Design review for portable mic preamp
Henry
designed a circuit around the LT1115 op amp. I thought it would be best
to get some feedback on this design from various experts before I
breadboard it, so I've put a schematic at
http://henryf.50megs.com/preamp so you can see what I have in mind. Do
you have any comments or suggestions for improvement?
Thanks,
-Henry
hen...@erols.com
Thomas Ludwig
The LT 1115 typically has an open loop gain of 70dB @ 20 kHz. So, at a
closed loop gain of 60 dB, your loop gain (feedback...) is 10dB, which
is too low. This will cause high frequency roll-off and
distortion/intermodulation at high frequencies. I wouldn't recommend a
gain higher than 40dB for a single stage.
At the lowest gain (do you really need this?), the impedance of the
feedback network (537.5 ohms) is a bit on the low side.
Finally: Are you sure that ~8kohms is a good input impedance for a
microphone preamp?
--
Thomas Ludwig
Institute of Mineralogy
University of Heidelberg, Germany
Eric Toline
They have a new 2 channel version comming out soon.
Norbert Hahn
>I'm thinking about building a portable microphone preamplifier and have
>designed a circuit around the LT1115 op amp. I thought it would be best
What kind of mics do you want to use whith this pre-amp? I don't know much
about instrument pick up, but for dynamic mics I would not use the protection
diodes D1 and D2 as shown but rather clamp them to the supply rails. This
ensures that under normal operating condition the diodes do not conduct.
What is C2 good for? Cancellation of RF pickup?
Have you looked at the application notes of the SSM 2017 ? This beast is a
true mic preamp for balanced mics. I've used its brother SSM 2016 in the past
which has lower noise at low impedances but it has been discontinued after PMI
was taken over by AMD. However, you should check that the SSM 2017 can be run
at +- 9V.
Norbert
David Josephson
>I'm thinking about building a portable microphone preamplifier and have
>designed a circuit around the LT1115 op amp. I thought it would be best
>to get some feedback on this design from various experts before I
>breadboard it, so I've put a schematic at
>http://henryf.50megs.com/preamp so you can see what I have in mind. Do
>you have any comments or suggestions for improvement?
It's hard to comment without knowing what your design goals were, but
it would seem that you intend using only unbalanced mics. If that's the
case (stealth mics using cheap electret capsules for instance) it's
probably okay, however most of those applications need a fairly sharp
low cut filter as an option, and you may need a lower gain setting
than 16 dB if the venue is loud. The LT1115 is a good part, rumored to
be the same die as the LT1028 but with relaxed DC specifications.
Be sure that C7-10 are physically adjacent to the chip.
--
David Josephson / Josephson Engineering / San Jose CA / da...@josephson.com
David Collins
<dav...@rahul.net> wrote:
>The LT1115 is a good part, rumored to
> be the same die as the LT1028 but with relaxed DC specifications.
> Be sure that C7-10 are physically adjacent to the chip.
Is that really true? In my admittedly limited use the 1115 sounded better
are was _slightly_ quieter than the 1028.
DC
--
Dave Collins Entropy just isn't what it used to be!
dcol...@earthlink.net dcol...@well.com
Henry
Thanks for looking at the circuit. I've added my responses to your
message below.
-Henry
Thomas Ludwig wrote:
> The LT 1115 typically has an open loop gain of 70dB @ 20 kHz. So, at a
> closed loop gain of 60 dB, your loop gain (feedback...) is 10dB, which
> is too low. This will cause high frequency roll-off and
> distortion/intermodulation at high frequencies. I wouldn't recommend a
> gain higher than 40dB for a single stage.
You are right about the typical gain, and the worst-case gain would be
about 4 dB less.
The pdf data sheet for the LT1115 shows typical performance curves
of THD + noise vs. frequency, for gains of +/-100 and +/-1000. For a
gain of +100, they show THD + noise of 0.01% at 20 kHz, and for a
gain of +1000, it looks like 0.025% at 20 kHz.
I don't understand how they are doing so well with so little feedback,
but that's what the typical performance curves say.
My results may vary, but if I could get 0.025% THD + noise with one
60 dB gain stage at 20 kHz, I'd be willing to live with that to nearly
double the battery life. With any luck, I won't be using full gain all the
time so the THD should be even better.
> At the lowest gain (do you really need this?), the impedance of the
> feedback network (537.5 ohms) is a bit on the low side.
