LH0002 Buffer Amp

[mcp6453[_2_]]

Has anyone here used an LH0002 buffer amp inside the feedback loop of op-amp to
create a high power op-amp for audio purposes? If so, how did it work for you?
The datasheet shows it driving 50 ohm loads at high frequencies, but I saw it
used somewhere in an audio circuit.

[geoff]

"mcp6453" > wrote in message
...
> Has anyone here used an LH0002 buffer amp inside the feedback loop of
> op-amp to
> create a high power op-amp for audio purposes? If so, how did it work for
> you?
> The datasheet shows it driving 50 ohm loads at high frequencies, but I saw
> it
> used somewhere in an audio circuit.

Why would you feel the need for a high-power opamp for audio ? And how
would it in a feedback loop acheive this ?

I assume you want more output than a 5523 can pump out ....

geoff

[geoff]

"geoff" > wrote in message
...
>
> "mcp6453" > wrote in message
> ...
>> Has anyone here used an LH0002 buffer amp inside the feedback loop of
>> op-amp to
>> create a high power op-amp for audio purposes? If so, how did it work for
>> you?
>> The datasheet shows it driving 50 ohm loads at high frequencies, but I
>> saw it
>> used somewhere in an audio circuit.
>
> Why would you feel the need for a high-power opamp for audio ? And how
> would it in a feedback loop acheive this ?
>
> I assume you want more output than a 5523 can pump out ....
>
> geoff


Ooops 5532 !

g

[PStamler]

On Friday, August 30, 2013 9:34:45 PM UTC-5, geoff wrote:

> Why would you feel the need for a high-power opamp for audio ? And how
>
> would it in a feedback loop acheive this ?

Maybe he wants to drive multiple outputs. Maybe he wants to drive a very low impedance. Maybe he wants to drive long lines with lots of capacitance. Who knows?

In any case, placing a buffer inside the loop of an opamp is a classic way to bolster its current-drive capacity. There are plenty of application notes about this. Often it's done with an integrated-circuit buffer; sometimes with a discrete buffer or a hybrid like the LH-0002. It requires a little extra attention to stability, but it works just fine.

Peace,
Paul

[William Sommerwerck]

"geoff" wrote in message
...

> How does a (say LH0002) buffer inside a feedback loop increase
> the output drive capability of the original amplifying device?

It does if the buffer has high drive capability. Right?

[geoff]

"William Sommerwerck" > wrote in message
...
> "geoff" wrote in message
> ...
>
>> How does a (say LH0002) buffer inside a feedback loop increase
>> the output drive capability of the original amplifying device?
>
> It does if the buffer has high drive capability. Right?

No, because the buffer, being in a feedback loop (of the original gain
stage), puts all that extra drive capability back to the original gain
device's input which is pointless. And the original device's output retains
it's original constraints.

Maybe he meant simply whop the buffer onto the output of the original
device, and the 'feedback' thing was a mis-speak.

geoff

[Don Pearce[_3_]]

On Sun, 1 Sep 2013 19:09:48 +1200, "geoff" >
wrote:

>
>"William Sommerwerck" > wrote in message
...
>> "geoff" wrote in message
>> ...
>>
>>> How does a (say LH0002) buffer inside a feedback loop increase
>>> the output drive capability of the original amplifying device?
>>
>> It does if the buffer has high drive capability. Right?
>
>No, because the buffer, being in a feedback loop (of the original gain
>stage), puts all that extra drive capability back to the original gain
>device's input which is pointless. And the original device's output retains
>it's original constraints.
>
>Maybe he meant simply whop the buffer onto the output of the original
>device, and the 'feedback' thing was a mis-speak.
>
>geoff
>

An op amp followed by a buffer, the whole enclosed within a single
feedback loop is how virtually every power amp is designed.

d

[William Sommerwerck]

"geoff" wrote in message
...
"William Sommerwerck" > wrote in message
...
> "geoff" wrote in message
> ...

>>> How does a (say LH0002) buffer inside a feedback loop increase
>>> the output drive capability of the original amplifying device?

>> It does if the buffer has high drive capability. Right?

> No, because the buffer, being in a feedback loop (of the original gain
> stage), puts all that extra drive capability back to the original gain
> device's input which is pointless. And the original device's output retains
> it's original constraints.

The buffer drives the load. The feedback loop consumes only a tiny bit of the
buffer's output. QED.

[Les Cargill[_4_]]

William Sommerwerck wrote:
> "geoff" wrote in message
> ...
>
>> How does a (say LH0002) buffer inside a feedback loop increase
>> the output drive capability of the original amplifying device?
>
> It does if the buffer has high drive capability. Right?

