysterious circumstances.
>
> I'd like everything on MicBuilders to at least "appear" good and more
>importantly, be reliable.
>
> Hence my urging people to clearly mark "Experimental" circuits. Your directory

>is a model of how it should be done.
>
>
>
>
>
>
> ------------------------------------
>
> Yahoo! Groups Links
>
>
>
>

--
---- Scott Helmke ---- scott@... ---- (734) 604-9340 ----
And you'll visualize not taking any chances
But meet them halfway with love, peace, and persuasion
And expect them to rise to the occasion
(from "Glad Tidings", by Van Morrison)

[Non-text portions of this message have been removed]




[Non-text portions of this message have been removed]

I have used 5.6v-6.8v zeners in my mic builds for protection, and there is no
apparent distortion from them, and neither should there be, because as they are
wired,they have a very high impedance compared to the low impedance output. I
have found the BC560Cs work really well as emitter followers. I always beta
match them as closely as possible, using the hFE socket on my LCR meter.

Kindest regards,

zephyrmic


--- In micbuilders@yahoogroups.com, francis dupuis <vitaeear@...> wrote:
>
> Hello Micbuilders
>
> I use BC560  too and one of my Alice mic became noisy after a while (without
> zeners) . I have to change the BJT .
>
>  Does zeners will had distorsion ? any idea?
>
> Francis
>
>
>
>
>
> ________________________________
> From: umashankar <umashanks@...>
> To: micbuilders@yahoogroups.com
> Sent: Mon, May 2, 2011 5:28:44 AM
> Subject: Re: [micbuilders] Re: Missing Diodes in Alice circuits
>
>
> the transistors i have always used are matched bc560s they are wonderful and
> cheap but have had several zener breakdowns. when i did not at first know what
> the reason for the suddenly noisy mic was, i tried to improve the voltage
> regulator zener, thinking it was coming from there! now i always have zeners
> (5v6)
>
> umashankar
>  i have published my poems. you can read (or buy) at
> http://stores.lulu.com/umashankar
>
> ________________________________
> From: Scott Helmke <scott@...>
> To: micbuilders@yahoogroups.com
> Sent: Mon, May 2, 2011 8:26:36 AM
> Subject: Re: [micbuilders] Re: Missing Diodes in Alice circuits
>
>
> I'm one of the guilty parties... though I don't think I actually have a
> schematic posted here.
>
> But I do have a couple comments about this. In general I'm in favor of
> these changes (or restorations, perhaps), though I mostly don't actually
> use them in my own mics. I've got the capacitor in a few mics, but
> don't have the diodes anywhere.
>
> The cap may well have been the difference between success and failure on
> a gig or two - funky house splitter snake leading to the Alice mics
> being unusable at one show. Haven't seen anything like that since I
> added the cap to my own show Alice/Audrey mics. So that's on my todo
> list to add to the standard circuit for new mics. For existing models,
> well, do I make a change that may affect the sound? "Beta Alice",
> maybe? Something to consider if I go back to building mics for money
> instead of as a hobby that generates roughly enough money to pay for
> itself but no profit.
>
> I've never had a problem with noise from damaged transistors, aside from
> one or two possible cases when powering up new mics for the first time.
> Maybe bad transistor instead of damage? I dunno. But there are
> something like 40-50 of my mics out there in the wild (either in my own
> working collection or ones I've sold to others), and I can't recall any
> of them ever coming back for repair. Granted most of them likely belong
> to studio people who are careful, but I've got at least one guy touring
> with several mics and no reported problems. At this point I'm not in a
> big hurry to add those protection diodes - I'll likely save that for a
> new production model, whenever that might happen.
>
> -Scott
>
> Richard wrote:
> > Thanks for this Jim.
> >
> > The 47u cap (C9 in Schoeps.pdf) does indeed improve performance but I didn't
> >want to re-design everyone else's circuit. There's no end to this and you
only
> >get a name for being a circuit Nazi.
> >
> > The reason I'm making an issue of the diodes is that it can result in really
> >bad noise under mysterious circumstances.
> >
> > I'd like everything on MicBuilders to at least "appear" good and more
> >importantly, be reliable.
> >
> > Hence my urging people to clearly mark "Experimental" circuits. Your
directory
>
> >is a model of how it should be done.
> >
> >
> >
> >
> >
> >
> > ------------------------------------
> >
> > Yahoo! Groups Links
> >
> >
> >
> >
>
> --
> ---- Scott Helmke ---- scott@... ---- (734) 604-9340 ----
> And you'll visualize not taking any chances
> But meet them halfway with love, peace, and persuasion
> And expect them to rise to the occasion
> (from "Glad Tidings", by Van Morrison)
>
> [Non-text portions of this message have been removed]
>
>
>
>
> [Non-text portions of this message have been removed]
>

> I've got the capacitor in a few mics, but don't have the diodes anywhere. ...

> The cap may well have been the difference between success and failure on a gig
or two - funky house splitter snake leading to the Alice mics being unusable at
one show. ...

> I've never had a problem with noise from damaged transistors, aside from one
or two possible cases when powering up new mics for the first time.

Actually the cap & "noisy BJT after a line short" are related. The 47u at the
collector is the one which reverse zaps Vbe.  If you haven't got one, the 0u47
base caps are 100x smaller. The other supply electrolytics do some damage but
there's a resistor & less volts.

The most dangerous situation is with 47u at collectors, no diodes and low
voltage transistors like BC560.

The zeners don't introduce distortion under normal working.

If you NEVER have line shorts, you don't have to worry about all this. Usual
cause is badly soldered & assembled XLRs.

If it happens, what you notice is the mike is suddenly permanently noisy for no
apparent reason but otherwise, works OK.

if you're concerned about the control node reverse breakdown voltage can't you
use a JFET instead? 


________________________________
From: Richard <ricardo@...>
To: micbuilders@yahoogroups.com
Sent: Monday, May 2, 2011 6:39 AM
Subject: [micbuilders] Re: Missing Diodes in Alice circuits


 
> I've got the capacitor in a few mics, but don't have the diodes anywhere. ...

> The cap may well have been the difference between success and failure on a gig
or two - funky house splitter snake leading to the Alice mics being unusable at
one show. ...