I don't think I'll need it for the type of recording I plan to do, since I
plan to do my recording at reasonable listening levels some distance
from performers. But, you never know... :^)
> Finally: Are you sure that ~8kohms is a good input impedance for a
> microphone preamp?
Normally, SW1 would be closed, so the input impedance would be
closer to 1580 ohms. If I ever need to amplify something with a
higher impedance, or if I just want to see what difference it makes,
then I can flip the switch.
Henry
> What kind of mics do you want to use whith this pre-amp? I don't know much
> about instrument pick up, but for dynamic mics I would not use the protection
> diodes D1 and D2 as shown but rather clamp them to the supply rails. This
> ensures that under normal operating condition the diodes do not conduct.
I have an Audio Technica AT-822 stereo mic that I was planning to use. It's a
battery-powered electret with 200 ohms impedance, and a 3-pin connector wired
for unbalanced stereo. AT includes an adapter that converts from 3-pin to an
1/8" stereo plug. So, I guess I'd expect decent unbalanced mics, suitable for
personal recording via minidisk or cassette.
The explanation of D1 and D2 is long. Note 4 in the LT1115 data sheet says:
"The inputs are protected by back-to-back diodes. Current limiting
resistors
are not used in order to achieve low noise. If differential input
voltage
exceeds +/- 1.8V, the input current should be limited to 25 mA"
My interpretation of this is that when the + input goes above 1.8 V, current will
start to flow throught those back-to-back diodes, out the - input, and through
R6 to ground. Since R6 is 84.5 ohms, at 25 mA there would be 2.1 V across
R6. So, I need to limit the voltage at the + input to less that 1.8V + 2.1V =
3.9V.
The backward-biased Zener should break down at 2.4 V, and the forward drop
of the other Zener might be 0.6 V, so it looks like the input would be limited at
2.4 + 0.6 = 3.0 V or so, safely less than the 3.9 V limit. The Zeners should be
good for 0.5 Watt, or 128 mA, of shunting before they fail.
I'm somewhat new at this, and the circuit may work perfectly fine in real life
without D1 and D2. But, I wouldn't want to find out later that the LT115 could
be ruined by a static zap on the mic input momentarily exceeding the 25 mA limit.
> What is C2 good for? Cancellation of RF pickup?
Yes, that's the hope, anyway.
> Have you looked at the application notes of the SSM 2017 ? This beast is a
> true mic preamp for balanced mics. I've used its brother SSM 2016 in the past
> which has lower noise at low impedances but it has been discontinued after PMI
> was taken over by AMD. However, you should check that the SSM 2017 can be run
> at +- 9V.
I've quickly looked at the SSM 2017 but was somewhat put off since it wasn't
stocked by Digikey. I've since found out that Newark carries it, so I'll take a
closer
look at it.
-Henry
Henry
About the low pass filter, what would you suggest? How about a
second order filter at 100 Hz?
Adding a switchable 20 dB pad should be easy enough.
-Henry
David Josephson wrote:
> It's hard to comment without knowing what your design goals were, but
> it would seem that you intend using only unbalanced mics. If that's the
> case (stealth mics using cheap electret capsules for instance) it's
> probably okay, however most of those applications need a fairly sharp
> low cut filter as an option, and you may need a lower gain setting
> than 16 dB if the venue is loud. The LT1115 is a good part, rumored to
mcp
> The explanation of D1 and D2 is long. Note 4 in the LT1115 data sheet says:
>
> "The inputs are protected by back-to-back diodes. Current limiting
> resistors
> are not used in order to achieve low noise. If differential input
> voltage
> exceeds +/- 1.8V, the input current should be limited to 25 mA"
>
> My interpretation of this is that when the + input goes above 1.8 V, current will
> start to flow throught those back-to-back diodes, out the - input, and through
> R6 to ground. Since R6 is 84.5 ohms, at 25 mA there would be 2.1 V across
> R6. So, I need to limit the voltage at the + input to less that 1.8V + 2.1V =
> 3.9V.
I've always heard that zeners create noise...even to the point that a
properly reverse biased zener can be used as a white noise source.
David Josephson
>Hi David,
>About the low pass filter, what would you suggest? How about a
>second order filter at 100 Hz?
Second order is not enough, and 100 Hz may not be the right frequency.
I would look for 3rd order at a minimum, switchable 40-80-160 or
something like that.