The original device remains unchanged by the addition of a buffer in
the feedback loop - it will still swing only so much voltage or
current.

--
Les Cargill

[geoff]

"William Sommerwerck" > wrote in message
...
> "geoff" wrote in message
> ...
> "William Sommerwerck" > wrote in message
> ...
>> "geoff" wrote in message
>> ...
>
>>>> How does a (say LH0002) buffer inside a feedback loop increase
>>>> the output drive capability of the original amplifying device?
>
>>> It does if the buffer has high drive capability. Right?
>
>> No, because the buffer, being in a feedback loop (of the original gain
>> stage), puts all that extra drive capability back to the original gain
>> device's input which is pointless. And the original device's output
>> retains
>> it's original constraints.
>
> The buffer drives the load. The feedback loop consumes only a tiny bit of
> the buffer's output. QED.

No . The OP had the buffer in the feedback loop - the output of the buffer
feeding back to the input of the original gain-stage. Poiuntless, but a
mis-speak I'm sure.

geoff

[William Sommerwerck]

"geoff" wrote in message
...

> The OP had the buffer in the feedback loop -- the output of the buffer
> feeding back to the input of the original gain-stage. Pointless, but a
> mis-speak I'm sure.

The buffer drives the load -- otherwise, what would the point of using it? --
and in common parlance, it's said to be "within" the loop, because the
feedback is taken /around/ the buffer. Right?

My electronic crossovers use exactly such an arrangement. The buffer allows
driving a low-impedance load that the op-amp could not handle by itself.

[Mike Rivers[_2_]]

[Not sure who wrote this] someone wrote:
>> The buffer drives the load. The feedback loop consumes only a tiny bit of
>> the buffer's output. QED.

On 9/1/2013 9:20 AM, geoff wrote:
> No . The OP had the buffer in the feedback loop - the output of the buffer
> feeding back to the input of the original gain-stage. Poiuntless, but a
> mis-speak I'm sure.

This would be so much less confusing if someone just drew a diagram and
posted a link to it. Then we'd know where the buffer was connected and
where the load was connected, and each person's interpretation of what
was asked.

When I first read the question, it sounded to me like the original
poster wanted to put a buffer in the feedback loop of the output stage.
That didn't make any sense. It makes more sense to put a higher current
driver ("buffer") after the output stage, and, if it makes any sense,
extend the feedback takeoff point to the output of the buffer rather
than where it was without the buffer.

Does anyone know what we're reall talking about here?



--
For a good time, call http://mikeriversaudio.wordpress.com

[William Sommerwerck]

"Mike Rivers" wrote in message ...

> Does anyone know what we're really talking about here?

I do, and I've explained it, even giving an example for a product I own.

[geoff]

"William Sommerwerck" > wrote in message
...
> "Mike Rivers" wrote in message ...
>> Does anyone know what we're really talking about here?
>
> I do, and I've explained it, even giving an example for a product I own.

No Willie. Re-read the original post and you will see that the normal
arrangement, obvious to you and me, is not what the OP was suggesting.

geoff

[gareth magennis]

"geoff" wrote in message
...


"William Sommerwerck" > wrote in message
...
> "Mike Rivers" wrote in message ...
>> Does anyone know what we're really talking about here?
>
> I do, and I've explained it, even giving an example for a product I own.

No Willie. Re-read the original post and you will see that the normal
arrangement, obvious to you and me, is not what the OP was suggesting.

geoff





I believe the OP may be describing this app sheet, specifically pages 4 and
18.

www.ti.com/lit/an/snoa725a/snoa725a.pdf




Gareth.

[PStamler]

Geoff, I'm afraid you're wrong. When a buffer is inserted into the feedback loop as the OP suggests, it's driven by the opamp's original output. If the buffer is capable of putting out 100mA of current, then the composite amp will put out up to 100mA of current, though its voltage output will be limited by the opamp's constraints. Say the buffer has a current gain of 100 (just to make the math easy). If the composite circuit puts out 100mA, then the opamp has to put out only 1mA into the buffer, which means ir's got an easy load to drive. But the full feedback is applied around the composite circuit, meaning the voltage gain will be the same as without the buffer. Assuming stability issues are managed properly, it's a very high quality way of getting an output circuit.

In essence, you're adding an additional current gain stage to the opamp's own circuitry. As Bill points out, this is essentially what most solid-state power amps do.