> I've never had a problem with noise from damaged transistors, aside from one
or two possible cases when powering up new mics for the first time.

Actually the cap & "noisy BJT after a line short" are related. The 47u at the
collector is the one which reverse zaps Vbe.  If you haven't got one, the 0u47
base caps are 100x smaller. The other supply electrolytics do some damage but
there's a resistor & less volts.

The most dangerous situation is with 47u at collectors, no diodes and low
voltage transistors like BC560.

The zeners don't introduce distortion under normal working.

If you NEVER have line shorts, you don't have to worry about all this. Usual
cause is badly soldered & assembled XLRs.

If it happens, what you notice is the mike is suddenly permanently noisy for no
apparent reason but otherwise, works OK.




[Non-text portions of this message have been removed]

Mike Rooke wrote:

> how to measure the impedance of a transformer
I'd mainly be interested in measuring the frequency response and gain of the
transformer in the application circuit.  That is, the transformer should be
driven by something less than 1 Ohm in the case of a transformer for a ribbon
microphone.  I have the figure of 0.22 Ohms for the ribbon in the BBC PGS
ribbon
microphone.  It would be easy to simply load the output of the generator by a
0.5 or 0.22 Ohm reisistor.  That would get the generator levels down nice and
low where you want them.

> For one experiment I cut a ribbon from a coffee packet
What I didn't say before (Mike and I have been corresponding off-list) is that
this is a cool idea.  I am assuming that you are trying to create a
high-durability ribbon, one that may perhaps be easier to apply crimps. 


It's my feeling that transformers probably only constitute a fair solution to
the ribbon microphone problem.  In my last post I came up with the figure of
-84
to -87 dBV/Pa for the sensitivity at the output of the ribbon.  I at least half
expected that someone would write and tell me that I was wrong, but apparently
not ... Since then I've looked at a fair number of ribbon microphone data sheets
and of modern ribbon microphones, the Royer 121 was most sensitive at -50
dBV/Pa
and the Nady RSM-5 was least sensitive at -60 dBV/Pa with the bulk of them being
around -55 dBV/Pa, implying that the sensitivity of the ribbon is -85 dBV/Pa. 
It's certainly believable that Royer would do a better job than most on making
the microphone more sensitivie.

I'm still thinking that the task of constructing a preamp that would directly
amplify the ribbon output with appropriate noise and distortion is only mostly
impossible (c.f. "The Princess Bride").  It would be a very interesting
project.  The 2SB737 transistor that I mentioned is essentially unavailable but
the 2SC3329 listed in Richard's preamp circuit is available, and it's cheap. 
The Toshiba spec sheet says that Rbb is 2 Ohms.  I find myself doubting all of
these spec sheets, but it should be relatively easy to measure.

> Antistatic tweezers are rather handy for moving delicate ribbons around
Are these the plastic ones?

Eric




________________________________
From: Mike Rooke <yg@...>
To: micbuilders@yahoogroups.com
Sent: Fri, April 29, 2011 12:44:30 PM
Subject: [micbuilders] Re: circuit for active ribbon mic

 
http://www.crystal-radio.eu/entrafometing.htm

Shows how to measure the impedance of a transformer - but is this also
applicable to a ribbon transformer? since the primary (which I assume is the
ribbon) will be much lower impedance than the signal generators 50 ohms?


I have a sacrificial T.Bone RB100 I've been "playing with" lately - and would
like to (try) and measure the primary / secondary impedance and resistance.


For one experiment I cut a ribbon from a coffee packet. This is PET/AL/PE,
basically remove/peel away the PE layer revealing the paint/logo - use acetone
to remove this and then if needed a soldering iron to remove some of the PET
mass (with plenty of ventilation) - this ribbon was around 0.3 ohms but I
suspect it was > 10um - Sensitivity was 10dB lower than a ribbon made from leaf,
perhaps 1um or less - I doubt theres much benefit in the coffee packet ribbon
since the stock mic's with a ribbon re-tension Im using outdoors covered with
Rycote large condensor foam's - so far they are not too sensitive to wind
blasts. Two of them sound rather nice in a 20cm faulker array.


Antistatic tweezers are rather handy for moving delicate ribbons around, plus a
deep breath.

BR
MIke.

--- In micbuilders@yahoogroups.com, "LESLIE WATTS" <leswatts@...> wrote:
>
> Never mind the offset on the transformer...I just misread the graph. Too many
>squiggly lines. Now to muck up the transistors...
>
> Les
> ----- Original Message -----
> From: LESLIE WATTS
> To: micbuilders@yahoogroups.com
> Sent: Friday, April 29, 2011 10:02 AM
> Subject: Re: [micbuilders] Re: circuit for active ribbon mic
>
>
>
> Neat!
> But really...I wish you'd stop holding back on using phantom current...use all
>you want!! :^D
>
> I threw it up on quickie 5spice with 2N4401...will get the other transistor
>models.
>
> Can't do dist on freebie 5spice but I have a full blown Protel seat.
> Just hate to use it. Will later though.
>
> Looks pretty good...pretty clean. Plenty of output. I am seeing 100mV DC
across
>the transformer though. Perhaps I screwed up something. It's just a 10 Henry
/10
>Henry with 1 ohm winding R fed by a voltage source in series with 300 ohm. And
a
>1G on the secondary to ground to keep Mr. Spice from blowing up.
>
> Let me check and see if I messed up...I'll want to muck up the transistor
betas
>etc and see what happens too.
>
> Les
> ----- Original Message -----
> From: Richard
> To: micbuilders@yahoogroups.com
> Sent: Thursday, April 28, 2011 9:38 PM
> Subject: [micbuilders] Re: circuit for active ribbon mic
>
> I've put a very headbanging circuit in my Ricardo Files directory.
>
> [Non-text portions of this message have been removed]
>
>
>
>
>
> [Non-text portions of this message have been removed]
>




[Non-text portions of this message have been removed]

Ok... All these circuits are very good and I've now got a lot of my own active
circuits together including some with JFETs, Opamps, and BJTs all mixed
together...  They work well in simulation anyway.