>Adding a switchable 20 dB pad should be easy enough.
Yes.
>-Henry
>David Josephson wrote:
>> It's hard to comment without knowing what your design goals were, but
>> it would seem that you intend using only unbalanced mics. If that's the
>> case (stealth mics using cheap electret capsules for instance) it's
>> probably okay, however most of those applications need a fairly sharp
>> low cut filter as an option, and you may need a lower gain setting
>> than 16 dB if the venue is loud. The LT1115 is a good part, rumored to
>> be the same die as the LT1028 but with relaxed DC specifications.
>> Be sure that C7-10 are physically adjacent to the chip.
--
michael gaster
I have mixed on several desk that offer different means of HPFs' and the
best are those that are sweepable from 20Hz to 400Hz granted I never filter
anything above 200 unless its is a high hat, anyway the best HPF/LPF's have
a slope of 12 to 18 dB per octave, using HPF with slopes steeper than 18dB
don't sound natural or smooth. i do live sound if that helps.
michael
Henry <hen...@erols.com> wrote in message
news:38714FD1...@erols.com...
> Hi David,
>
> About the low pass filter, what would you suggest? How about a
> second order filter at 100 Hz?
>
> Adding a switchable 20 dB pad should be easy enough.
>
Henry
> I've always heard that zeners create noise...even to the point that a
> properly reverse biased zener can be used as a white noise source.
I've heard that too, but I thought it only applied when the voltage on
the zener was high enough for it to conduct. If the voltage on the zeners
gets high enough for them to conduct in this case, there are going to be
bigger problems with the sound than the diode noise. ;^)
Does anyone know offhand how to properly model the noise
from a zener that is biased below Vz, or where to find such
measurements?
Thanks,
Henry
Norbert Hahn
>I have an Audio Technica AT-822 stereo mic that I was planning to use. It's a
>battery-powered electret with 200 ohms impedance, and a 3-pin connector wired
>for unbalanced stereo.
So forget the SSM 2017. It's intended for balanced mics. If you use it in an
unbalanced setup you loose signal to noise ration because 1/2 of the amp does
not receive a signal.
>My interpretation of this is that when the + input goes above 1.8 V,
... of the - input ...
>current will
>start to flow throught those back-to-back diodes, out the - input, and through
>R6 to ground. Since R6 is 84.5 ohms, at 25 mA there would be 2.1 V across
>R6. So, I need to limit the voltage at the + input to less that 1.8V + 2.1V =
>3.9V.
Unfortunately you calculation is correct. The problem lies in the value of R6
which is way too lows. Why don't you use 8.45 kOhms. I agree with Thomas Ludwig
who pointed out that your feedback network has too low an impedance.
>
>The backward-biased Zener should break down at 2.4 V,
Have you had a look at the i/v diagram of such a Zener diode? It is actually
made of 4 "normal" diodes connected in series... Hm, I don't like such a beast
in any audio application. You could replace it by the CE part of a transistor,
however, that "zener" voltage is about 9 V (which is too much in your
current concept). The main advantage of a transistor is that it cuts off a lot
better than a 8.2 V Zener, the only disadvantage is, that the CE break down
voltage is not closely specified.
>I'm somewhat new at this, and the circuit may work perfectly fine in real life
>without D1 and D2.
.. but your assumption is correct: Real life can be brutal.
>> What is C2 good for? Cancellation of RF pickup?
>
>Yes, that's the hope, anyway.
I would reduce its value and insert a coil in the input lead. Thus you get a
real LC low pass with a higher slope and less influence in to audio band.
>I've quickly looked at the SSM 2017 but was somewhat put off since it wasn't
>stocked by Digikey. I've since found out that Newark carries it, so I'll take a
>closer
>look at it.
The SSM 2017 would be quite usefull with balanced feed. In case that you need a
very low noise design you might think of a step up transformer. One of the main
draw backs of the 2017 is its current consumption. You should compare it with
2 * LT 1115. As Thomas Ludwig wrote, a single LT 1115 might not be sufficient.
>
>-Henry
>
Norbert
Norbert Hahn
>I've always heard that zeners create noise...even to the point that a
>properly reverse biased zener can be used as a white noise source.
That's true for true Zeners when they start conducting. However, the
diodes in question do not use the avalanche effect. The are made up
of 4 p/n junctions in series to give the 2.4 V. They are used in
a conduction setup, hence some current flows no matter what voltage
is applied. And p/n is non-linear...