Peace,
Paul

[William Sommerwerck]

"geoff" wrote in message
...

"William Sommerwerck" > wrote in message
...
> "Mike Rivers" wrote in message ...
>> Does anyone know what we're really talking about here?
>
> I do, and I've explained it, even giving an example for a product I own.

No Willie. Re-read the original post and you will see that the normal
arrangement, obvious to you and me, is not what the OP was suggesting.

Well, to me, "inside the feedback loop" has a specific meaning that is what
you and I would agree on.

[William Sommerwerck]

"Gareth Magennis" wrote in message ...

> I believe the OP may be describing this app sheet,
> specifically pages 4 and 18.

> www.ti.com/lit/an/snoa725a/snoa725a.pdf

The schematics on pages 14 and 18 exactly what we're talking about -- the
buffer is inside the feedback loop.

Of course, the buffer doesn't /have/ to go inside the feedback loop -- but
it's usually put there.

[geoff]

"William Sommerwerck" > wrote in message
...
> "Gareth Magennis" wrote in message ...
>
>> I believe the OP may be describing this app sheet,
>> specifically pages 4 and 18.
>
>> www.ti.com/lit/an/snoa725a/snoa725a.pdf
>
> The schematics on pages 14 and 18 exactly what we're talking about -- the
> buffer is inside the feedback loop.
>
> Of course, the buffer doesn't /have/ to go inside the feedback loop -- but
> it's usually put there.

In which case the terminology could have been more complete. The scenario I
immediated envisaged from the description was the buffer in the feedback
path, which is of course silly.

geoff.

[Trevor]

"geoff" > wrote in message
...
>>>>> How does a (say LH0002) buffer inside a feedback loop increase
>>>>> the output drive capability of the original amplifying device?
>>
>>>> It does if the buffer has high drive capability. Right?
>>
>>> No, because the buffer, being in a feedback loop (of the original gain
>>> stage), puts all that extra drive capability back to the original gain
>>> device's input which is pointless. And the original device's output
>>> retains
>>> it's original constraints.
>>
>> The buffer drives the load. The feedback loop consumes only a tiny bit of
>> the buffer's output. QED.
>
> No . The OP had the buffer in the feedback loop - the output of the buffer
> feeding back to the input of the original gain-stage.

Where did you see that? I saw no circuit diagram.

>Poiuntless, but a mis-speak I'm sure.

Your "mis-speak" could also be a mis-interpretation.

Trevor.

[geoff]

"Trevor" > wrote in message
...
>
> "geoff" > wrote in message
> ...
>>>>>> How does a (say LH0002) buffer inside a feedback loop increase
>>>>>> the output drive capability of the original amplifying device?
>>>
>>>>> It does if the buffer has high drive capability. Right?
>>>
>>>> No, because the buffer, being in a feedback loop (of the original gain
>>>> stage), puts all that extra drive capability back to the original gain
>>>> device's input which is pointless. And the original device's output
>>>> retains
>>>> it's original constraints.
>>>
>>> The buffer drives the load. The feedback loop consumes only a tiny bit
>>> of the buffer's output. QED.
>>
>> No . The OP had the buffer in the feedback loop - the output of the
>> buffer feeding back to the input of the original gain-stage.
>
> Where did you see that? I saw no circuit diagram.
>
>>Poiuntless, but a mis-speak I'm sure.
>
> Your "mis-speak" could also be a mis-interpretation.


Granted.

I'd have said put a LH0002 buffer on the output, and take the feedback
source from it's output. But that's just me.

geoff

[Mike Rivers[_2_]]

On 9/1/2013 5:29 PM, Gareth Magennis wrote:

> I believe the OP may be describing this app sheet, specifically pages 4
> and 18.

> www.ti.com/lit/an/snoa725a/snoa725a.pdf

I think that the answer here are describing Figure 32 in that
application note. I'm not sure if that's the hookup the original poster
was talking about, or if he had a different idea of what "in the
feedback loop" means. My first impression of the question was that he
was asking about putting a buffer between the output of the op amp and
its input. I couldn't see how that would do anything good.



--
For a good time, call http://mikeriversaudio.wordpress.com

[Mike Rivers[_2_]]

On 9/1/2013 6:36 PM, PStamler wrote:

> Say
> the buffer has a current gain of 100 (just to make the math easy). If
> the composite circuit puts out 100mA, then the opamp has to put out
> only 1mA into the buffer, which means ir's got an easy load to drive.
> But the full feedback is applied around the composite circuit,
> meaning the voltage gain will be the same as without the buffer.