Since I'm not anywhere near a studio or ribbon mic I've got some questions...

How much gain do you need after the transformer?

If you could have all the gain you might need, what would be a reasonable S/N
ratio?

Ron, the issue isn't S/N but how close you can get to a noiseless amp. Ribbons
like all passive mikes, have a signal power output. There is also a noise power
which for a given resistance R, translates to

0.129nV rt(f R)   f: frequency bandwidth considered

A nominal 300R ribbon will have noise over 20kHz
0.129nV rt(20000 x 300) = 315nV which is -130dBV

Even with a noiseless amp, you would still see this -130dBV noise as it is
inherent in the resistance. If you saw -128dBV Equivalent Input Noise (output
noise divided by gain) through your amp, it would only have degraded this
inherent noise by 2dB, a NOISE FACTOR of 2dB.  This would be an excellent amp.
as it gets to within 2dB of a noiseless amp.

The best I've done is 1.5 - 1.7dB NF for source 200R in a Calrec Mixer Mike
input with a Lundahl transformer & NE5534.

A dedicated ribbon input using 1510 should be able to get near to 1dB NF.

The noise performance of an amp is usually represented by a Ein and a noise
current, In. You want Ein to be smaller than the source voltage noise and In to
be smaller than the source current noise.

Amps have a ideal source resistance where they give the best NF.  This is given
by    Rideal = Ein / In

In = 0.129nA rt(f / R)

If you can use a transformer, you try to transform the source resistance to
somewhere near Rideal for the following amp.  For a 5534, this is about 8k.  The
transformer needs to introduce minimal losses & noise too.

So the answer to part of your question is

NF of better than 2dB with the typical nominal 300R ribbon.
____________________________

> In my last post I came up with the figure of -84 to -87 dBV/Pa for the
sensitivity at the output of the ribbon. ..

> Since then I've looked at a fair number of ribbon microphone data and of
modern ribbon microphones, the Royer 121 was most sensitive at -50 dBV/Pa and
the Nady RSM-5 was least sensitive at -60 dBV/Pa with the bulk of them being
around -55 dBV/Pa, implying that the sensitivity of the ribbon is -85 dBV/Pa. 
It's certainly believable that Royer would do a better job than most on making
the microphone more sensitive.

Eric, have you got a good sensitivity for Coles 4038?  I dun unnerstan dis
dynes/cm stuff in the BBC paper though I think that has a 300R transformer like
the Coles.  Most STC versions were nominal 30R.  If we believe Coles, their 4038
is -65dBV/Pa with nominal 300R

The answer to 1st part of Ron's question is "How loud do you want it to be?"  I
think the upper limit is Schoeps MK2 at 15mV/Pa which is
-36.5dBV/Pa

This would need 19dB gain to boost the usual -55dBV/Pa ribbon to the same
loudness as MK2.

> I'm still thinking that the task of constructing a preamp that would directly
amplify the ribbon output with appropriate noise and distortion ...

> The 2SB737 transistor that I mentioned is essentially unavailable but the
2SC3329 listed in Richard's preamp circuit is available, and it's cheap. The
Toshiba spec sheet says that Rbb is 2 Ohms.

My Moving Coil preamp in Files is good for testing rbb but only for
complementary pairs.  You can up the gain easily to 1000x to help.

I've never tested 2sc3329 but 2sc2546/2sa1084 at 3mA in my circuit measures
Rnv=5R.  This is average rbb of 5R7 for the pair. No improvement beyond 3mA
which ties in.

Doing this little exercise has put me off trying to do a transformerless ribbon.
20x more devices & current is 60mA in my Moving Coil Preamp and only gives Rnv
0R25; still more noise than the ribbon.

From playing with Ribbon.gif & Ribbon2.gif, I'm nervous about having this large
current near an unbuffered ribbon

Ricardo wrote:
> have you got a good sensitivity for Coles 4038
From their own web site, I have -65 dBV/Pa It's a little difficult to read
because of the way that they write it, but based on the Coles web site and on
the data sheet from AEA:
http://www.audiorents.com/library/specsheets/5_AEA_Coles4038.pdf

I think that's right.  It makes sense that it's a little less sensitive than
the
more modern microphones simply due to the Neodymium magnets.

>this little exercise has put me off trying to do a transformerless ribbon
Aw, come on.  Where's your sense of adventure?



________________________________
From: Richard <ricardo@...>
To: micbuilders@yahoogroups.com
Sent: Mon, May 2, 2011 8:41:35 PM
Subject: [micbuilders] Re: circuit for active ribbon mic

 
Ron, the issue isn't S/N but how close you can get to a noiseless amp. Ribbons
like all passive mikes, have a signal power output. There is also a noise power
which for a given resistance R, translates to

0.129nV rt(f R) f: frequency bandwidth considered

A nominal 300R ribbon will have noise over 20kHz
0.129nV rt(20000 x 300) = 315nV which is -130dBV

Even with a noiseless amp, you would still see this -130dBV noise as it is
inherent in the resistance. If you saw -128dBV Equivalent Input Noise (output
noise divided by gain) through your amp, it would only have degraded this
inherent noise by 2dB, a NOISE FACTOR of 2dB. This would be an excellent amp. as
it gets to within 2dB of a noiseless amp.

The best I've done is 1.5 - 1.7dB NF for source 200R in a Calrec Mixer Mike
input with a Lundahl transformer & NE5534.

A dedicated ribbon input using 1510 should be able to get near to 1dB NF.

The noise performance of an amp is usually represented by a Ein and a noise
current, In. You want Ein to be smaller than the source voltage noise and In to
be smaller than the source current noise.

Amps have a ideal source resistance where they give the best NF. This is given
by Rideal = Ein / In

In = 0.129nA rt(f / R)

If you can use a transformer, you try to transform the source resistance to
somewhere near Rideal for the following amp. For a 5534, this is about 8k. The
transformer needs to introduce minimal losses & noise too.

So the answer to part of your question is

NF of better than 2dB with the typical nominal 300R ribbon.
____________________________

> In my last post I came up with the figure of -84 to -87 dBV/Pa for the
>sensitivity at the output of the ribbon. ..