Norbert
Norbert Hahn
>In <38714FD1...@erols.com> Henry <hen...@erols.com> writes:
>
>>Hi David,
>
>>About the low pass filter, what would you suggest? How about a
>>second order filter at 100 Hz?
>
>Second order is not enough, and 100 Hz may not be the right frequency.
>I would look for 3rd order at a minimum, switchable 40-80-160 or
>something like that.
You should as well consider the recording device. I use two DAT decks,
the Sony DTC ZA5 ES goes down to 2 Hz and definitely needs a HP filter
in a live recording. My mic amp cuts below 18 Hz @ 18 dB /octave
(Chebychev type filter). My other deck, a TASCAM DA-30 mkII has already
a build in LP filter, at about 20 Hz.
So I have the option to filter in the digital domain after the recording
giving me an undo function....
Norbert
Monte P McGuire
Norbert Hahn <ha...@hrz.tu-darmstadt.de> wrote:
>Unfortunately you calculation is correct. The problem lies in the
>value of R6 which is way too lows. Why don't you use 8.45 kOhms. I
>agree with Thomas Ludwig who pointed out that your feedback network
>has too low an impedance.
I disagree. R6 is right at the negative input terminal, so any noise
caused by that resistance goes directly into the amplifier. What's
worse is that any input noise current from the 1115s input stage will
be impressed upon R6 (paralleled with the feedback resistor): make R6
large and you get more noise.
Most low En amplifiers are designed to drive low impedance loads
precisely because the feedback networks have to have low impedance to
get low total noise. 84.5 ohms is a reasonably good value, since it's
a bit smaller than the 200 ohm source he'll have at the other input
terminal. Driving 500 ohms or so to 1V or thereabouts only requires a
few milliamps of drive current, and that's not a biggie. Sure,
distortion would be a little worse than if the load were higher, but
the only way out is to spend more of the supply current budget on a
buffer stage. 10mA of class A buffer stage bias would clean things
up, but then you're dissipating 10mA all the time to chase down a tiny
bit of distortion. If it wasn't an AT electret mike at the source,
then maybe it's worthwhile, but I wouldn't sweat it...
One thing that I think should be added is a coupling cap at the input.
I don't know if the AT mike has an output coupling cap, but it just
seems like a prudent thing to do to prevent DC from some other sort of
powered mike from getting into the input stage. I'd use something
like 10-20uF of metallized polypropylene in series with the input.
I'd also put an RF choke in series with the input to the box, right at
the jack, to further prevent RF from spoiling your day. C2 seems a
little large and it's not really needed if you have C3 there too. C2
will only make a difficult load for your electret mike to drive at
high frequencies and this will make the mike generate more distortion.
So, the input stage will have a series RF choke, then C2 as a shunt RF
capacitor to chassis and then a series coupling cap. Put the load
resistors, protection diodes and the amp after that.
About your load resistors: your AT mike doesn't need them to produce
the proper frequency response. It's an active mike and any load there
just makes it harder for the circuitry to drive. So, I'd skip the
switch and R4 and just use R3. The other (more important) purpose of
of R3 is to provide a path for input stage bias current for the 1115.
Yes, the 1115 is bias current compensated, but there will still be a
little bit of net input bias current, and if you use an input coupling
cap, the only place this current can come from is R3. So, you do need
it, just make it large (like the 7870 ohms that you already have) so
it won't be a significant load. Don't worry about noise from this
resistor, since it's in parallel with the 200 ohm mike.
About distortion performance: the 1115 has inherently low open loop
distortion, so it doesn't need a lot of feedback to make it linear.
That's why it still performs well even with low feedback. This is a
good thing...
A completely different approach you could take would be to use an
input transformer. However, since your mike is unbalanced, it's not
clear that you even need any sort of common mode rejection. I guess
the main thing you get from a transformer input is free gain and the
ability to use higher impedance circuitry around the amplifier with no
noise penalty. Transformers are not cheap however and they are not
that small either. So, for your application, the design you have
seems to be a good one.
One final thing: the gain switch you choose should be a 'shorting'
type, also called 'make before break'. The reason is that your
feedback loop opens up if the switch opens, and the amp goes to
maximum gain. This will happen each time you change gain, unless your
switch is a 'shorting' type.
