How is variable impedance being implemented on some of the mic pres
(Vipre, Digital MPA etc)

I would guess it would be a dual resistance to ground, but it would
have to be on the inboard side of the transformer (or capacitors in a
non-transformer design).

It seems that all the pres have some set impedance just inside the dc
blocking like 10k or so.

What range of impedance is appropriate?

R. Foote wrote:
How is variable impedance being implemented on some of the mic pres
(Vipre, Digital MPA etc)

Some are using multitap transformers. Some are using shunt resistors.

I would guess it would be a dual resistance to ground, but it would
have to be on the inboard side of the transformer (or capacitors in a
non-transformer design).

Shunt resistors are not a good thing from a S/N standpoint, so if you
are going transformerless, a transformer with a multitap primary is
the way to go.

It seems that all the pres have some set impedance just inside the dc
blocking like 10k or so.

Often much lower. Often the input stage is fairly low Z to begin with
in a transformerless type.
--scott


--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Thanks Scott.
I guess if I can find a multi-tapped primary Jensen transformer I will
try refitting my diff amp input with that. (It has coupling capacitors
now) Right now, the amps inputs are at 10k, but the transformer will
probably want it's own specific secondary impedance? I wonder if
Jensen supplies that info when you but a transformer.

Obviously, the multi tap transformer is the way to go.

Roger Foote

R. Foote wrote:
I guess if I can find a multi-tapped primary Jensen transformer I will
try refitting my diff amp input with that. (It has coupling capacitors
now) Right now, the amps inputs are at 10k, but the transformer will
probably want it's own specific secondary impedance? I wonder if
Jensen supplies that info when you but a transformer.

Yes. Look at the Jensen web site. Each of the transformers basically want
to see a particular load impedance, and in the case of the input transformers,
if the load impedance isn't right, they will ring. Output transformers are
usually designed to work into a wider range of load impedances, at the expense
of efficiency and noise performance, because that's a lot more important in
that application.

Obviously, the multi tap transformer is the way to go.

If you like the sound of transformers.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

In article writes:

How is variable impedance being implemented on some of the mic pres
(Vipre, Digital MPA etc)

I the case of the ViPre there's switch that selects taps on the input
transformer from about 50 ohms up to about 3K I think. In the case of
the MPA, I don't know. Some simply use a resistor, which also has an
effect on the sound of the mic - a different effect than a transformer
tap, but an effect none the less. And that's what it's all about -
changing something and listening to see if you like it.



--
I'm really Mike Rivers - )

Well I found a transformer (JT-MB-CPCA) that seems suited to work with
my preamp which has a differential input structure. Right now the diff
amp is at 20k impedance and this transformer wants 3k, so it would be
an easy thing to replace the 2- 10k resistors with 1.5k resistors...

I wonder if there is any benefit going with a single ended transformer
tied to an AD743 for instance versus retro-fitting this existing
differential amp?

Of course this is not going to be a variable Z pre, but that seems a
little tough to get parts for, and I can't just mess with the
secondary loading of the transformer without risk of ringing.

Anyway I just want to say thanks for taking the time to help me out on
this Scott!

Roger Foote

R. Foote wrote:
Well I found a transformer (JT-MB-CPCA) that seems suited to work with
my preamp which has a differential input structure. Right now the diff
amp is at 20k impedance and this transformer wants 3k, so it would be
an easy thing to replace the 2- 10k resistors with 1.5k resistors...

Sounds reasonable. Run a square wave through and see what it looks like
on the other end. Set the load capacitance and resistance so you get a
nice-looking 1 KC square wave on the scope.

I wonder if there is any benefit going with a single ended transformer
tied to an AD743 for instance versus retro-fitting this existing
differential amp?

Depends on the stage itself. A differential amp stage, if it's well
designed, can give you some linearity benefits over a simple
one-transistor input stage.

All op-amps basically have a differential front-end going into an
output stage (usually push-pull but sometimes single-ended).

Of course this is not going to be a variable Z pre, but that seems a
little tough to get parts for, and I can't just mess with the
secondary loading of the transformer without risk of ringing.