> Since then I've looked at a fair number of ribbon microphone data and of
modern
>ribbon microphones, the Royer 121 was most sensitive at -50 dBV/Pa and the
Nady
>RSM-5 was least sensitive at -60 dBV/Pa with the bulk of them being around -55
>dBV/Pa, implying that the sensitivity of the ribbon is -85 dBV/Pa. It's
>certainly believable that Royer would do a better job than most on making the
>microphone more sensitive.

Eric, have you got a good sensitivity for Coles 4038? I dun unnerstan dis
dynes/cm stuff in the BBC paper though I think that has a 300R transformer like
the Coles. Most STC versions were nominal 30R. If we believe Coles, their 4038
is -65dBV/Pa with nominal 300R

The answer to 1st part of Ron's question is "How loud do you want it to be?" I
think the upper limit is Schoeps MK2 at 15mV/Pa which is

-36.5dBV/Pa

This would need 19dB gain to boost the usual -55dBV/Pa ribbon to the same
loudness as MK2.

> I'm still thinking that the task of constructing a preamp that would directly
>amplify the ribbon output with appropriate noise and distortion ...

> The 2SB737 transistor that I mentioned is essentially unavailable but the
>2SC3329 listed in Richard's preamp circuit is available, and it's cheap. The
>Toshiba spec sheet says that Rbb is 2 Ohms.

My Moving Coil preamp in Files is good for testing rbb but only for
complementary pairs. You can up the gain easily to 1000x to help.

I've never tested 2sc3329 but 2sc2546/2sa1084 at 3mA in my circuit measures
Rnv=5R. This is average rbb of 5R7 for the pair. No improvement beyond 3mA which
ties in.

Doing this little exercise has put me off trying to do a transformerless ribbon.
20x more devices & current is 60mA in my Moving Coil Preamp and only gives Rnv
0R25; still more noise than the ribbon.

From playing with Ribbon.gif & Ribbon2.gif, I'm nervous about having this large
current near an unbuffered ribbon




[Non-text portions of this message have been removed]

On May 3, 2011, at 1:36 AM, Eric Benjamin wrote:
> Ricardo wrote:
>> this little exercise has put me off trying to do a transformerless ribbon
> Aw, come on.  Where's your sense of adventure?


Exactly! : )

Some folks around here (Cambridge MA) made an active ribbon mike about 30 years
ago, and their idea was to use a current to voltage converter to load the
ribbon.  Essentially, you're loading the ribbon into a dead short, so it
modifies the kinematics of the ribbon, essentially stiffening it massively.  So,
you can totally redesign the mechanics, (i.e. the resonance point) for better or
worse.  Unlike condenser mikes, which usually run the capsule below its resonant
frequency, most ribbon mikes tune the ribbon so that the audio passband is above
the ribbon's fundamental resonance; they usually place resonance around 30-40Hz.
Never heard this particular active ribbon, or know how they tuned it, but it was
something sorta memorable for a bunch of folks, so I bet it didn't fail totally.

As for the electronics, how about doing something stupid like using a generic
TIP-32 transistor as the front end for such an amp, and seeing how it works? 
That transistor, like any generic, medium wattage bipolar, would probably have a
low Rbb, but not so much bandwidth, Hfe, or Hfe linearity... but, it might be in
the ballpark for what you'd ideally want for a ridiculously low voltage noise,
current to voltage converter.  I'd use what I'd call "half of a Trans-Amp"
circuit to place the transistor inside of an op amp feedback loop - look at the
Neotek Elite or Elan schematics to see the single ended version, or the Valley
People TransAmp mike amp circuit to see the balanced version.

Personally, a good ribbon motor into a quality transformer, then into a quality
conventional mike preamp is my preference.  But, there are some fun things you
can do with an active setup...! ;-)


Regards,

Monte McGuire
montemcguire@...

> their idea was to use a current to voltage converter to load the ribbon. 
Essentially, you're loading the ribbon into a dead short

That's Ribbon.gif & Ribbon2.gif.  A virtual earth or I to V convertor is a
headbanging Low Noise circuit which most people don't know. No evil resistors.

> kinematics of the ribbon, essentially stiffening it massively.

Actually it introduces resistance rather than stiffness. Not a lot as most
ribbons, even modern ones, are inefficient.  See BBC Engineering Monograph 4 on
the 4038.

> Unlike condenser mikes, which usually run the capsule below its resonant
frequency

That's omni condensors. Condensor Cardioid/Fig-8s etc have a midrange resonance.
See the Neuman pdf book which is the only accurate guide to mike design I've
found. Readable by non-gurus and useful to gurus too.

> using a generic TIP-32 transistor as the front end for such an amp,

In the early 80's I looked at a lot of medium power BJTs for my moving coil
preamp and so did Baxandall. The prodigy thread too.

Some were OK but none came near 2sa1084/2sc2546 so I stopped looking.

There's good stuff on the last page of the 2sb737 datasheet. Also bcarso on the
prodigy thread.

For Common Emitter, high hfe gives low current noise.  Old LN transistors were
usually just high hfe small signal transistors with some care in manufacture. eg
BC109, 214, 184, 560 etc.  These give good performance for eg 10k source down to
10uA current.  But for lower source resistance and high current, rbb is too
high.

Some medium power devices have low rbb but hfe is quite low so In is high. More
importantly, the process used is such that most large devices have high 1/f
noise.  There are exceptions I found 30yrs ago but none were as good as
2sa1084/2sc2546. You need very "clean" processing for LN.

> are some fun things you can do with an active setup...! ;-)

like blow your ribbon ... 8>D

I think the gain should be enough to drive the following piece of gear to
maximum undistorted output. As a case in point last week
I was in a studio where the board's mic pres had a max gain of only 50 dB and
with the sub 1mV/Pa experimental mic I was trying
a quiet guitar recording simply couldn't get to a level beyond -40 dBFS in their
converters. They could have had beter gain staging of course.

The quietest thing I record is my sitar...

So if I had to put a number on it I would say at least 2 mV/Pa with standard mic
pre imedances. That would be 6 dB gain over a typical ribbon.
A little more would be better...10 to 20 dB I'd say.

