Are multiple amplifier stages better than one ???

[jalbers@bsu.edu]

I have a few questions about the design of the circuit in Figure 1
page 71 located at https://ilocker.bsu.edu/users/jalbers/WORLD_SHARED/WireTracer.PDF

The gain of the first op amp stage is 10 and the gain fo the second op
amp stage is 33 for a total net gain of 330. Why not just use one op
amp and set the gain to 330?

When designing a circuit like this, why use a low gain of 330, why not
a higher value? How did the author arrive at 300?

Is it better to have multiple stages in an amplifier rather than just
using one op amp to get the job done? If so, why?

Any help would be greatly appreciated. Thanks

 

[Michael Black]

On Tue, 9 Sep 2008, jalbers@bsu.edu wrote:

I have a few questions about the design of the circuit in Figure 1
page 71 located at https://ilocker.bsu.edu/users/jalbers/WORLD_SHARED/WireTracer.PDF

The gain of the first op amp stage is 10 and the gain fo the second op
amp stage is 33 for a total net gain of 330. Why not just use one op
amp and set the gain to 330?

When designing a circuit like this, why use a low gain of 330, why not
a higher value? How did the author arrive at 300?

Is it better to have multiple stages in an amplifier rather than just
using one op amp to get the job done? If so, why?

Any help would be greatly appreciated. Thanks

One big reason for using multiple stages is because one stage won't

provide enough gain.

Don't be misled by the op-amps, think of all the stages inside those
op-amps that are needed to provide the gain. Before op-amps, and
even today, discrete components would require multiple stages to get
enough gain, precisely because a single stage couldn't.

Remember, an op-amp may be able to supply a lot of gain, but it
comes at the cost of bandwidth. The higher the gain, the lower the
bandwidth (though in more recent times the gain-bandwidth has improved
significantly from the days of the 1MHz 741, so you can get more gain
at a higher frequency than back then). So if the op-amp has a 1MHz
gain-bandwidth, divide that 1MHz by the gain and you'll see how
performance tops off. A gain of ten would allow flat response to
100KHz, a gain of 33 would allow a flat response to 30KHz, and a gain
of 330 would allow a flat response to 3KHz. That latter may really
impact things if you hope for good audio response.

Another reason for multiple stages is so the stages can be optimized
for their place in the chain. A first stage may be optimized for
low noise, especially if the input signal is very weak, like a radio
signal or the audio coming out of a phono cartridge. You want
to get the signal level up to a decent level without adding much noise
because otherwise that noise will be amplified along with the desired
signal.

Or, an output stage may be optimized for driving a speaker or something,
so it's more important than it can supply current gain rather than voltage
gain.

As for the amount of gain, surely whoever put it together decided
on the gain because that was what's needed. The circuit must
have some purpose, and thus the design specs specify the gain needed.

Michael

 

[Tim Wescott]

jalbers@bsu.edu wrote:

I have a few questions about the design of the circuit in Figure 1
page 71 located at https://ilocker.bsu.edu/users/jalbers/WORLD_SHARED/WireTracer.PDF

The gain of the first op amp stage is 10 and the gain fo the second op
amp stage is 33 for a total net gain of 330. Why not just use one op
amp and set the gain to 330?

As Bob Masta mentioned, the biggest reason is the opamp's gain bandwidth
product. That part has a GBW of about 1MHz, so a stage with a gain of
330 would have a bandwidth of just 3kHz, and it wouldn't be well
controlled, either.

There are additional difficulties, to, such as input offset (and its
drift), distortion products not getting beaten down because of
insufficient gain, and the oddball saturation effects that Bob mentioned.

All in all, it's often better to limit the amount of gain per stage to
keep things in better control.

When designing a circuit like this, why use a low gain of 330, why not
a higher value? How did the author arrive at 300?

High gain circuits oscillate. The higher the gain, the more readily
they oscillate, so the harder you have to work at decoupling the input
from the output.

The author probably found that a gain of 330 worked well enough, and
didn't want to mess with the ultra-careful circuit layout necessary to
keep the oscillations down in a higher gain circuit.