The other way around this problem is to switch R6 and use a fixed
resistor for the feedback network. The advantage there is that if the
switch opens up, the amp goes to unity gain, not infinity gain. The
disadvantage is that at low gains, R6 will get somewhat large-ish and
the amp's noise figure gets worse. Fortunately, the amp also is run
at low gain in that condition, so the total noise at the output isn't
so large. Most switchable gain amps use this scheme just to make them
more reliable and having a worse noise figure at low gains is
surprisingly common. The other advantage is that you don't want a lot
of stray reactances around the feedback resistor. A rotary switch
network is gonna be a lot messier at a couple MHz than a simple 1/8W
feedback resistor right next to the amp, and that's another reason why
most people switch R6 and not the feedback resistor.
Best of luck,
Monte McGuire
mcg...@world.std.com
Norbert Hahn
>I disagree. R6 is right at the negative input terminal, so any noise
>caused by that resistance goes directly into the amplifier.
Umm, that's right. Now we should have the amount of self noise of the
mic, which can be expressed as a resistor...
> What's
>worse is that any input noise current from the 1115s input stage will
>be impressed upon R6 (paralleled with the feedback resistor): make R6
>large and you get more noise.
I wasn't aware that the LT 1115 has a bipolar input stage.
I downloaded the data sheet in the meanwhile.
The input noise current density is 1 pA sr Hz, thus a value of 1 kOhm
of R6 would add 1 nV sr Hz which is in the same magnitude as the input
voltage noise density (specified 0.9 nV sr Hz).
Thanks for the clarification!
-----
What about the input protection diodes? I stiff feel uncomfortable
with them, and unconfortable without them.
Norbert
Henry
I'm concerned about the input protection diodes now also. I've just
emailed Vishay to try to find out more about the 1N5221 than the data
sheets show.
-Henry
-----------------------------
Henry
The AT-822 mic has a switchable high-pass filter on it. Maybe on this first
go-around I'll use the KISS principle and leave the HPF off the mic preamp.
For an active HPF it seems that I'd need two extra op amps per channel, and
I'm not sure I'd want to budget that much power for the filters.
A quick look at switched-capacitor filters gives me the impression that
they'd add noise and possibly have aliasing problems. Is there some
brilliant solution to adding a fourth-order HPF to a portable mic preamp
that I'm missing?
-Henry
Norbert Hahn
>The AT-822 mic has a switchable high-pass filter on it. Maybe on this first
>go-around I'll use the KISS principle and leave the HPF off the mic preamp.
That depends on the cut-off frequency and on the slope. Usually the HPF of
mics are designed to give a slight roll off for close up miking of voice.
It may be 120 Hz at -6 dB per octave. It serves for the intended purpose
but think of a live recording in a medium and someone slams the door...
>
>For an active HPF it seems that I'd need two extra op amps per channel, and
>I'm not sure I'd want to budget that much power for the filters.
You are right, unfortunately. You may savely use the OPA 2604 or OPA 2134 for
that, but you need to pay for the condensors too. And you need to "pay" for
their power consumption if you want to feed your amp with batteries.
>
>A quick look at switched-capacitor filters gives me the impression that
>they'd add noise and possibly have aliasing problems.
I find switched capacitor filters horrible for that purpose too.
>Is there some
>brilliant solution to adding a fourth-order HPF to a portable mic preamp
>that I'm missing?
You should use a filter that has a good phase response in the passed band.
IMHO, this is more important than a ruler flat frequency response below
100 Hz. That's why I used a Chebychev design.
>
>-Henry
>
Later,
Norbert
Ken Nelson
I've been playing with breadboarding a similar amp, just using 5532A's - two
stages... it's going to be used with cheap electrets and a MD portable...use is
90% hobby, 10% multimedia sound-gathering.
For simplicity's sake, I've just been changing a series capacitance on the
preamp input to provide a simple HPF (6dB/oct when teamed with source impedance
& input load) , choosing values to provide selectable -3dB points of 20 Hz /150
Hz.
My reasons for this approach are
1) simple and cheap (like it's application... and it's designer ;^) )
2) it's before any preamp so I am possibly able to apply more gain without
clipping on extreme LF
3) it's similar to what many mic manufacturers do for onboard HPF in their mics
4) it's very smooth and no phase problems
5) it seems to be complementary to cheap electret omni's that often have too
much LF pick-up, especially outdoors
Initial tests of a prototype have been good, for me anyways.
