Homemade Mic Pre-preamp Project

[Chris Hornbeck]

Beginning in a condenser mic thread, several folks have
been contributing ideas and input to a construction
project intended for use with ribbon or other low-output
mic's. The project pre-preamp needs to run on phantom
power and give 20dB of low-noise gain.

Initial attempts to build the electronics into an A3M/A3F
shell pair with mic cable between have been sobering. It
*can* be done, but nobody in their right mind would want
to spend the time and effort. And they seem so roomy when
you're only soldering in wire.....

Next I'm trying two chassis options. First, a Hammond
cast aluminum box, size 1590C, with two Neutrik NC3MD, two
NC3FD, and the electronics for two pre-preamps.

The alternate chassis I'm trying is the Switchcraft S3FM,
after a commercial design. Found a source!

In the cast box two channel design, we're trying hand-matched
2SK369 BL's. In the S3FM sleeve, we're trying a 2SK389 GR,
a monolithic (?) dualFET.

To the adventurous folks who've volunteered to test the
prototypes, thanks, and hang in there another week, to get
proper chassis in.

And to everyone, comments welcomed.
Thanks,


Chris Hornbeck

[Kurt Albershardt]

Chris Hornbeck wrote:

> Beginning in a condenser mic thread, several folks have
> been contributing ideas and input to a construction
> project intended for use with ribbon or other low-output
> mic's. The project pre-preamp needs to run on phantom
> power and give 20dB of low-noise gain.

20 dB of gain for ribbons? As a pre-preamp, maybe but I think you're
going to want more than that if you're going straight to tape or ADC.





> To the adventurous folks who've volunteered to test the
> prototypes, thanks, and hang in there another week, to get
> proper chassis in.

I'm interested, tell me (or point me at) more...

[P Stamler]

>The best method I can think of would cost a little more to build, but
>would also be a lot simpler than an active pre-preamp. Why not just
>put some JEnsen JT-115K input transformers in a box in front of a
>typical transformerless mic preamp? The 1:10 step-up ratio will give
>you your 20dB gain, and the mic preamp could have a rather high input
>impedance, which would be just fine when you're not using the
>"pre-pre."

I'm not sure whether you're suggesting simply putting a transformer in a box,
for use with existing preamps, or creating a new design that's
transformer-coupled for ribbons and transformerless for condensers. But neither
will really wash.

A transformer is really picky about loading. The JT-115K-E wants to see 150k at
its secondary; the typical input of a standard transformerless preamp is about
1.5k-3k. This would translate into a 15-30 ohm load at the primary, which most
ribbons would hate. It would also damp the transformer's response very heavily,
probably killing most of your high end.

So what about a changeable circuit. Well, if it's transformerless with a
switcheable transformer, presumably you can design it with two legs at 75k
each, letting the phantom power resistors form the microphone termination. But
you'd have to do some interesting switching. Naah, just design a good
transformer-coupled input stage, and a good transformerless one, give each a
jack and...well, you've just built two preamps. So build two preamps.

I'm rambling, so I'll go to bed.

Peace,
Paul

[Justin Ulysses Morse]

P Stamler > wrote:

> I'm not sure whether you're suggesting simply putting a transformer
> in a box, for use with existing preamps, or creating a new design
> that's transformer-coupled for ribbons and transformerless for
> condensers. But neither will really wash.

Yeah, I don't really know either. In my defense, I never found the
original thread that generated this discussion.

> A transformer is really picky about loading. The JT-115K-E wants to
> see 150k at its secondary; the typical input of a standard
> transformerless preamp is about 1.5k-3k. This would translate into a
> 15-30 ohm load at the primary, which most ribbons would hate. It
> would also damp the transformer's response very heavily, probably
> killing most of your high end.