Or the author felt that he was asking enough of his amplifier, and
didn't want to add more stages.

Is it better to have multiple stages in an amplifier rather than just
using one op amp to get the job done? If so, why?

Any engineering question that starts with "is it better to ..." has the

answer "that depends". At best the answer will be "usually".

With circuits like this it's often better to limit the gain per stage,
and accept more stages as necessary. But there are always oddball
requirements that demand oddball solutions.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" gives you just what it says.
See details at http://www.wescottdesign.com/actfes/actfes.html

 

[Jamie]

jalbers@bsu.edu wrote:

I have a few questions about the design of the circuit in Figure 1
page 71 located at https://ilocker.bsu.edu/users/jalbers/WORLD_SHARED/WireTracer.PDF

The gain of the first op amp stage is 10 and the gain fo the second op
amp stage is 33 for a total net gain of 330. Why not just use one op
amp and set the gain to 330?

When designing a circuit like this, why use a low gain of 330, why not
a higher value? How did the author arrive at 300?

Is it better to have multiple stages in an amplifier rather than just
using one op amp to get the job done? If so, why?

Any help would be greatly appreciated. Thanks

Capacitance is designed in OP-AMPS in the feed back especially
, to help eliminate oscillation that would occurs.

In many references, this is called the Miller effect. some may
look at it a slightly different way.. Which is why, even though
comparator will work for basic OP-AMPS, they normally do not have
this kind of prevention circuit designed into them and can oscillate but
allows for a fast switching time to take place over using an
Op-amp as a comparator as long as the proper hysteresis is engineered
into it.

(Back to Op-amps).

This is when the Band Width comes into play, having a high gain
stage would involve using high value ratio&resistor networks and thus
the capacitance in the OP-AMP would have adverse affects on the gain at
higher frequencies.

Using much lower value resistors and still trying to get the gain
from a single stage then places the OP-amp into a stress state, if you
want to look at it that way.

Normally to get high gain and low noise, you select a practical
resistor network that will not allow the capacitance to place your band
width out of spec and keep the Op-amp in safe operating range.

Thus, multiple stages are created to smoothly bring it up.

You'll also find this kind of normanclature in basic transistor design
also.
http://webpages.charter.net/jamie_5"

 

[Eeyore]

Bob Masta wrote:

"jalbers@bsu.edu" <jalbers@bsu.edu> wrote:

I have a few questions about the design of the circuit in Figure 1
page 71 located at https://ilocker.bsu.edu/users/jalbers/WORLD_SHARED/WireTracer.PDF

The gain of the first op amp stage is 10 and the gain fo the second op
amp stage is 33 for a total net gain of 330. Why not just use one op
amp and set the gain to 330?

When designing a circuit like this, why use a low gain of 330, why not
a higher value? How did the author arrive at 300?

Is it better to have multiple stages in an amplifier rather than just
using one op amp to get the job done? If so, why?

Any help would be greatly appreciated. Thanks

The main reason for separate stages has to do with
an op-amp parameter called Gain Bandwidth Product.
The available gain of the chip falls with
frequency, so if you need a wide bandwidth you
must use a lower gain.

Yes, but for best noise performance the first stage should have the higher gain (making
certain assumptions about source impedance and op-amps used but almost invariably the
case).

Graham

 

[Eeyore]

Michael Black wrote:

On Tue, 9 Sep 2008, jalbers@bsu.edu wrote:

I have a few questions about the design of the circuit in Figure 1
page 71 located at https://ilocker.bsu.edu/users/jalbers/WORLD_SHARED/WireTracer.PDF

The gain of the first op amp stage is 10 and the gain fo the second op
amp stage is 33 for a total net gain of 330. Why not just use one op
amp and set the gain to 330?

When designing a circuit like this, why use a low gain of 330, why not
a higher value? How did the author arrive at 300?

Is it better to have multiple stages in an amplifier rather than just
using one op amp to get the job done? If so, why?

Any help would be greatly appreciated. Thanks

One big reason for using multiple stages is because one stage won't
provide enough gain.