The signal to noise is pretty much set by the tranducer ideally. As you
transform it up voltage increases with turns ratio, impedance increases as the
square,
but thermal noise goes up as square root...so you have the same S/N with an
ideal transformer. As far as noise contribution of a pre...my feeling is that we
pick
a noise factor like one or two dB and stop due to diminishing returns. At that
point most of the noise is from the source and even an ideal noiseless preamp
would
give little audible improvement.

On the other end there is max spl for the microphone. There's nothing worse than
a mic that overloads, and I feel that the extreme SPL handling of ribbons and
dynamics
(at least mid band) is one of their best virtues and must not be degraded by
active electronics if possible.

The lowly Shure SM-57 can handle 175 dB+ SPL mid band with about 18 dBA self
noise. Even at diaphragm resonance (It's mostly mass controlled like a ribbon)
the SPL handling
is extreme.

How much is enough? Well I get outputs corresponding to about 145 dBSPL in close
drum overheads.To be clear  this does not mean the actual sound pressure is at
that level...
it means that the pressure/pressure gradient microphone is responding as it
would in a plane wave with the same pressure. Often close mics are subjected to
reactive sound fields.

Ok, so i've been thinking about the booster circuits. Richard said probably the
configuration with the least CONS is the extreme ratio tranformer with FET
followers. So I looked
at the Royer circuit...it's a center tapped high ratio transformer driving a
balanced JFET/bipolar  pair ...
http://www.google.com/patents?id=m2CbAAAAEBAJ&zoom=4&pg=PA1#v=onepage&q&f=true

Uses a tapewound amorphous core toroid.

When Richard mentioned the high ratio transformer being not unlike the jurassic
days of driving a vacuum tube grid with a ribbon mic I commented that in that
case the step up was shared
by two transformers...one in the mic and one in the pre. Well, The SE active
ribbon (Rupert Neve designed electronics) DOES use two transformers:
http://recordinghacks.com/microphones/SE-Electronics/RNR1
I have no circuit details, but peculiarly there is mention of the tronics being
between the two transformers

Then there's the Lundahl 55:55 center tapped ribbon transformer...The highest
ratio I see. Pretty expensive.

So for diy I see the Lundahl or perhaps two lower ratio transformers driving
either a balanced royer type circuit or perhaps an unbalanced input
Schoeps (modified front end of course) or Richard's ( or PRR) circuit. Richard's
latest seems aproaching viability.

For me (not diy) the goals would be:

(1) NF 1dB for the tronics
(2) gain 10 dB min
(3) No overload under any concievable use (might require atten. mechanism)
(4) no catastrophic failure modes with any wiring fault conditions (like zener
breakdown of emitter base junctions)

If I can't have those four things I probably won't do it. :^)

P.S. I am actually talking to my transformer guy about high ratios...I said
he'll prob have a cow but what will really happen is that
he will be delighted...and I'll have a cow when we talk price.

Les
L M Watts Technology



   ----- Original Message -----
   From: RSN
   To: micbuilders@yahoogroups.com
   Sent: Monday, May 02, 2011 8:07 PM
   Subject: [micbuilders] Re: circuit for active ribbon mic



   Ok... All these circuits are very good and I've now got a lot of my own active
circuits together including some with JFETs, Opamps, and BJTs all mixed
together... They work well in simulation anyway.

   Since I'm not anywhere near a studio or ribbon mic I've got some questions...

   How much gain do you need after the transformer?

   If you could have all the gain you might need, what would be a reasonable S/N
ratio?





[Non-text portions of this message have been removed]

Well I've posted a schematic with JFETs on the input instead of caps and that
seems to protect the transformer better then no caps at all. I get a little more
gain out of that then the Ribbon.gif and Ribbon2.gif as well, but I'm not sure
how this would load down the ribbon if at all...  I'm not sure if the resistor
values are correct or even if it would work in a real life situation without
going up in smoke, but here it is...

http://groups.yahoo.com/group/micbuilders/photos/album/843802531/pic/1344401803/\
view


I've also created a version with WAY too much gain that uses JFET input, NE5534
pair for gain and 2n4401 to replace the output caps that would have been needed.
Works in simulation but I'm not sure it would be usable with all that gain.
Maybe could eliminate gain on the receiving end?  I think I was able to reduce
the final gain to about 40dB and that means with a 1:37 transformer I get ~72dB
gain (from ribbon to output) with 86dB s/n on output... Anyway, it's much more
complicated and I'm guessing that it's still a bit excessive? :P

http://groups.yahoo.com/group/micbuilders/photos/album/843802531/pic/1781993885/\
view

--- In micbuilders@yahoogroups.com, Monte McGuire <montemcguire@...> wrote:
>
>
> On May 3, 2011, at 1:36 AM, Eric Benjamin wrote:
> > Ricardo wrote:
> >> this little exercise has put me off trying to do a transformerless ribbon
> > Aw, come on.  Where's your sense of adventure?
>
>
> Exactly! : )
>
> Some folks around here (Cambridge MA) made an active ribbon mike about 30
years ago, and their idea was to use a current to voltage converter to load the
ribbon.  Essentially, you're loading the ribbon into a dead short, so it
modifies the kinematics of the ribbon, essentially stiffening it massively.
> Regards,
>
> Monte McGuire
> montemcguire@...
>

I was there :) Not sure it was quite 30yr.ago. They used the complementary RHOM
2SB/C transistors now unavailable. I might have the circuit somewhere in my
archives, but last I looked they never pursued their patent. Their argument that
it would work the same magic on moving coil phono cartridges made no sense. The
low frequency resonance there is the arm/cartridge compliance at 10Hz or so and
I never saw the point in trying to damp that with the motor/generator even if
you could.

Hi

Back in the distant past I remember a basic pre-amp circuit for a low impedance
i/p using a transistor in common base configuration (IIRC).  I wonder if
there’s any mileage in this route – or is it just a dead end?  My
electronics design isn’t up to sorting out a practical circuit.

I may just possibly have the original article in my archives – but I’ve not
seen it for a while – and can’t get to check at present.