To me, "buffer" implies unity voltage gain and as much current gain as
it's designed to handle and the power supply can support. That may not
be the case any more. Languages are subject to change without notice.

--
For a good time, call http://mikeriversaudio.wordpress.com

[William Sommerwerck]

"geoff" wrote in message
...

> I'd have said put a LH0002 buffer on the output, and take
> the feedback source from it's output. But that's just me.

That's a correct description. It's just that "inside the feedback loop" is
commonly taken to mean that.

[William Sommerwerck]

"Neil Gould" wrote in message ...

> I was also unclear about the OP's meaning. So, I took a look at the
> schematic examples for the LH002. Those who accept the "common parlance"
> apparently read the question in a reasonable way, but offered no insights as
> to whether it was a good idea to create such a circuit. So, after all the
> chatter I'm still pondering the benefits of such a layout vs. having the
> buffer independent of the source op amp's feedback circuit and set to
> unity gain, which is how I would do it.

The general theory is that any time you put an amplifying device within a
feedback loop, you get lower distortion, and reduced sample-to-sample
variation.

Here's what the spec sheet says about putting the buffer in the feedback
loop...

Closed-Loop Feedback Operation
-------------------------------------------
Any of the buffer amplifiers may be used inside an op amp
feedback loop. When this is done, the additional phase lag
introduced by the buffer must be included in loop stability
consideration. With most op amps, the bandwidth of these
buffers is so great that the op amp totally controls the loop
stability. However, when using very wide-band op amps such
as the LH0024, LH0032, and LH0062, the small additional
phase lag of the buffers should be taken into consideration.
Figure 26 presents the Bode plots of gain and phase for the
three buffers. The phase margin and open loop frequency response
is altered by the additional pole(s) contributed by the
buffer. The buffer phase shift is algebraically summed with the
op amp phase shift, and may cause a stable op amp loop to
become marginally stable depending upon the relative positions
of the op amp and buffer poles. In general, the buffer
bandwidth should significantly exceed that of the op amp, so
that the loop performance will be determined solely by the op
amp.

This, by the way, is an example of overwriting -- explaining everything twice.

[Neil Gould]

Mike Rivers wrote:
> On 9/1/2013 5:29 PM, Gareth Magennis wrote:
>
>> I believe the OP may be describing this app sheet, specifically
>> pages 4 and 18.
>
>> www.ti.com/lit/an/snoa725a/snoa725a.pdf
>
> I think that the answer here are describing Figure 32 in that
> application note. I'm not sure if that's the hookup the original
> poster was talking about, or if he had a different idea of what "in
> the feedback loop" means. My first impression of the question was
> that he was asking about putting a buffer between the output of the
> op amp and its input. I couldn't see how that would do anything good.
>
I was also unclear about the OP's meaning. So, I took a look at the
schematic examples for the LH002. Those who accept the "common parlance"
apparently read the question in a reasonable way, but offered no insights as
to whether it was a good idea to create such a circuit. So, after all the
chatter I'm still pondering the benefits of such a layout vs. having the
buffer independent of the source op amp's feedback circuit and set to unity
gain, which is how I would do it.
--
best regards,

Neil

[Mike Rivers[_2_]]

On 9/2/2013 10:18 AM, William Sommerwerck wrote:

> That's a correct description. It's just that "inside the feedback loop"
> is commonly taken to mean that.

But since nearly every op-amp configured as an amplifier has a feedback
path, it's not very specific unless this is designer lingo where
"feedback loop" has a specific connotation of what it includes. We must
have blown at least 1,000 words on this discussion, where a picture made
it very clear.


--
For a good time, call http://mikeriversaudio.wordpress.com

[Don Pearce[_3_]]

On Mon, 02 Sep 2013 11:30:04 -0400, Mike Rivers >
wrote:

>On 9/2/2013 10:18 AM, William Sommerwerck wrote:
>
>> That's a correct description. It's just that "inside the feedback loop"
>> is commonly taken to mean that.
>
>But since nearly every op-amp configured as an amplifier has a feedback
>path, it's not very specific unless this is designer lingo where
>"feedback loop" has a specific connotation of what it includes. We must
>have blown at least 1,000 words on this discussion, where a picture made
>it very clear.

I think there are two confused terminology here.l

Inside the feedback loop I take to mean that the main output and
feedback are both taken from the output of the buffer - this would be
the normal configuration.