Why not? You can see ringing on the scope. You can "select on test"
until it looks good on the scope.

Anyway I just want to say thanks for taking the time to help me out on
this Scott!

The Jensen applications notes have a couple nice designs for transformer
input preamps with various technologies.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

I am studying the Jensen schematics as we speak. Thanks Scott and Mike
for the invaluable input!

Well I ordered a Sowter #3678 multi-tap mic transformer. Tapped at 50,
200 and 600 ohms. Wish it had some higher taps (1k or so) but at least
it has some variation.

All of the Jensens seem to have between 1.2 and 1.4k primaries, so if
a person was to put a resistance3 across pins 2 and 3, it would lower
the impedance even more.. So maybe 600 ohms won't be so bad.

I think I will try a transformerless design with variable z. I guess
Panasonic bipolar caps would be best?

R. Foote wrote:

Well I ordered a Sowter #3678 multi-tap mic transformer. Tapped at 50,
200 and 600 ohms. Wish it had some higher taps (1k or so) but at least
it has some variation.

the people who do it "right" (Great River, myself, certainly others)
simply use a relay to strap a dual-primary transformer for series or
parallel connection. This gives you nominally 150 or 600 ohm
impedance.
I make a circuitboard that handles this function along with polarity
switching, -20dB pad, phantom power, and XLR input jacks all on a 3x3"
2-channel board. http://www.rollmusic.com/projects/3125.shtml

All of the Jensens seem to have between 1.2 and 1.4k primaries, so if
a person was to put a resistance3 across pins 2 and 3, it would lower
the impedance even more.. So maybe 600 ohms won't be so bad.

Ultimately these numbers are pretty meaningless most of the time. A
transformer doesn't have an actual impedance of its own, only an
impedance ratio. So the actual source and load impedances are
dependent on the microphone, the transformer ratio, and the amplifier.
Along with any terminating resistance and/or capacitance.

I think I will try a transformerless design with variable z. I guess
Panasonic bipolar caps would be best?

Transformerless preamps with a variable input impedance are simply a
bad idea if you ask me. Unles you come up with some very sophisticated
circuit to vary the actual input impedance of the amplifier, then
you're forced to use loading resistors which throw away precious
mic-level signal voltage and really defeat the purpose of the whole
concept. The reason impedance-variance works so well with transformers
is that varying the ratio lets you trade impedance for gain without a
noise penalty. You still have to face the limitations of high-ratio
transformers, but if you're buying new transformers from the likes of
Jensen, Sowter, or Lundahl then you've got a big technology advantage
over the iron of yore which suffered phase, frequency, and distortion
penalties in high-ratio arrangements.

ulysses

R. Foote wrote:

I think I will try a transformerless design with variable z. I guess
Panasonic bipolar caps would be best?

What are you going to do with the bipolar caps?

ulysses

Justin Ulysses Morse wrote in message ...
R. Foote wrote:

I think I will try a transformerless design with variable z. I guess
Panasonic bipolar caps would be best?

What are you going to do with the bipolar caps?

ulysses

The de-coupling caps in a transformerless preamp's input structure.

Thanks

R. Foote wrote:
Justin Ulysses Morse wrote in message ...
R. Foote wrote:

I think I will try a transformerless design with variable z. I guess
Panasonic bipolar caps would be best?

What are you going to do with the bipolar caps?

The de-coupling caps in a transformerless preamp's input structure.

Decoupling or coupling? Decoupling caps should be the last thing to worry
about.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Scott Dorsey wrote:

R. Foote wrote:
The de-coupling caps in a transformerless preamp's input structure.


Decoupling or coupling? Decoupling caps should be the last thing to worry
about.

Perhaps he's referring to DC blocking caps, to keep the phantom supply
out of the inputs. Non-polar would be good for when the phantom power
is off. Maybe an input amp that can run with +48 on its inputs could
even eliminate those. After all, it *is* common-mode.