Don't be misled by the op-amps, think of all the stages inside those
op-amps that are needed to provide the gain. Before op-amps, and
even today, discrete components would require multiple stages to get
enough gain, precisely because a single stage couldn't.

Remember, an op-amp may be able to supply a lot of gain, but it
comes at the cost of bandwidth. The higher the gain, the lower the
bandwidth (though in more recent times the gain-bandwidth has improved
significantly from the days of the 1MHz 741, so you can get more gain
at a higher frequency than back then). So if the op-amp has a 1MHz
gain-bandwidth, divide that 1MHz by the gain and you'll see how
performance tops off. A gain of ten would allow flat response to
100KHz, a gain of 33 would allow a flat response to 30KHz, and a gain
of 330 would allow a flat response to 3KHz. That latter may really
impact things if you hope for good audio response.

Not to mention that with essentially no feedback, the distortion is likely to be quite
high.

Graham

 

[Eeyore]

Bob Eld wrote:

jalbers@bsu.edu> wrote in message
I have a few questions about the design of the circuit in Figure 1
page 71 located at
https://ilocker.bsu.edu/users/jalbers/WORLD_SHARED/WireTracer.PDF

The gain of the first op amp stage is 10 and the gain fo the second op
amp stage is 33 for a total net gain of 330. Why not just use one op
amp and set the gain to 330?

When designing a circuit like this, why use a low gain of 330, why not
a higher value? How did the author arrive at 300?

Is it better to have multiple stages in an amplifier rather than just
using one op amp to get the job done? If so, why?

Any help would be greatly appreciated. Thanks

There is no reason this circuit would not operate perfectly well with a
single op-amp with a gain of 330 times. The discussion about gain bandwidth
product, while true, is a nonsequitur in this application. There is
sufficient gain-bandwidth to amplify mid range audio signals for your
purposes.

However, high gain implies a wide spread of resistance values. Finding a
suitable gain control pot might be an issue.

This thing has a low frequency cutoff of 1590Hz. What frequencies is it
supposed to amplify? It would be very poor at 50 or 60 Hz.

One poster mentioned DC characteristics such as drift and offset as a reason
for keeping the gain low per stage, but this circuit has a DC gain of unity
in both stages. A single stage version would be the same, meaning the DC
characteristics are immaterial.

Someone also mentioned noise. The more stages and the more TOTAL gain the
more noise. Spreading the gain amongst more stage adds noise it does not
reduce it. The total gain is the same.

These guys used two amplifiers because they felt like it not because of any
technical reason. They had a dual amplifier so that's what they used. A
better circuit would use one amplifier to establish the battery center
voltage with high resistance and the other amplifier as a single gain stage.
In this way the battery drain could be minimized to microamps in the quiet
state.

In this particular application, it looks like you're quite right.

Graham

 

[Eeyore]

Phil Allison wrote:

"Eeysore"
Bob Masta wrote:

The main reason for separate stages has to do with
an op-amp parameter called Gain Bandwidth Product.
The available gain of the chip falls with
frequency, so if you need a wide bandwidth you
must use a lower gain.

Yes, but for best noise performance the first stage should have the higher
gain (making certain assumptions about source impedance and op-amps used but
almost
invariably the case).

** Nonsense.

Once the first stage has a gain of a few times, then its ouput noise will
swamp the input noise of the following stage - no matter what gain it has.

** Nonsence.

What I say is true at ALL times. It's simple math.

Graham

 

[Charlie Siegrist]

On Tue, 9 Sep 2008 08:08:30 -0700 (PDT), "jalbers@bsu.edu"
<jalbers@bsu.edu> wrote:

I have a few questions about the design of the circuit in Figure 1
page 71 located at https://ilocker.bsu.edu/users/jalbers/WORLD_SHARED/WireTracer.PDF

The gain of the first op amp stage is 10 and the gain fo the second op
amp stage is 33 for a total net gain of 330. Why not just use one op
amp and set the gain to 330?

When designing a circuit like this, why use a low gain of 330, why not
a higher value? How did the author arrive at 300?

Is it better to have multiple stages in an amplifier rather than just
using one op amp to get the job done? If so, why?