Right, I guess I was thinking that if you were using a transformer for
occasions when you want a lower input impedance and a gain step-up,
then on the occasions when you don't use the transformer, you could
probably have a much higher-than-usual input impedance. I bet a TLM103
would be perfectly happy driving a 150k input. The coupling capacitors
could be downright tiny.

> So what about a changeable circuit. Well, if it's transformerless with a
> switcheable transformer, presumably you can design it with two legs at 75k
> each, letting the phantom power resistors form the microphone termination. But
> you'd have to do some interesting switching. Naah, just design a good
> transformer-coupled input stage, and a good transformerless one, give each a
> jack and...well, you've just built two preamps. So build two preamps.

As long as your phantom switch disconnected the two 6k8 resistors from
each other, you'd be okay I think. "Build two preamps" is still the
wiener, imho.

> I'm rambling, so I'll go to bed.

I should be so wise.


ulysses

[Kurt Albershardt]

Ben Bradley wrote:

> In rec.audio.pro, Chris Hornbeck >
> wrote:
>
>
>> On Mon, 08 Sep 2003 17:41:58 -0700, Kurt Albershardt >
>> wrote:
>>
>>
>>> 20 dB of gain for ribbons? As a pre-preamp, maybe but I think you're
>>> going to want more than that if you're going straight to tape or ADC.
>>
>> Hi Kurt,
>> Yes, the intended use is in a mic line, upstream of a conventional
>> mic preamp, to (hopefully!) reduce noise, quantization, and RFI
>> and to lightly load the mic's transformers.
>
>
> So it already has the usual ribbon-mic step-up transformer in it,
> you just want some extra gain in between that and the 'regular' pre,
> and have your circuitry 'properly' load (or not-load) the transformer.
>
>
>> It also has to be a simple enough project that anyone on the
>> group who can solder could build it.
>
>
> Does that mean a thru-hole PCB design? Your size constraints tends
> to point towards an SMT design (maybe even with parts on both sides),
> which may be possible, but certainly a challenge, for 'anyone who can
> solder.'
> What does your "pre-pre" look like? [no comments from the peanut
> gallery, please...]


Maybe it could even run on phantom power? If it needs just a tad more
oomph, maybe T-power?

[Chris Hornbeck]

On Tue, 09 Sep 2003 19:16:59 GMT,
(Ben Bradley) wrote:

> Does that mean a thru-hole PCB design? Your size constraints tends
>to point towards an SMT design (maybe even with parts on both sides),
>which may be possible, but certainly a challenge, for 'anyone who can
>solder.'

Hi Ben,
This will be so simple that it won't need a PCB. Just a pair of
FET's and a few resistors, an input bifilar choke and caps for
RF suppression if desired. Nothing fancy; it can be built right
on its XLR connectors.

> What does your "pre-pre" look like? [no comments from the peanut
>gallery, please...]

[Arf!] I guess anybody interested will choose their own
construction technique, but I'm going to try two versions to
test the concept, and two chassis types to test ease of
construction.

One version will require the builder to buy a batch of
2SK369 FET's and match them (instructions will be posted).
The other version will use dual FET's, type 2SK389,
factory matched in a common package.

One chassis to be tried is a "sleeve" from Switchcraft with
XLR male and female ends. Might be tight....
The other chassis I'm trying is a small cast aluminum box
made by Hammond that I use a lot and like. It has room for
two pre-preamp channels.

The circuit is of course just a push-pull stage with the
phantom resistors providing drain loads.

Thanks,

Chris Hornbeck

[Chris Hornbeck]

On Tue, 09 Sep 2003 15:38:17 -0700, Kurt Albershardt >
wrote:

>Maybe it could even run on phantom power? If it needs just a tad more
>oomph, maybe T-power?

Yes, phantom power of about 6mA per "leg" is used, 12mA per
pre-preamp.

Thanks,

Chris Hornbeck

[David Satz]

Chris Hornbeck wrote:

> Yes, phantom power of about 6mA per "leg" is used, 12mA per pre-preamp.