Steve O'Neill wrote:
Scott Dorsey wrote:
R. Foote wrote:
The de-coupling caps in a transformerless preamp's input structure.
Decoupling or coupling? Decoupling caps should be the last thing to worry
about.

Perhaps he's referring to DC blocking caps, to keep the phantom supply
out of the inputs. Non-polar would be good for when the phantom power
is off. Maybe an input amp that can run with +48 on its inputs could
even eliminate those. After all, it *is* common-mode.

For input blocking caps, I see no reason not to just go with huge mylars.
Depends on your input Z of course, but 22 uF is plenty.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Howdy Steve...

I meant DC blocking. That's the trouble with looking at a schematic
for one thing and talking about something else.

Scott, I might try the mylar idea. I got to thinking about non-polars
from a recommendation from Deane Jensen's application notes.

thanks all!

In article ,
R. Foote wrote:
I meant DC blocking. That's the trouble with looking at a schematic
for one thing and talking about something else.

Scott, I might try the mylar idea. I got to thinking about non-polars
from a recommendation from Deane Jensen's application notes.

The upside is that they'll last forever, unlike an electrolytic. The
downside is that they're physically large and somewhat costly. The
physical size can cause some problems with coupling between different
parts of a circuit, especially if you're retrofitting film caps into
an existing circuit, floating them above the circuit board. If you're
making a new layout, this isn't a problem at all, but make sure not to
run other stuff near the input caps, especially the amp's output
traces.

Finally, IMHO, mylar won't sound as good as a fresh electrolytic, but
a better dielectric like polypropylene will. 10uF is probably the
bare minimum size and 22uF will be enough.


Best of luck,

Monte McGuire

Finally, IMHO, mylar won't sound as good as a fresh electrolytic, but
a better dielectric like polypropylene will. 10uF is probably the
bare minimum size and 22uF will be enough.


Best of luck,

Monte McGuire


What are your thoughts on solid tantalum?

In article ,
R. Foote wrote:
What are your thoughts on solid tantalum?

They're pretty fragile and they _must_ be run forward biased. If
phantom is _always_ on, then one might be able to get away with using
one as an input blocking cap, but they are very very picky about being
reverse biased. The only place I'll use them nowadays is in single
supply, class A amplifiers as cathode or emitter bypasses. They have
an interesting sound there...

But, even if you think it can almost always be forward biased,
consider what happens when it fails in a mike pre. Your preamp front
end or the mike or both might get seriously annoyed by a randomly
shorted out component that's biased up by P48. That could be pretty
expensive.

For this application, phantom power DC blocking, I'd never consider a
tantalum electrolytic. It's just too risky, and 40uF total of 60V
high quality (Kemet brand) tantalum is pretty darn expensive anyway.

These days, a nice Panasonic FC (or the obsolete HFQ) series aluminum
electrolytic seems to make the most sense. You'll probably have to
swap them out every 5-10 years or so, but they do sound good and you
can get a lot of capacitance in a small place. In some cramped
layouts, the small size of an aluminum coupling cap makes the circuit
work better than a film, if only because the thing isn't marginally
unstable from random coupling to other parts of the circuit.

For a new design where you have the space to avoid coupling problems,
a quality metallized polypropylene (or two or three) would work quite
well. Use 10-20uF per input leg at the smallest voltage rating you
can get - probably 100V but maybe you could find 63V rated parts
(which would be smaller). The best caps are noninductively wound and
use pure copper leads (not copper clad steel). You can go for
MultiCaps and other exotica, but I see no reason as long as the film
is clean, the foil is clean and the cap is put together competently.

A simple dielectric absorption test seems to weed out the bad caps and
it's easy to do. Take a 9V battery, charge up the cap, remove the
battery, short the cap for exactly one second and then measure the
voltage that ends up across the cap. It should be in the millivolt
range with a good cap, the lower the better. BTW, this is an easy way
to identify the type of dielectric used in a cap when you're at a
hamfest or surplus store. A 9V battery and a cheapo multimeter work
just fine. You'll find that polystyrene, polypropylene and teflon
have the lowest recovery voltage, polycarbonate are higher and mylar
is higher still. I've seen some defective polypropylenes that read
high with this test, so it seems to be a good thing to do.