My question/comment centers around whether it's really necessary or useful to
have an elaborate 3rd or 4th order HPF filter on such a portable preamp. My
reasons against having such a one would be
1) too complex for a portable device (difficulty to build and tune, parts count,
power)
2) more potential to screw up the sound from filter/phase artifacts or too much
cut
3) it's common to find more accurate, sweepable HPF or parametric filters in
consoles or inexpensive audio software (eg CoolEdit) that can be used
post-recording.
So, I would not myself choose to put an elaborate HPF in a basic field preamp. A
more elaborate film production field mixer, maybe...
Agree? Disagree? Discuss...
ken
Henry wrote:
> Norbert Hahn wrote:
>
> > David Josephson <dav...@rahul.net> wrote:
> >
> > >In <38714FD1...@erols.com> Henry <hen...@erols.com> writes:
> > >
> > >>Hi David,
> > >
> > >>About the low pass filter, what would you suggest? How about a
> > >>second order filter at 100 Hz?
> > >
> > >Second order is not enough, and 100 Hz may not be the right frequency.
> > >I would look for 3rd order at a minimum, switchable 40-80-160 or
> > >something like that.
> >
> > You should as well consider the recording device. I use two DAT decks,
> > the Sony DTC ZA5 ES goes down to 2 Hz and definitely needs a HP filter
> > in a live recording. My mic amp cuts below 18 Hz @ 18 dB /octave
> > (Chebychev type filter). My other deck, a TASCAM DA-30 mkII has already
> > a build in LP filter, at about 20 Hz.
> >
> > So I have the option to filter in the digital domain after the recording
> > giving me an undo function....
>
> The AT-822 mic has a switchable high-pass filter on it. Maybe on this first
>
> go-around I'll use the KISS principle and leave the HPF off the mic preamp.
>
> For an active HPF it seems that I'd need two extra op amps per channel, and
> I'm not sure I'd want to budget that much power for the filters.
>
> A quick look at switched-capacitor filters gives me the impression that
> they'd add noise and possibly have aliasing problems. Is there some
> brilliant solution to adding a fourth-order HPF to a portable mic preamp
> that I'm missing?
>
> -Henry
David Josephson
mic, you don't need them.
>-Henry
>-----------------------------
>Norbert Hahn wrote:
--
Thomas Ludwig
> The pdf data sheet for the LT1115 shows typical performance curves
> of THD + noise vs. frequency, for gains of +/-100 and +/-1000. For a
> gain of +100, they show THD + noise of 0.01% at 20 kHz, and for a
> gain of +1000, it looks like 0.025% at 20 kHz.
>
> I don't understand how they are doing so well with so little feedback,
> but that's what the typical performance curves say.
I have to admit that these values are impressive, I did not look much
beyond the open-loop-gain diagram...
But you will still get a 2..3 dB rolloff at 20 kHz.
>
> > At the lowest gain (do you really need this?), the impedance of the
> > feedback network (537.5 ohms) is a bit on the low side.
>
> I don't think I'll need it for the type of recording I plan to do, since I
> plan to do my recording at reasonable listening levels some distance
> from performers. But, you never know... :^)
The 84.5 ohms give you a noise density of 1.2 nV/sqrt(Hz).
So for high gain this is a good idea.
But: Think of the output voltage swing.
At Vs= +/-18V an Rload = 600ohms, the datasheet states +/-14V typ. and
+/- 10V min. Now imagine your batteries are not fully charged (8V). Than
you get an output voltage swing of +/- 4V typ. and 0V (well...) min. At
a load of 2 kohms these values (mostly the guaranteed ones) look much
better. This is why thought it would be a good idea to have a higher
impedance feedback network for the low gain.
Norm Strong
>>> What about the input protection diodes? I still feel uncomfortable
>>> with them, and unconfortable without them.
For the input protection diodes to do you any good, there must be a resistance
in the circuit. You have none. I'd recommend 2 diodes in series in each
direction, and make sure they're epoxy packaged. If glass packages are used,
any light will cause noise in the input. Now put a 100 ohm resistor in series,
so the diodes have something to clamp. This should not add appreciably to the
noise, since the noise resistance of your amplifier is already 750 ohms.
Norm Strong (nh...@aol.com)
2528 31st South, Seattle WA 98l44