Any help would be greatly appreciated. Thanks

As nobody else has yet mentioned a couple details about the circuit,
I'll take a bit of time to address them. First, the gain of the
second stage is 30.3, not 33. A minor point, but relevent. Second,
the gain of the first stage as coupled to the second stage is
variable, from zero to ten. These two points show that maybe you
hadn't looked at the circuit all that closely before posting your
questions.

As far as choosing the gain value, for this particular circuit, the
gain need only be high enough to efficiently and sufficiently drive
the load. That's the primary purpose of any amplifier. Why have a
gain of one thousand if one hundred is all you need? Choosing a
maximum gain value close to the reality of the circuit application
results in greater accuracy through the usable range.

This isn't a particularly demanding design, it's a simple application.
It may very well be that the design consideration included having a
dual package op-amp on hand, and/or the desire for a stable and simple
means of variable gain control. Sometimes deducing the reasoning
behind electronic design is like interpreting poetry - there isn't
always one simple and correct answer.

 

[extremesoundandlight@yaho]

On Sep 10, 5:51 am, "Phil Allison" <philalli...@tpg.com.au> wrote:

"Eeysore"







Phil Allison wrote:
"Eeysore"

Yes, but for best noise performance the first stage should have the
higher
gain (making certain assumptions about source impedance and op-amps
used but
almost invariably the case).

**  Nonsense.

Once the first stage has a gain of a few times, then its output noise
will
swamp the input noise of the following stage  -  no matter what gain it
has.

** Nonsence.

What I say is true at ALL times.

**  Nope  -   it is just more of your inexhaustible supply

 of   RIDICULOUS  &  FALSE   INFORMATION  !

.....    Phil- Hide quoted text -

- Show quoted text -

the more stages just amplafies your thermal distortion and passes it
to the next stage and thats not an effeicient way and you end up with
much more distortion over all

 

[Bill]

On Wed, 10 Sep 2008 06:19:32 -0700 (PDT),
"extremesoundandlight@yahoo.com" <extremesoundandlight@yahoo.com>
wrote:

the more stages just amplafies your thermal distortion and passes it
to the next stage and thats not an effeicient way and you end up with
much more distortion over all

Man, you are a crack. I hope you are not trying to live off this.

 

[Michael A. Terrell]

"extremesoundandlight@yahoo.com" wrote:

On Sep 10, 8:45 am, Bill <z...@z.z> wrote:
On Wed, 10 Sep 2008 06:19:32 -0700 (PDT),
"extremesoundandli...@yahoo.com" <extremesoundandli...@yahoo.com
wrote:

the more stages just amplafies your thermal distortion and passes it
to the next stage and thats not an effeicient way and you end up with
much more distortion over all

Man, you are a crack. I hope you are not trying to live off this.

Explain why you’re not passing distortion from your 4055 or any other
op amp and then amplifying that distortion in multiple stages.

I work for Yamaha Roland, fender and 5 other companies that make
amplifiers as a factory authorized repair tech I think I know
something about the first question that was purposed.

It doesn't take much training to repair amps. A lot of the tube
changers who used o screw up TV sets moved on to torturing band
equipment. In fact, we tested new techs by sticking them in our audio
department before letting them touch anything difficult. You need to
study circuit design if you really want to understand your problems.

I used to clean up a lot of crap that factory authorized shops had
butchered. None of them had any real test equipment, and had no idea if
it was just barely working or properly repaired. Most of my customers
had to depend on their audio gear 24/7 and had already learned to stay
away from the 'butchers'.


--
http://improve-usenet.org/index.html

aioe.org, Goggle Groups, and Web TV users must request to be white
listed, or I will not see your messages.

If you have broadband, your ISP may have a NNTP news server included in
your account: http://www.usenettools.net/ISP.htm


There are two kinds of people on this earth:
The crazy, and the insane.
The first sign of insanity is denying that you're crazy.