Best of luck,

Monte McGuire

Monte P McGuire wrote:

These days, a nice Panasonic FC (or the obsolete HFQ) series aluminum
electrolytic seems to make the most sense.

For a new design where you have the space to avoid coupling problems,
a quality metallized polypropylene (or two or three) would work quite
well. Use 10-20uF per input leg at the smallest voltage rating you
can get - probably 100V but maybe you could find 63V rated parts
(which would be smaller). The best caps are noninductively wound and
use pure copper leads (not copper clad steel).

I was wondering about this lately while flipping through the Digikey
catalog. Other than the fact that it's usually the better caps that
they bother using OFC leads for, how much does it really matter if the
leads are copper-clad steel? I mean obviously copper is a better
conductor than steel, is the difference measurable?

In any event, taking a small, strong magnet to the suprlus shop sounds
like a good idea as well. I went through my capacitor bin with a
magnet recently and found about half the stuff I had on hand was
copper, half steel. Not counting electrolytics, which are all steel.
Is the steel cheaper, or is it a matter of current handling?

A simple dielectric absorption test seems to weed out the bad caps and
it's easy to do. Take a 9V battery, charge up the cap, remove the
battery, short the cap for exactly one second and then measure the
voltage that ends up across the cap. It should be in the millivolt
range with a good cap, the lower the better. BTW, this is an easy way
to identify the type of dielectric used in a cap when you're at a
hamfest or surplus store. A 9V battery and a cheapo multimeter work
just fine. You'll find that polystyrene, polypropylene and teflon
have the lowest recovery voltage, polycarbonate are higher and mylar
is higher still. I've seen some defective polypropylenes that read
high with this test, so it seems to be a good thing to do.

Good idea. But I bet it's very hard to short each cap the exact same
amount of time. Differences you measure could be due to timing errors.

It's interesting to note how in general, all the relevant properties of
capacitors tend to follow one another among different dielectrics.
Mylar, polypropylene, polystyrene, teflon seems to be an ascending
order of preference almost no matter the qualitative parameter. Size
and cost go the other direction. Monte, what do you know about the PPS
film caps, such as the Panasonic ECHS series? I gather there's been
some supply disruptions due to materials availability. But are they
worth it?

ulysses

Justin Ulysses Morse wrote:
Monte P McGuire wrote:

These days, a nice Panasonic FC (or the obsolete HFQ) series aluminum
electrolytic seems to make the most sense.

For a new design where you have the space to avoid coupling problems,
a quality metallized polypropylene (or two or three) would work quite
well. Use 10-20uF per input leg at the smallest voltage rating you
can get - probably 100V but maybe you could find 63V rated parts
(which would be smaller). The best caps are noninductively wound and
use pure copper leads (not copper clad steel).

I was wondering about this lately while flipping through the Digikey
catalog. Other than the fact that it's usually the better caps that
they bother using OFC leads for, how much does it really matter if the
leads are copper-clad steel? I mean obviously copper is a better
conductor than steel, is the difference measurable?

The argument is that the steel/copper junction makes a little semiconductor
and causes a little bit of nonlinearity at low levels.

I don't know about cap leads, but I know I can tell the difference in sound
between a couple hundred feet of RG-174 which has a steel-clad center
conductor, and a run of a similar copper-core cable. There's a difference
between a couple hundred feet and a 1/8 long cap lead, though.

In any event, taking a small, strong magnet to the suprlus shop sounds
like a good idea as well. I went through my capacitor bin with a
magnet recently and found about half the stuff I had on hand was
copper, half steel. Not counting electrolytics, which are all steel.
Is the steel cheaper, or is it a matter of current handling?

The steel is cheaper and easier to manufacture. Also since it is stiffer,
it's less apt to get deformed in transit and jam-up on pick and place machines.
All of the parts you buy today are basically designed for machine insertion.