In the feedback path means just that - it is in the path that returns
the output to the comparator, but is not part of the output driving
circuitry. This would be a stupid and pointless configuration
guaranteed both to increase distortion and minimize drive power.

d

[None]

"William Sommerwerck" > wrote in message
...
> snip>
> This, by the way, is an example of overwriting -- explaining
> everything twice.

Apparently, the writer is writing for the benefit of the reader. For
someone who writes only to impress himself, this must seem
so strange!

[William Sommerwerck]

"None" wrote in message
...
"William Sommerwerck" > wrote in message
...

>> This, by the way, is an example of overwriting -- explaining everything
>> twice.

> Apparently, the writer is writing for the benefit of the reader. For
> someone who writes only to impress himself, this must seem so
> strange!

For someone who has probably never written anything of import, you display
your ignorance quite crudely.

Explaining something twice is a mark of poor organization. Doing it once,
correctly, takes fewer words and is easier to understand.

I used to edit "Electronic Design" articles. This careless redundancy was a
common problem.

[gareth magennis]

"William Sommerwerck" wrote in message ...


Snip...

Explaining something twice is a mark of poor organization. Doing it once,
correctly, takes fewer words and is easier to understand.





Ah, I note the use of the word "correctly".

I'm not sure this has any meaning whatsoever.



Gareth.

[Scott Dorsey]

mcp6453 > wrote:
>Has anyone here used an LH0002 buffer amp inside the feedback loop of op-amp to
>create a high power op-amp for audio purposes? If so, how did it work for you?
>The datasheet shows it driving 50 ohm loads at high frequencies, but I saw it
>used somewhere in an audio circuit.

It works fine, but why not just use two power transistors? That way you can
make the compromise between drive capacity and current output yourself.
--scott

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

[Scott Dorsey]

geoff > wrote:
>"William Sommerwerck" > wrote in message
...
>> "geoff" wrote in message
>> ...
>>
>>> How does a (say LH0002) buffer inside a feedback loop increase
>>> the output drive capability of the original amplifying device?
>>
>> It does if the buffer has high drive capability. Right?
>
>No, because the buffer, being in a feedback loop (of the original gain
>stage), puts all that extra drive capability back to the original gain
>device's input which is pointless. And the original device's output retains
>it's original constraints.

No. You're thinking of voltage drive.... because it's in the feedback loop,
the voltage drive is still limited by the supply rails on the op-amp. But
the current is not... you could control thousands of amps with one op amp
and a big enough buffer.
--scott

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

[William Sommerwerck]

I have rarely seen so much pointless discussion over something so trivial.

[Scott Dorsey]

In article >, Mike Rivers > wrote:
>When I first read the question, it sounded to me like the original
>poster wanted to put a buffer in the feedback loop of the output stage.
>That didn't make any sense. It makes more sense to put a higher current
>driver ("buffer") after the output stage, and, if it makes any sense,
>extend the feedback takeoff point to the output of the buffer rather
>than where it was without the buffer.
>
>Does anyone know what we're reall talking about here?

Yes. He wants more current drive from an op-amp. He wants the standard
output stage that hundreds of manufacturers have used: the 5532 with the
buffer.

The buffer is inside the feedback loop so many forms of distortion in the
buffer are compensated for. The buffer is needed because the 5532 will not
swing rail-to-rail into a 600 ohm load, let alone two in parallel.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[hank alrich]

William Sommerwerck > wrote:

> I have rarely seen so much pointless discussion over something so trivial.

Language is a tool for discussion, and for understanding or
misunderstanding. Language is flexible, and many people speak
differently, even speak different languages.

If the only stupid question is the question not asked, then each
consideration bears equal weight, in the big picture.

Understanding is shared with patience. Patience may be derived
personally by realizing and accepting that one knoweth not all of it,
and never shall.

--
shut up and play your guitar * HankAlrich.Com
HankandShaidriMusic.Com
YouTube.Com/WalkinayMusic

[William Sommerwerck]

"hank alrich" wrote in message
...
William Sommerwerck > wrote:

>> I have rarely seen so much pointless discussion over something so trivial.

> Language is a tool for discussion, and for understanding or
> misunderstanding. Language is flexible, and many people speak
> differently, even speak different languages.

> If the only stupid question is the question not asked, then each
> consideration bears equal weight, in the big picture.

> Understanding is shared with patience. Patience may be derived
> personally by realizing and accepting that one knoweth not all of it,
> and never shall.

I won't gainsay your remarks. But in this particular case, the argument was
over the usage of the word "inside", which has an accepted meaning in the
context of feedback loops. As I've grown older, I've gradually (but not
completely) grown out of arguing with accepted usage.