I have a pair of 30 ohm Coles 4038’s that I need to get a useable signal from.
(I also need to sort out some mounting hardware)

Every Blessing

Tony

From: userno232000
Sent: Tuesday, May 03, 2011 7:02 PM
To: micbuilders@yahoogroups.com
Subject: [micbuilders] Re: circuit for active ribbon mic




--- In mailto:micbuilders%40yahoogroups.com, Monte McGuire <montemcguire@...>
wrote:
>
>
> On May 3, 2011, at 1:36 AM, Eric Benjamin wrote:
> > Ricardo wrote:
> >> this little exercise has put me off trying to do a transformerless ribbon
> > Aw, come on. Where's your sense of adventure?
>
>
> Exactly! : )
>
> Some folks around here (Cambridge MA) made an active ribbon mike about 30
years ago, and their idea was to use a current to voltage converter to load the
ribbon. Essentially, you're loading the ribbon into a dead short, so it modifies
the kinematics of the ribbon, essentially stiffening it massively.
> Regards,
>
> Monte McGuire
> montemcguire@...
>

I was there :) Not sure it was quite 30yr.ago. They used the complementary RHOM
2SB/C transistors now unavailable. I might have the circuit somewhere in my
archives, but last I looked they never pursued their patent. Their argument that
it would work the same magic on moving coil phono cartridges made no sense. The
low frequency resonance there is the arm/cartridge compliance at 10Hz or so and
I never saw the point in trying to damp that with the motor/generator even if
you could.





[Non-text portions of this message have been removed]

I really doubt there's any advantage using a common-base topology.
It presents a low input impedance so noise is less than optimum, without
the benefit of reducing distortion (as opposed to common-emitter with
current NFB).
The only advantage of this topology is increased frequency response (as
in MHz), so is sometimes used in RF applications.

Re. 4038, 30 ohms impedance is manageable with a pair or two of BJT's
running at 10/15mA core current.

Le 03/05/2011 21:14, Rev Tony Newnham a écrit :
>
> Hi
>
> Back in the distant past I remember a basic pre-amp circuit for a low
> impedance i/p using a transistor in common base configuration (IIRC).
> I wonder if there’s any mileage in this route – or is it just a dead
> end? My electronics design isn’t up to sorting out a practical circuit.
>
> I may just possibly have the original article in my archives – but
> I’ve not seen it for a while – and can’t get to check at present.
>
> I have a pair of 30 ohm Coles 4038’s that I need to get a useable
> signal from. (I also need to sort out some mounting hardware)
>
> Every Blessing
>
> Tony
>
> From: userno232000
> Sent: Tuesday, May 03, 2011 7:02 PM
> To: micbuilders@yahoogroups.com <mailto:micbuilders%40yahoogroups.com>
> Subject: [micbuilders] Re: circuit for active ribbon mic
>
> --- In mailto:micbuilders%40yahoogroups.com, Monte McGuire
> <montemcguire@...> wrote:
> >
> >
> > On May 3, 2011, at 1:36 AM, Eric Benjamin wrote:
> > > Ricardo wrote:
> > >> this little exercise has put me off trying to do a
> transformerless ribbon
> > > Aw, come on. Where's your sense of adventure?
> >
> >
> > Exactly! : )
> >
> > Some folks around here (Cambridge MA) made an active ribbon mike
> about 30 years ago, and their idea was to use a current to voltage
> converter to load the ribbon. Essentially, you're loading the ribbon
> into a dead short, so it modifies the kinematics of the ribbon,
> essentially stiffening it massively.
> > Regards,
> >
> > Monte McGuire
> > montemcguire@...
> >
>
> I was there :) Not sure it was quite 30yr.ago. They used the
> complementary RHOM 2SB/C transistors now unavailable. I might have the
> circuit somewhere in my archives, but last I looked they never pursued
> their patent. Their argument that it would work the same magic on
> moving coil phono cartridges made no sense. The low frequency
> resonance there is the arm/cartridge compliance at 10Hz or so and I
> never saw the point in trying to damp that with the motor/generator
> even if you could.
>
> [Non-text portions of this message have been removed]
>
>


[Non-text portions of this message have been removed]

http://groups.yahoo.com/group/micbuilders/photos/album/843802531/pic/1344401803/\
view

Ron, I'm not sure this is an improvement over PRR's BJT circuit w/o capacitors. 
Yours is slightly noisier cos the extra FETs.  You will need an output trim cos
the FETs will vary a lot.

Apart from that, your circuit has PRR's pros & cons.

One con is high output Z.  PRR uses 2x150R across p2&3 to reduce this but this
reduces gain & increases noise.

http://groups.yahoo.com/group/micbuilders/photos/album/843802531/pic/1781993885/\
view

What is this one supposed to do?  Is it P48V?  What do T3/4 do?

Can you explain how you get your S/N ratio?
_________________

BTW you can get very high gain in Ribbon.gif & Ribbon2.gif by varying R1/2.
Change C1/2 so R1*C1, R2*C2 remain about the same.
Higher gain (bigger resistors) also puts less stuff in the ribbon with a line
short.
_________________

> at least 2 mV/Pa with standard mic pre imedances.

That's -54dBV/Pa which is average according to Eric's little survey.

> I get outputs corresponding to about 145 dBSPL in close drum overheads.

That's a bit difficult with P48V w/o gain switching.  A modified Schoeps circuit
can ju.uust squeeze 2.1Vp on each leg.  135dB spl at MK2 15mV/Pa.  You can get a
tiny bit more with a lot more circuit complexity.

I think you need 2 different mikes.  Passive for headbanging drums.  Active for
sensible instruments and Blumlein.

> I am actually talking to my transformer guy about high ratios...

Try Marik (Mark Fouxman) 
http://www.samaraudiodesign.com/RibbonMicrophoneTransformers.html

His stuff is smaller than Lundahl for "similar" spec so I suspect he cheats. 
Probably less inductance to get lower losses & size.  But this might be a good
exchange.  Amorphous taped wound cores.  Winding huge ratios is not a big deal
if you know how.  For a ribbon, you probably shouldn't section but use Cu tape
instead. Easier to get nice neat windings.
__________________

> The low frequency resonance there is the arm/cartridge compliance at 10Hz or
so and I never saw the point in trying to damp that with the motor/generator
even if you could.

