Operational amplifier highest gain restriction in practical case

[Lahari]

Is it possible to get gain of 1000 in a LM324 or OP07 practically? If not why? What gain can an OPAMP provide maximum? Why this restriction?

 

[audioguru]

The datasheet for an opamp shows its voltage gain vs frequency when it has no negative feedback. For a very slow LM324 with a supply that is 10V to 15V the typical gain is 400,000 forom DC to about 2Hz, then the gain drops for higher frequencies. At 1kHz the typical gain is about 250. At about 300kHz the gain is 1 like a piece of wire.

The typical voltage gain for an OP07 opamp is almost 1 million from DC to about 3Hz, 1000 at 1kHz and 1 at 1MHz.

The gain is reduced at higher frequencies by a "frequency compensation" capacitor. The gain is reduced so that the phase shift in the opamp does not cause oscillation when negative feedback is used. Oscillation will occur when the phase shift causes negative feedback to be positive feedback at a high frequency when there is gain at that frequency.

 

[Winterstone]

Is it possible to get gain of 1000 in a LM324 or OP07 practically? If not why? What gain can an OPAMP provide maximum? Why this restriction?

Hi Lahari, there is no theoretical restriction - but, certainly, a practical one.
An opamp could - theoretically - be operated without feedback at its maximum gain - provided that there is no offset !
But this never can be ensured in practice.
Therefore, negative feedback is necessary to bias the device in its linear region - more or less independent on offset effects.
In addition, negative feedback offers some other important advantages:
* less sensitivity to active tolerances (open-lop gain, temperature and offset effects)
* less sensitive to supply voltage variations
* less distortions (better THD)
* larger bandwidth
* smaller output resistance (better voltage source properties)
__________
* Disadvantage: Increased tendency to oscillations (less phase margin).
* Comment: The reduced gain due to negative feedback is no real "disadvantage", because this reduction is desired and leeds to less sensitivity of the final gain value to open-loop gain variations (uncertainty).

W.

 

[unclejed613]

so... if you look at the frequency vs gain curves AG posted, and draw a line across at 60db (gain of 1000), you will see that you would have a gain of 1000 up to 1khz, and after 1khz, the gain would drop off at 20db/decade. if you want a gain of 1000 up to 20khz, you need an op amp with 20Mhz unity gain bandwidth. this can be calculated very simply with Acl*Fbw=GBW or (1000*20khz)=20Mhz where Acl is the desired closed loop gain, and Fbw is the desired bandwidth of the amplifier and GBW is the minimum gain-bandwidth product of the op amp.

 

[ericgibbs]

The datasheet for an opamp shows its voltage gain vs frequency when it has no negative feedback. For a very slow LM324 with a supply that is 10V to 15V the typical gain is 400,000 forom DC to about 2Hz, then the gain drops for higher frequencies. At 1kHz the typical gain is about 250. At about 300kHz the gain is 1 like a piece of wire.

The typical voltage gain for an OP07 opamp is almost 1 million from DC to about 3Hz, 1000 at 1kHz and 1 at 1MHz.

hi agu,
I would agree that overall an OP07 has a better specification, especially at the very low frequency end, but I would suggest you are not comparing fairly the OP07 and LM324 gains with your Open Loop gains posted from the datasheets.

If you rationalise the two plots to the same X/Y axis scales, its clear the higher gain of the OP07 falls off rapidly above a few Hz. After that the slopes are very similar.

E.

 

[MrAl]

Hi Eric,


It is a well known fact that audioguru would rather use a rock with three wires glued to it than use one of those 'nasty' LM324/358 type op amps
He just doesnt like them. Today's op amps are much better though.

 

[ericgibbs]

Hi Eric,


It is a well known fact that audioguru would rather use a rock with three wires glued to it than use one of those 'nasty' LM324/358 type op amps
He just doesnt like them. Today's op amps are much better though.

hi Al,
I would agree the modern OPA's out perform the older versions, sadly in some parts of the world the older versions are the only ones which are taught and are readily available.