A simple dielectric absorption test seems to weed out the bad caps and
it's easy to do. Take a 9V battery, charge up the cap, remove the
battery, short the cap for exactly one second and then measure the
voltage that ends up across the cap. It should be in the millivolt
range with a good cap, the lower the better. BTW, this is an easy way
to identify the type of dielectric used in a cap when you're at a
hamfest or surplus store. A 9V battery and a cheapo multimeter work
just fine. You'll find that polystyrene, polypropylene and teflon
have the lowest recovery voltage, polycarbonate are higher and mylar
is higher still. I've seen some defective polypropylenes that read
high with this test, so it seems to be a good thing to do.

Good idea. But I bet it's very hard to short each cap the exact same
amount of time. Differences you measure could be due to timing errors.

I strongly recommend the ESR meter from Dick Smith electronics. Doesn't do
real DA tests, but it does a great job for selecting capacitors and also
for in-circuit capacitor testing. I have been amazed how much easier
troubleshooting is with that thing.

It's interesting to note how in general, all the relevant properties of
capacitors tend to follow one another among different dielectrics.
Mylar, polypropylene, polystyrene, teflon seems to be an ascending
order of preference almost no matter the qualitative parameter. Size
and cost go the other direction. Monte, what do you know about the PPS
film caps, such as the Panasonic ECHS series? I gather there's been
some supply disruptions due to materials availability. But are they
worth it?

The PPS films aren't so bad. The polycarbonate stuff was wonderful... density
almost as good as mylar, sound quality and DA specs almost as good as
polystrene, and low price. Then the one manufacturer of the film raised his
prices and most folks discontinued them and the ones that are left are not
cheap any more.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Steve O'Neill wrote:

Perhaps he's referring to DC blocking caps, to keep the phantom supply
out of the inputs. Non-polar would be good for when the phantom power
is off. Maybe an input amp that can run with +48 on its inputs could
even eliminate those. After all, it *is* common-mode.

FWIW, I played around with the idea of letting the 48Vdc just ride on
through to eliminate the capacitors, but I could never achieve the
necessary CMRR to avoid contaminating the audio. My latest attempt
involves dropping the 48Vdc down to about 5Vdc, since a smaller voltage
ought to be easier to make disappear, or at least not cause problems at
the first stage output.

Bill

In article ,
Justin Ulysses Morse wrote:
I was wondering about this lately while flipping through the Digikey
catalog. Other than the fact that it's usually the better caps that
they bother using OFC leads for, how much does it really matter if the
leads are copper-clad steel? I mean obviously copper is a better
conductor than steel, is the difference measurable?

As Scott mentioned, the possibility of a nonlinear junction exists,
and also at the very least, you have a little bit extra inductance
from the ferrous lead material. Probably not a big deal, but what the
heck. I have used some of the Panasonic caps with copper clad steel
leads and they seem to work OK, but I'm not crazy about the concept.

Good idea. But I bet it's very hard to short each cap the exact same
amount of time. Differences you measure could be due to timing errors.

Try it out and see. When you short out the main charge on a cap, the
secondary charge doesn't get shorted out right away and it slowly
recovers in a few seconds. So, the amount of recovery voltage is not
all that dependent on how long you short out the cap. In practice, I
found it good enough to identify different dielectrics. And heck,
you're an engineer, right? You've probably got a good enough sense of
timing from punching in tape machines, so you can probably short some
leads pretty consistently.

It's interesting to note how in general, all the relevant properties of
capacitors tend to follow one another among different dielectrics.
Mylar, polypropylene, polystyrene, teflon seems to be an ascending
order of preference almost no matter the qualitative parameter. Size
and cost go the other direction. Monte, what do you know about the PPS
film caps, such as the Panasonic ECHS series? I gather there's been
some supply disruptions due to materials availability. But are they
worth it?

PPS is useful because it can withstand the heat of surface mount
reflow. Most all of the other dielectrics can't, although I'm hoping
for a Teflon comeback for this reason alone... ;-)

I don't know how good it sounds, but it may be OK... never tried it
yet. The numbers aren't fantastic, but it may still sound OK.