Actually that's exactly where you want to damp a arm/cartridge system.  Damping
at the pivot is crap.  The Shure V15-5 damped brush attempts to put it near the
right place.

Resistance reduces mid trackability but IIRC, there's a famous Shure paper which
shows mid tracking is usually ample.  Les will give us chapter & verse.

Don't think you can introduce enough to matter.

Don Barlow and Tony Emerson did the LEAK MC cartridges of the 60s.  LEAK never
did another commercial cartridge but they were certainly up to date with
developments.  Some prototypes circa 1980 when I was R&D teaboy at Rank
LEAK/Wharfedale.  I was never as successful with cartridges as with mikes &
speakers ... 8>D
___________________

> I have a pair of 30 ohm Coles 4038's that I need to get a useable signal from.

Tony, I'll cautiously suggest Ribbon2.gif  ...  maybe even upping the gain to
30dB.
_______________

> I really doubt there's any advantage using a common-base topology.  It
presents a low input impedance so noise is less than optimum, without the
benefit of reducing distortion (as opposed to common-emitter with current NFB).

Jerry, you have to analyse each circuit and application.  My Moving Coil Preamp
is common-base and is the lowest noise such device in the known universe. 
Common base has a max NF of 3dB but in this application there is no advantage to
CE with current NFB.  CB distortion is also much less than CE for this
application.

> Ron, I'm not sure this is an improvement over PRR's BJT circuit w/o
capacitors.  Yours is slightly noisier cos the extra FETs.  You will need an
output trim cos the FETs will vary a lot.

Under testing it may not be an improvement but it doesn't show voltage on the
ribbon transformer which was the real goal for that particular circuit.


> What is this one supposed to do?
It's just another crazy P48 circuit for balanced input/output. I messed up the
numbering on the transistors so I've updated the circuit here:
http://groups.yahoo.com/group/micbuilders/photos/album/843802531/pic/574237475/v\
iew

>Is it P48V?
The 'Way Too Much Gain' amp is a P48 circuit that is a hybrid of PRR's circuit,
a 1510/OPA217 and a P48 extractor circuit from one of Rod Elliott's projects...

>What do T3/4 do?
I'm assuming you mean T5/6 here in the updated version.  The idea came from Rod
Elliott's Project 93, Figure 7.  He explains it on his site here:
http://sound.westhost.com/project93.htm

Just scroll down, he explains it under Figure 7.

> Can you explain how you get your S/N ratio?
I may have fudged the numbers a bit.  I was up late experimenting with the new
TINA-TI v9 and it occasionally has some visual problems (or I do at least), but
it does do simulations a lot faster.

Anyway, the goal for that circuit was to get as much gain as I possibly could
without getting power into the ribbon transformer.  When you consider S/N it
doesn't seem all that great...

I've been using a rough approximation of a Lundahl 2912 in all the circuits for
testing. For testing the voltage/amps across the transformer I replace it with a
resistor and calculate the voltage across the resistor instead.  I used TINA-TI
for all the numbers.  Also, I've posted a S/N graph and a Gain/Phase graph of
the NE5534 circuit generated with TINA-TI.

Honestly, I really may not know what I'm doing here and everything I've done may
be completely wrong...  I guess I've just been experimenting with a lot of
different circuits looking at pros and cons, not really knowing what I'm doing. 
I would like to find a nice balance between gain, noise and fool proofing
though.  That way I can build it and maybe sell it to some friends when I'm done
without any worries it's going to turn out as a ribbon killer.  I could just be
too paranoid though...

> > I really doubt there's any advantage using a common-base topology.
> It presents a low input impedance so noise is less than optimum,
> without the benefit of reducing distortion (as opposed to
> common-emitter with current NFB).
>
> Jerry, you have to analyse each circuit and application. My Moving
> Coil Preamp is common-base and is the lowest noise such device in the
> known universe. Common base has a max NF of 3dB but in this
> application there is no advantage to CE with current NFB. CB
> distortion is also much less than CE for this application.
>
That's exactly what I did, analysing the circuit and the application. A
MC cartridge is much more efficient than a ribbon transducer, by a
factor of 10 at least (10% vs. 1%), so loading a ribbon mic wastes 99%
of the available signal. I'm not advocating CE with current NFB either,
for the same reason.
>
> __._,_.__
> ._,___


[Non-text portions of this message have been removed]

http://groups.yahoo.com/group/micbuilders/photos/album/843802531/pic/1344401803/\
view

> Under testing it may not be an improvement but it doesn't show voltage on the
ribbon transformer which was the real goal for that particular circuit.

Ron, does this mean SPICE shows no ribbon voltage even when one line shorted?

> It's just another crazy P48 circuit for balanced input/output. I messed up the
numbering on the transistors so I've updated the circuit here:

http://groups.yahoo.com/group/micbuilders/photos/album/843802531/pic/574237475/v\
iew

.....  He explains it on his site here:

http://sound.westhost.com/project93.htm

OK.  I didn't spot that connection to the power supply on T5/6.  The jpg is very
small so the text is poor even when magnified.

If you are going to the trouble to use several OPAs, you should try to get a lot
for it.  2 x NE5534 will take 2x4mA with no signal.  I'm not sure what happens
if there is signal into a 2k mixer input.  T3/4 might be better replaced by 2
electrolytics like the ESP fig 7.

You might consider replacing T1/2, the NE5534s & T3/4 with THAT 1570. Presently,
you have Ein of 2 FETs, 4 BJTs and a 3k3 resistor in series with your source. 
1570 can replace that with 2 very low noise BJTs and maybe 10R.

If fact R7 3k3 has more than 10x the noise power of your 300R ribbon losing you
10dB S/N immediately before the active devices are considered.

1570 takes 7.5-10mA so you have to check if you get enough voltage from P48V. 
Don't know how an ribbon connected directly to 1570 is protected if one line
shorts.
_________________________________

> Can you explain how you get your S/N ratio?

I meant what is the signal?  Input?  Output?  What conditions?

I assume your noise is output noise cos you say it gets worse with gain.

> For testing the voltage/amps across the transformer I replace it with a
resistor and calculate the voltage across the resistor instead.