 

[Bill]

On Wed, 10 Sep 2008 14:34:07 -0700 (PDT),
"extremesoundandlight@yahoo.com" <extremesoundandlight@yahoo.com>
wrote:

On Sep 10, 8:45 am, Bill <z...@z.z> wrote:
On Wed, 10 Sep 2008 06:19:32 -0700 (PDT),
"extremesoundandli...@yahoo.com" <extremesoundandli...@yahoo.com
wrote:

the more stages just amplafies your thermal distortion and passes it
to the next stage and thats not an effeicient way and you end up with
much more distortion over all

Man, you are a crack. I hope you are not trying to live off this.

Explain why you’re not passing distortion from your 4055 or any other
op amp and then amplifying that distortion in multiple stages.

I work for Yamaha Roland, fender and 5 other companies that make
amplifiers as a factory authorized repair tech I think I know
something about the first question that was purposed.
Be constructive and tell me why I am wrong maybe a explanation other
thanbtbant
Man, you are a crack. I hope you are not trying to live off this.

Non-technical: my apologies. I don't like what I wrote. Again, sorry.

Technical: if each of the stages has a bandwidth that includes the
audio bandwidth and a response that is sufficiently flat, which is the
usual case, the stages do not amplify the input distortion, because
distortion is a ratio, and each stage amplifies both wanted and
unwanted components in more or less equal amounts. Of course it will
slightly favor one or the others, but it is hard to tell which ones,
and the "average" of the "gain factor" applied to the input distortion
is closer to 1 than to G (the gain of that stage). Another thing is
that that stage will also create and add new distortion products.

Best,

 

[George Herold]

On Sep 10, 2:21 am, "Phil Allison" <philalli...@tpg.com.au> wrote:

"Eeysore"

Bob Masta wrote:

The main reason for separate stages has to do with
an op-amp parameter called Gain Bandwidth Product.
The available gain of the chip falls with
frequency, so if you need a wide bandwidth you
must use a lower gain.

Yes, but for best noise performance the first stage should have the higher
gain (making
certain assumptions about source impedance and op-amps used but almost
invariably the
case).

**  Nonsense.

Once the first stage has a gain of a few times, then its ouput noise will
swamp the input noise of the following stage  -  no matter what gain it has.

.....    Phil

Once the first stage has a gain of a few times, then its ouput noise will
swamp the input noise of the following stage - no matter what gain it has.

Hmmm, Well this could depend on the resistor values choosen for the
second stage. A low noise front end coupled to a second stage with a
10k ohm resistor could make things worse.
George

 

[George Herold]

On Sep 10, 9:19 am, "extremesoundandli...@yahoo.com"
<extremesoundandli...@yahoo.com> wrote:

On Sep 10, 5:51 am, "Phil Allison" <philalli...@tpg.com.au> wrote:





"Eeysore"

Phil Allison wrote:
"Eeysore"

Yes, but for best noise performance the first stage should have the
higher
gain (making certain assumptions about source impedance and op-amps
used but
almost invariably the case).

**  Nonsense.

Once the first stage has a gain of a few times, then its output noise
will
swamp the input noise of the following stage  -  no matter what gain it
has.

** Nonsence.

What I say is true at ALL times.

**  Nope  -   it is just more of your inexhaustible supply

 of   RIDICULOUS  &  FALSE   INFORMATION  !

.....    Phil- Hide quoted text -

- Show quoted text -

the more stages just amplafies your thermal distortion and passes it
to the next stage and thats not an effeicient way and you end up with
much more distortion over all- Hide quoted text -

- Show quoted text -

amplafies your thermal distortion
What is thermal distortion?
George

 

[Eeyore]

George Herold wrote:

"Phil Allison" wrote:
"Eeysore"
Bob Masta wrote:

The main reason for separate stages has to do with
an op-amp parameter called Gain Bandwidth Product.
The available gain of the chip falls with
frequency, so if you need a wide bandwidth you
must use a lower gain.

Yes, but for best noise performance the first stage should have the higher
gain (making
certain assumptions about source impedance and op-amps used but almost
invariably the
case).

** Nonsense.

Once the first stage has a gain of a few times, then its ouput noise will
swamp the input noise of the following stage - no matter what gain it has.

Once the first stage has a gain of a few times, then its ouput noise will
swamp the input noise of the following stage - no matter what gain it has.