Urp. When I wrote earlier that 6 mA was too much for many preamps and
mixers, I meant 6 mA per microphone input total! Totally shorting a
(correctly implemented) P 48 circuit would draw only 2 mA more than
your proposed 12. How many preamps or mixers can withstand an indefinite
short circuit on multiple inputs with the phantom powering turned on?

I don't know the answer. It's a reasonable enough requirement, but I can't
think of a single U.S. manufacturer whose specifications say that it's OK.
(I just did a quick check of Millenia, John Hardy and Great River ...)
With a lot of portable equipment that I know, using a pair of these devices
would almost certainly cause a sacrifice, complete with burnt offering.

The beefiest outboard phantom power supplies that I know of are rated at
"only" 10 mA/channel (the maximum current permitted for a P 48 microphone
according to the IEC specification)--and those are relatively rare. I can
also show you a nice Neumann brand phantom supply (model N 80) that offers
a grand total of 2 mA for a _pair_ of microphones. It was widely sold in
the KM 84 era because the heavier ones were so expensive and at the time,
anything beyond about 0.7 mA per microphone was generally unnecessary.

Apart from that, there are many preamps and mixers in which the filtering
of the 48 Volt supply introduces several kOhms additional series resistance
prior to the 6.8 kOhm phantom resistors. With increasing current, there's
naturally a larger and larger DC voltage drop across the filter. 2 kOhms
extra at that point (and I've certainly seen greater) would present quite
a barrier to a pair of 12 mA devices ... the filter would practically be
trying to suck power out of the microphones at that point.

I don't want to discourage you from building the lowest-noise design if
it can be used with some carefully selected preamps--once we find out
which preamps those are--but is there a workable fallback approach that
would use, say, 4 - 5 mA per microphone? That much is available from the
better preamps and mixers, including some (though certainly not all)
portable equipment.

I think we're beginning to see why this hasn't become a commercial product
before now, and why the only announced commercial product of its type isn't
specified to be a super-high-performance piece.

--best regards

[Kurt Albershardt]

David Satz wrote:

> Chris Hornbeck wrote:
>
>> Yes, phantom power of about 6mA per "leg" is used, 12mA per pre-preamp.
>
>
> Urp. When I wrote earlier that 6 mA was too much for many preamps and
> mixers, I meant 6 mA per microphone input total! Totally shorting a
> (correctly implemented) P 48 circuit would draw only 2 mA more than
> your proposed 12.


Which is why I suggested T-power as a fallback--if you're going to wind
up building an external power source anyway, that would at least give
some users the option to self-power. Also, building external T-power
boxes is pretty simple stuff.

[Chris Hornbeck]

On 10 Sep 2003 06:14:42 -0700, (David Satz) wrote:

>The beefiest outboard phantom power supplies that I know of are rated at
>"only" 10 mA/channel (the maximum current permitted for a P 48 microphone
>according to the IEC specification)--and those are relatively rare.

I'm too used to my own designs, I guess. And take things as trivial as
half watt power supplies as a given. Oh, well. The only way this will
still feel like a fun project for me is if it *does* run off phantom
power. So, for me, the phantom supplies will just have to be up to
snuff.

>I don't want to discourage you from building the lowest-noise design if
>it can be used with some carefully selected preamps--once we find out
>which preamps those are--but is there a workable fallback approach that
>would use, say, 4 - 5 mA per microphone? That much is available from the
>better preamps and mixers, including some (though certainly not all)
>portable equipment.

Sure; the operating current will have to be set by the builder anyway.
Noise increases and gain decreases more or less linearly with reducing
current, but 6dB increase in noise is *not* 6dB increase in noise
figure; it's much less. For your test prototype, I can include a
choice of common source resistors, which set bias, for you to
clip out and experiment. All builders will need is a DVM.

>I think we're beginning to see why this hasn't become a commercial product
>before now, and why the only announced commercial product of its type isn't
>specified to be a super-high-performance piece.