Use the DC of the secondary for input offset (voltage across ribbon).  For
noise, use the nominal impedance of the mike; eg 300R

--- In micbuilders@yahoogroups.com, Richard Lee <ricardo@...> wrote:
>
>
http://groups.yahoo.com/group/micbuilders/photos/album/843802531/pic/1344401803/\
view
>
>ribbon).  For noise, use the nominal impedance of the mike; eg 300R
>

Just for info the guys at MIT were using the ribbon ALONE into a complementary
CB amp (virtual short). Massively parallel bipolars at lots of current and no
fault protection, I would not put a $1000 MC cart on it, at least ribbons are
replaceable. It was a vintage RCA 44(?) and it sounded fine.

>
> Just for info the guys at MIT were using the ribbon ALONE into a
> complementary CB amp (virtual short).
>
Can you be a little more specific about the actual topology?
Virtual short implies current NFB, which a CB amp does not cater for
naturally.
>
>


[Non-text portions of this message have been removed]

--- In micbuilders@yahoogroups.com, Jerry Lee Marcel <jerryleemarcel@...> wrote:
>
>
>
> >
> > Just for info the guys at MIT were using the ribbon ALONE into a
> > complementary CB amp (virtual short).
> >
> Can you be a little more specific about the actual topology?
> Virtual short implies current NFB, which a CB amp does not cater for
> naturally.
> >
> >
>
>
> [Non-text portions of this message have been removed]

Think of two simple current mirrors (one NPN one PNP) facing each other with the
ribbon tied between ground and the two diode connected transistors and a gain
resistor from the collectors back to the input. Now bias them with current
sources from the top and bottom. If the bias current is big enough re will be
very small and the ribbon current is mirrored back through the gain resistor
which is much larger than the ribbon R. There was circuitry to servo the bias to
"do the right thing". Yes the transistors are at 0 Vcb which is OK for those
RHOM devices (Vsat < .1V).

Le 05/05/2011 18:42, userno232000 a écrit :
>
>
>
> --- In micbuilders@yahoogroups.com
> <mailto:micbuilders%40yahoogroups.com>, Jerry Lee Marcel
> <jerryleemarcel@...> wrote:
> >
> >
> >
> > >
> > > Just for info the guys at MIT were using the ribbon ALONE into a
> > > complementary CB amp (virtual short).
> > >
> > Can you be a little more specific about the actual topology?
> > Virtual short implies current NFB, which a CB amp does not cater for
> > naturally.
> > >
> > >
> >
> >
> > [Non-text portions of this message have been removed]
>
> Think of two simple current mirrors (one NPN one PNP) facing each
> other with the ribbon tied between ground and the two diode connected
> transistors and a gain resistor from the collectors back to the input.
>
So, there would be positive feedback?


[Non-text portions of this message have been removed]

> Think of two simple current mirrors (one NPN one PNP) facing each other
.......

Here's a similar circuit with much fewer the devices for the same performance
and no need to servo.

http://f1.grp.yahoofs.com/v1/4B3DTcikRL0ZUMDhX5Awy87qgN7bvQk-rBZ-Z3qa6mU9n9eTvcq\
IbRGXo349jGuYgtsLV1TBKuUH1k8l_Dw3dQ/Ricardo/MCjul81.pdf

More stuff in Readme.doc

Common Base needs Ri = Rs for best noise (& power transfer) which is a maximum
of 3dB NF with "perfect" (rbb << Rs) devices.  Running 60mA gives Ri = 0R22 eg
Coles 4038 ribbon.

Using 40x 2sa1084 (Unobtainium) and 40x 2sc2547 (available), rbb 5R7, give Rnv
0R18.

30x 2sd786 (Unobtainium) and 30x 2sc3329 (available), rbb 2R gives Rnv 0R14

Feeding the input to the bases has more distortion and evilytic noise is more
pronounced. (My blind panel said musicality is reduced!)  Noise is NOT improved.
Taking the 1k load to the bases gives current feedback but there is insufficient
gain to make much difference to CE distortion.

Distortion drops like a stone with level in the simple CB version.

I wouldn't use this for a ribbon but its probably safer than the MIT circuit
which I haven't seen.
__________________

Marcel, in this circuit, CB gives very low noise and the best distortion & other
stuff.

Ribbon.gif is CE virtual earth, and gives the best noise.  Feedback lowers
distortion and reduces Zo, important for a mike.

Both are low Ri but give better noise than high Ri solutions.  The input R is
actually a side issue for low noise.  Feedback can give you any input R but
doesn't really affect noise (unless evil resistors are introduced  ...  8>). 
The important factors for low noise are Power/Noise matching and the absence of
evil resistors.

What configuration gives best results depends on the exact circuit and the
application.

> > Under testing it may not be an improvement but it doesn't show voltage on
the ribbon transformer which was the real goal for that particular circuit.
>
> Ron, does this mean SPICE shows no ribbon voltage even when one line shorted?

It's a very very low number across a 300R load...  5.29E-23V @ 1.76E-25A.


> OK.  I didn't spot that connection to the power supply on T5/6.  The jpg is
very small so the text is poor even when magnified.

Not sure if this is a solution there but at the bottom of the page there is a
link to the original size image which is much clearer.  It should say "View:
small | medium | Large" or something similar.


> You might consider replacing T1/2, the NE5534s & T3/4 with THAT 1570.
Presently, you have Ein of 2 FETs, 4 BJTs and a 3k3 resistor in series with your
source.  1570 can replace that with 2 very low noise BJTs and maybe 10R.

I've looked for a good spice model of those THAT circuits but I've never found
any.  If anyone knows where I can find them I would much appreciate it.  I have
one for a INA163 and I think a INA217 but no luck finding any for THAT 1570...


> > Can you explain how you get your S/N ratio?
> I meant what is the signal?  Input?  Output?  What conditions?
> I assume your noise is output noise cos you say it gets worse with gain.

I put a 1KHz sine wave at 500uV p-p signal into the primary of the transformer. 
500uV was a random guess for the ribbon output.  The resistance of the input
signal (VG2 in WTMG schematic) is set at 10m Ohms and that is also a guess. 
Then the S/N is then measured at the output of XLR-3.