[Paul[_13_]]

I promise I won't pretend to understand all
of this presentation if you don't!



http://www.thatcorp.com/datashts/AES...ic_Preamps.pdf

It's still very interesting reading.

I'm gonna assume they can't be too worried about
revealing too many trade secrets in an online document,
as reverse-engineering is very easy to do, and as we
all know, all companies hire engineers from their
competition anyways!

SECRETS DON'T STAY SECRET FOR VERY LONG!

And I'm sure all of this stuff is common
knowledge anyways.....

[Scott Dorsey]

In article , Paul wrote:

http://www.thatcorp.com/datashts/AES...ic_Preamps.pdf

It's still very interesting reading.

Gary has been giving this presentation for several years in various forms,
and it gets better every year.

Most of it is common knowledge but some of it (like phantom fault protection)
really isn't very well-known in the industry at all.

I'm gonna assume they can't be too worried about
revealing too many trade secrets in an online document,
as reverse-engineering is very easy to do, and as we
all know, all companies hire engineers from their
competition anyways!

The people from THAT want people to be building as many mike preamps
as possible using their ICs, and they don't want those preamps to have poor
performance or poor reliability because that would affect the reputation of
their components. So it's in their best interest to teach designers how to
make the best possible preamplifiers that they can.

SECRETS DON'T STAY SECRET FOR VERY LONG!

And I'm sure all of this stuff is common
knowledge anyways.....

Some of it is... the whole long-tailed pair preamp topology has been around
in the audio industry since the 1970s when the Trans-Amp came out. But an
awful lot of engineers never got told about all the things that affect CMRR,
and I don't think any of us ever thought about having to protect from impulses
when the phantom power caps discharge. They aren't trade secrets by any means,
they're just things that nobody ever mentions in school.

But... you'll notice that just about everything discussed in the first half
of the paper.... CMRR issues, feedback resistor issues, DC bias issues,
and protection issues... all of them can be readily solved by dropping a
transformer in front. Of course, you get a whole other pile of problems
with the transformer but for that you need to read the Jensen app notes instead
of the THAT ones.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Phil Allison[_4_]]

Scott Dorsey wrote:



Most of it is common knowledge but some of it (like phantom fault protection)
really isn't very well-known in the industry at all.


** Not in my experience.

A transformerless pre-amp that had no protection from phantom voltage accidents would soon be toast.

Most only have a couple of 1 amp diodes to stop the input transistors from being reverse biased ( C-E or B-E ) by a short on the signal line when the phantom supply is on - but that is all you need in practice.

What seems to be true, however, is that single chip mic pre-amps ( eg Analog Devices SSM2017 ) need more elaborate protection.


..... Phil

[Paul[_13_]]

On 2/21/2017 10:45 AM, Scott Dorsey wrote:
In article , Paul wrote:

http://www.thatcorp.com/datashts/AES...ic_Preamps.pdf

It's still very interesting reading.

Gary has been giving this presentation for several years in various forms,
and it gets better every year.

Most of it is common knowledge but some of it (like phantom fault protection)
really isn't very well-known in the industry at all.

I'm gonna assume they can't be too worried about
revealing too many trade secrets in an online document,
as reverse-engineering is very easy to do, and as we
all know, all companies hire engineers from their
competition anyways!

The people from THAT want people to be building as many mike preamps
as possible using their ICs, and they don't want those preamps to have poor
performance or poor reliability because that would affect the reputation of
their components. So it's in their best interest to teach designers how to
make the best possible preamplifiers that they can.

"THAT" makes sense! PUN INTENDED!


SECRETS DON'T STAY SECRET FOR VERY LONG!

And I'm sure all of this stuff is common
knowledge anyways.....

Some of it is... the whole long-tailed pair preamp topology has been around
in the audio industry since the 1970s when the Trans-Amp came out. But an
awful lot of engineers never got told about all the things that affect CMRR,
and I don't think any of us ever thought about having to protect from impulses
when the phantom power caps discharge. They aren't trade secrets by any means,
they're just things that nobody ever mentions in school.

But... you'll notice that just about everything discussed in the first half
of the paper.... CMRR issues, feedback resistor issues, DC bias issues,
and protection issues... all of them can be readily solved by dropping a
transformer in front. Of course, you get a whole other pile of problems
with the transformer but for that you need to read the Jensen app notes instead
of the THAT ones.

Ah yes, a good review for me:

Energy stored in a capacitor = (0.5)*C*V**2

http://www.thatcorp.com/datashts/AES...tom_Menace.pdf

And it's interesting that they are using back-to-back Schottky
and Zener diodes for phantom fault protection and ESD. We used
Schottky for ESD protection on GSM/DCS cellular phone power amplifiers.

Very, very similar problems and solutions.

So if you put the transformer in the front, can you safely
leave out the back-to-back protection diodes?

[Scott Dorsey]

In article , Paul wrote:
And it's interesting that they are using back-to-back Schottky
and Zener diodes for phantom fault protection and ESD. We used
Schottky for ESD protection on GSM/DCS cellular phone power amplifiers.

The problem with back to back diodes is that they are never totally off
so with a very high-Z input the leakage can be an issue. It still gives me
the willies even with a mike preamp, but I can live with it. 1N4148s to the
supply rails don't leak because the diodes are totally turned off, and they
don't act as parasitic varactors because they are totally turned off so the
capacitance is fixed.

So if you put the transformer in the front, can you safely
leave out the back-to-back protection diodes?

Think about it. You have a 1:5 step-up transformer in the front, so now you
can ditch the long-tailed pair and just have a single transistor for your
input stage. You get "free gain" from the better match with the step-up,
and now _all_ of the differential stuff is being done by the transformer.
Plus you get free low-pass from the transformer.

You don't need _any_ common-mode clamping on the input, because the
transformer is going to eat any common mode transient, even hundreds of
volts worth. You don't need to protect the transformer. So you can
put your protection on the secondary of the transformer and now the protection
is on a single lead. It doesn't need to be matched or balanced in any way,
it's just a single clamp now. And you can just clamp to the rails with
1N4148s, you don't need anything else.

Switch the phantom on or off, the transformer is going to eat it all.
Maybe you get a little peak on the output if the phantom resistors are
very badly mismatched.

Now... because of the step-up you do need clamps that can handle higher
voltage, but the job of the clamps becomes a lot easier.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."