That makes good sense to me. Also, if I were making a commercial
product "like" this, I'd have a dedicated AC power supply anyway.
I'd want complimentary DC supplies and enough current to get to
Paul's recommended 1dB noise figure with 150 ohm source, and
output buffering, and.. and.... Enough flash to be able to charge
money for.

But DIY has its own joys.

Thanks,

Chris Hornbeck

[David Satz]

Chris Hornbeck wrote:

> The only way this will still feel like a fun project for me is if it
> *does* run off phantom power. So, for me, the phantom supplies will
> just have to be up to snuff.

I agree--and that's the only way it would be really convenient for most
of us to use, I think.

--Last night I received email from Dan Kennedy of Great River Electronics,
who is following this thread although apparently his newsreader software
doesn't currently support posting messages. He said (quoted by permission):

"The Great River pre's will all withstand a short of either or both legs to
ground for an infinite length of time, so it will be fine with the pre-pre
suggested. [ ... ] I've not specifically mentioned it in the spec sheets
because I understood that capability to be an implicit requirement of the
P48 standard. [ ... T]he supplies will do 48 volts +/-2% into the top of
the 6K81 resistors all day, shorted, all channels active."

So, "yay." (quoted by me from me)

[LeBaron & Alrich]

David Satz > wrote:

> Chris Hornbeck wrote:

> > The only way this will still feel like a fun project for me is if it
> > *does* run off phantom power. So, for me, the phantom supplies will
> > just have to be up to snuff.

> I agree--and that's the only way it would be really convenient for most
> of us to use, I think.

> --Last night I received email from Dan Kennedy of Great River Electronics,
> who is following this thread although apparently his newsreader software
> doesn't currently support posting messages. He said (quoted by permission):

> "The Great River pre's will all withstand a short of either or both legs to
> ground for an infinite length of time, so it will be fine with the pre-pre
> suggested. [ ... ] I've not specifically mentioned it in the spec sheets
> because I understood that capability to be an implicit requirement of the
> P48 standard. [ ... T]he supplies will do 48 volts +/-2% into the top of
> the 6K81 resistors all day, shorted, all channels active."

> So, "yay." (quoted by me from me)

Thanks, David. (Been wondering where Dan was, but figured it was
hunkered down on an EQ to mate with the NV.) It's exactly that kind of
intelligent design and construction, one that aims to deliver
bulletproof kit, that puts the cost in the good stuff. Folks who need
tools that not only sound great but work day in day out under pressure,
where there will be no second chance at a take, understand and
appreciate this concern for quality sound and reliable operation.

--
hank alrich * secret mountain
audio recording * music production * sound reinforcement
"If laughter is the best medicine let's take a double dose"

[Kurt Albershardt]

David Satz wrote:
> Chris Hornbeck wrote:
>
>
>> The only way this will still feel like a fun project for me is if it
>> *does* run off phantom power. So, for me, the phantom supplies will
>> just have to be up to snuff.
>
>
> I agree--and that's the only way it would be really convenient for most
> of us to use, I think.
>
> --Last night I received email from Dan Kennedy of Great River Electronics,
> who is following this thread although apparently his newsreader software
> doesn't currently support posting messages. He said (quoted by permission):
>
> "The Great River pre's will all withstand a short of either or both legs to
> ground for an infinite length of time, so it will be fine with the pre-pre
> suggested. [ ... ] I've not specifically mentioned it in the spec sheets
> because I understood that capability to be an implicit requirement of the
> P48 standard. [ ... T]he supplies will do 48 volts +/-2% into the top of
> the 6K81 resistors all day, shorted, all channels active."

You might also consider a circuit (a la Josephson) which can take and
use higher current if it's available (via lower resistor values in the
preamp.) Looking forward to trying this when my Series Sixes show up.