Hmmm, Well this could depend on the resistor values choosen for the
second stage. A low noise front end coupled to a second stage with a
10k ohm resistor could make things worse.

Well spotted. I'm working on an ultra low noise differential input right now (no -
none of the IC solutions are good enough) and resistive noise is truly a PITA !
Never underestimate it, so many people do then wonder why their circuit's 'hissy'.



Graham

 

[Eeyore]

George Herold wrote:

What is thermal distortion?

Certain devices when driving heavy loads can get quite warm (even ona cycle by
cycle basis). If the device is not designed to equalise this heat across specific
areas of the chip, it can affect the sensitive bits at the front end and lead to
thermally induced distortion.

I've even seen the effect in dummy load resistors with highish coeffs of temp. You
measure THD at the amp and it's fine (say 0.03%) and measure it at the load and
it's 0.1% ! The reason is the combination of the non-linearity of the load and the
finite resistance of the connecting wires.

Graham

 

[Bill Bowden]

On Sep 9, 6:38 pm, Eeyore <rabbitsfriendsandrelati...@hotmail.com>
wrote:

Bob Masta wrote:
"jalb...@bsu.edu" <jalb...@bsu.edu> wrote:

I have a few questions about the design of the circuit in Figure 1
page 71 located athttps://ilocker.bsu.edu/users/jalbers/WORLD_SHARED/WireTracer.PDF

The gain of the first op amp stage is 10 and the gain fo the second op
amp stage is 33 for a total net gain of 330. Why not just use one op
amp and set the gain to 330?

When designing a circuit like this, why use a low gain of 330, why not
a higher value? How did the author arrive at 300?

Is it better to have multiple stages in an amplifier rather than just
using one op amp to get the job done? If so, why?

Any help would be greatly appreciated. Thanks

The main reason for separate stages has to do with
an op-amp parameter called Gain Bandwidth Product.
The available gain of the chip falls with
frequency, so if you need a wide bandwidth you
must use a lower gain.

Yes, but for best noise performance the first stage should have the higher gain (making
certain assumptions about source impedance and op-amps used but almost invariably the
case).

Graham

Why would a higher gain first stage improve noise performance? Seems
like identical op-amps would produce the same total noise regardless
of how the gain was distributed? Wouldn't using a low noise op-amp in
the front end with somewhat higher gain be a better idea? The next
stage (a cheaper op-amp) wouldn't care much about the noise because
the signal level would be many times higher.

-Bill

 

[Eeyore]

Bill Bowden wrote:

Why would a higher gain first stage improve noise performance? Seems
like identical op-amps would produce the same total noise regardless
of how the gain was distributed?

Let me simplifiy this for you.

Let's say the input noise of your amp stage is 1 uV ?

You set your fist stage gain to ONE. It theoretically produces 1uV of output noise. This then
adds to the second stage noise of 1uV making 1.4uV equivalent input noise (addition of random
sources).

So you just made a noisier amplifier. Now do the calculations for other gain distributions.

Graham

 

[John Larkin]

On Tue, 9 Sep 2008 08:08:30 -0700 (PDT), "jalbers@bsu.edu"
<jalbers@bsu.edu> wrote:

I have a few questions about the design of the circuit in Figure 1
page 71 located at https://ilocker.bsu.edu/users/jalbers/WORLD_SHARED/WireTracer.PDF

The gain of the first op amp stage is 10 and the gain fo the second op
amp stage is 33 for a total net gain of 330. Why not just use one op
amp and set the gain to 330?

When designing a circuit like this, why use a low gain of 330, why not
a higher value? How did the author arrive at 300?

Is it better to have multiple stages in an amplifier rather than just
using one op amp to get the job done? If so, why?

Any help would be greatly appreciated. Thanks

Interesting aside: if you want to make an amp out of some number of
simple RC-coupled tube stages, the plate resistor determines the gain,
where the gain-bandwidth is pretty much fixed. If you want the maximum
gain-bandwidth from a cascade of stages, the optimum gain per stage is
e, the base of the natural logs.

e keeps popping up all over the place.

John