Best of luck,

Monte McGuire

In article ,
Justin Ulysses Morse wrote:
I think it makes sense to build separate preamps for use with and
without phantom power. Build a transformerless preamp with phantom
power (and electrolytic blocking caps) for use with condensers, and you
never have to turn off the phantom power, so the caps are always
biased. Build a separate preamp for dynamics and eliminate the
coupling caps entirely in that one.

A great idea... there's no cap better than no cap at all.

The design challenge in a mic preamp is trying to accommodate such a
huge range of input signals without sacrificing performance. The
trade-off between input overload and maximum gain is a tough one when
some mikes put out 5mV and others put out 5V. I think if you're
building your own mike preamps, you'll get much better results if you
don't make compromises to accommodate every imaginable microphone with
one preamp design. A BK-5 has different needs than a TLM103. If
you're building preamps for your own personal use, it makes sense to
think about what you're going to use them with and build them
accordingly. For example, you could build one preamp with a hefty
step-up transformer and no phantom power on the input and a gain range
of 50-70dB just for use with low-output ribbons on vocals. You could
build another preamp that would have no gain at all, just phantom power
(and maybe a JT-16B input transformer), for use with ultra-sensitive
mikes like the TLM103. The whole point of DIY is you get exactly what
you want and need. You just have to decide what that is.

Great ideas here - worth quoting again...

For 'colorless' preamps, perhaps this isn't so important, because it's
not that hard to get a similar tone across a wide input / gain range
if you're striving for complete accuracy. Of course, reality sets in
and nothing's all that colorless, but you get my drift: many of the
transformerless designs are pretty much uniform sounding across their
gain range, as long as you aren't asking for gobs of gain.

But... for deliberately colored preamps, it is definitely worth
optimizing a preamp to a mike to an input signal. Many of the preamps
I use for lead vocals have a sweet spot that isn't all that wide, and
if you're lucky enough to get a singer and mike and preamp that hits
the right spot, things work really well. If you fall out of that
level/gain pocket though, the preamp won't sound very good at all...

There are some topologies that help you widen that sweet spot, and
this can sometimes help out a lot. The A-Designs preamp does this
quite well. It's a fixed gain amp preceded by a volume control and a
transformer, so you can hit the sweet spot pretty easily by adjusting
the attenuator after the input tranny. Turn the pot up for more
drive, turn it down for less. Pretty simple.

I made some low gain mike preamps designed for drum use that use a
similar design - a fixed gain amp driven by a transformer feeding a
pot. It's a great idea for a limited range amp - this thing is
designed for drum use, so maximum gain is pretty low, around 30dB or
so. But, minimum gain is also quite low (negative infinity, because
of the input pot), so it's easy to handle high outputs from close
mikes on loud drums.

The only thing left 'flapping in the breeze' is the input transformer,
which has to handle the full mike input signal. Fortunately, there
are quite a few low to medium ratio transformers that can handle high
input levels without problems. I used the UTC A20 as the input tranny
for this preamp, wired as a 1:2 step up, and it can handle +15dBm
according to the manufacturer. In practice, it's worked just fine.

Another concept related to all of this is "what are you tracking to"
and what levels it can deal with. When I ran analog tape, the signals
I asked from my mike preamps were a good bit lower than what I ask of
them when tracking to a modern ADC. This means that if you have a
signal at a preamp's sweet spot that causes an output of +10dBV, then
it might be tough to deal with this preamp in an analog tape
situation, but it might work just fine with digital where converter
sensitivities are a good bit lower. Sometimes it's worth
recalibrating a converter to a different sensitivity to deal with a
specific preamp's sweet spot.

As you said initially, all this stuff is under your control if you
make your own preamps, but also keep in mind that the same gain
staging concepts apply to storebought electronics too. Evaluating a
preamp has to include deciding what sort of input range the pre is
happy with and what sort of output levels it's best at providing, and
somehow scaling that to your recording devices with pads, trims or
clever console use.


All the best,

Monte McGuire