[GKB]

I haven't followed all the posts , does the poster
want to make a circuit like the royer active 122 ?
I don't think anyone can give you the circuit !
regards Greg



Chris Hornbeck wrote:

> Beginning in a condenser mic thread, several folks have
> been contributing ideas and input to a construction
> project intended for use with ribbon or other low-output
> mic's. The project pre-preamp needs to run on phantom
> power and give 20dB of low-noise gain.
>
> Initial attempts to build the electronics into an A3M/A3F
> shell pair with mic cable between have been sobering. It
> *can* be done, but nobody in their right mind would want
> to spend the time and effort. And they seem so roomy when
> you're only soldering in wire.....
>
> Next I'm trying two chassis options. First, a Hammond
> cast aluminum box, size 1590C, with two Neutrik NC3MD, two
> NC3FD, and the electronics for two pre-preamps.
>
> The alternate chassis I'm trying is the Switchcraft S3FM,
> after a commercial design. Found a source!
>
> In the cast box two channel design, we're trying hand-matched
> 2SK369 BL's. In the S3FM sleeve, we're trying a 2SK389 GR,
> a monolithic (?) dualFET.
>
> To the adventurous folks who've volunteered to test the
> prototypes, thanks, and hang in there another week, to get
> proper chassis in.
>
> And to everyone, comments welcomed.
> Thanks,
>
> Chris Hornbeck

[Chris Hornbeck]

On Thu, 11 Sep 2003 12:26:47 -0700, Kurt Albershardt >
wrote:

>You might also consider a circuit (a la Josephson) which can take and
>use higher current if it's available (via lower resistor values in the
>preamp.) Looking forward to trying this when my Series Sixes show up.

If I'm thinkin what you're thinkin, that's a pretty cool idea. Can't
seem to get details from Josephson's website, which is in flux, but
combined with David's timely heads-up, a much better perspective is
appearing.

Since modern and well-engineered mic preamps like the Great River have
less need for an add-on contraption, and less expensive or older gear
have likely spottier phantom supplies, the first step for this
project must be to test and evaluate the phantom supply.

Two values of loading and observation with a DVM should tell us a lot.
Then we can decide on an individual current drain and safely use the
phantom supply to sort and match FET's.

And those willing to dedicate a channel or two to higher phantom
current could push the noise voltage envelope even further. I'm agin'
it on constitutional grounds, but some folks may want to. More
is more.

Thanks,


Chris

[David Satz]

Chris Hornbeck wrote:

> And those willing to dedicate a channel or two to higher phantom
> current could push the noise voltage envelope even further. I'm agin'
> it on constitutional grounds, but some folks may want to. More is more.

I'm agin' it as well (beyond a few percent, anyway)--but a word to those
who'll do it nonetheless: The ability of a preamp to reject common-mode
hum and hash depends on the exact match of the two phantom supply resistors
(on each microphone input) to each other.

The resistors don't have to be exactly 6.8 kOhms, and the ones on the left
input don't have to be exactly the same as the ones on the right input--
but the two on the left input _MUST_ be the same as each other, and the
two on the right input _MUST_ be the same as each other.

There, have I said it enough times, and clearly enough?

Even 1% tolerance resistors (which means +/- 1%, for a total tolerance of
nearly 2%) aren't automatically close enough. The IEC standard specifies
that the pairs shall be matched so that the difference within any given
pair is less than 0.5% (= 34 Ohms).

But even closer matching is nearly always worthwhile, because that helps
to keep the phantom supply's own residual hum and hash from entering the
input in differential mode.

--best regards

[John Cafarella]

"Kurt Albershardt" > wrote in message
...
> It is an intriguing idea. Maybe more intriguing would be if enough
> microphones started supporting it and we started seeing preamps with
> switchable phantom buildout resistance...
>
It is kinda telco-ish isn't it :-)

--
John Cafarella
EOR Studio
Melbourne Australia
[ cafarellaj at powertel dot com dot au ]