Perhaps I am being naive, but I believe what we post here should be technically factual and not be biased with personal opinions.

The LM324 has many features which make it a versatile OPA

To answer the OP's question, as the LM324 is a quad, 2 or 3 OPA's could be configured to give a gain 0f 1000.

E.

 

[audioguru]

The LM324 is too noisy (hiss) to be an audio amplifier.
The LM324 has crossover distortion which sounds awful.
The LM324 has a very slow slew rate that causes high output level frequencies above 2kHz to be triangle waves which produces severe distortion in audio.
Parts of the world that cannot obtain better opamps can use them in DC circuits or distorted narrow bandwidth AM radios.

The LM358 is a dual opamp version of the LM324 quad. There is (was?) a single version that I have never seen.

The LM324 and LM358 are the first low power opamps so they are distorted and have a narrow bandwidth.
MC33171 single, MC33172 dual and MC33174 quad opamps have exactly the same low power and output voltage levels but do not have crossover distortion and have a bandwidth up to 35kHz which is fine for audio.

 

[Winterstone]

To answer the OP's question, as the LM324 is a quad, 2 or 3 OPA's could be configured to give a gain 0f 1000.

E.

Hi Eric,
I am not quite sure if this (gain of 1000 with 2 or 3 opamp units) was the background of the OP's question :

Quote Lahari; What gain can an OPAMP provide maximum? Why this restriction?

Rather, I think he knows that each opamp is operated with negative feedback (reduced gain) - and his question is simply WHY ?
Therefore, my answer in post #3.

W.

 

[MrAl]

The LM324 is too noisy (hiss) to be an audio amplifier.
The LM324 has crossover distortion which sounds awful.
The LM324 has a very slow slew rate that causes high output level frequencies above 2kHz to be triangle waves which produces severe distortion in audio.
Parts of the world that cannot obtain better opamps can use them in DC circuits or distorted narrow bandwidth AM radios.

The LM358 is a dual opamp version of the LM324 quad. There is (was?) a single version that I have never seen.

The LM324 and LM358 are the first low power opamps so they are distorted and have a narrow bandwidth.
MC33171 single, MC33172 dual and MC33174 quad opamps have exactly the same low power and output voltage levels but do not have crossover distortion and have a bandwidth up to 35kHz which is fine for audio.

Hi again audioguru,

I think you are being too hard on the LM358 especially after i had explained how to deal with the crossover distortion in another thread which im sure you had read. What you are describing is an LM358 (or similar) in an application where the designer didnt know how to use it. That's when the 'problems' show up. True, it's not the best op amp in the world, but it's also not the worst.

You always mention that the LM358 cant be used in an audio application, but that's only if it isnt used right. I've used an LM358 in a headphone amplifier in the past and it worked pretty well even though it was driving headphones. It sounded good too with little noise or 'hiss'.
Again, i know it is not the best op amp in the world, but i think the better wording would be something like:
"The LM358 is not suitable for audio applications unless care is taken to use it within its natural limits".

 

[audioguru]

If you want to bother, you can bias an LM324 or LM358 opamp so that its output transistor works in class-A with low crossover distortion.
But why not use a modern opamp that has extremely low distortion instead?

Deaf people cannot hear the hiss, distortion and poor bandwidth of an LM324 or LM358 opamp.
Some people think an old awful-sounding AM radio sounds fine.

 

[MrAl]

Hi audioguru,

Well all i am saying is if it is designed right in the first place (and within the limits of the device) it works ok. Not the audiophiles choice perhaps, but done right it wont matter that much.

It's a matter of choice. If it works within reason, why not use it.

"Poor bandwidth" is still not exactly the real picture. It has 1MHz unity gain bandwidth which is a lot higher than some other more expensive chips made for low input offset. That means a gain of 10 at 100kHz as long as the output voltage is limited accordingly.

There are better transistors there are worse transistors, there are better LEDs there are worse LEDs, there are better op amps there are worse op amps. We dont always use the best possible part it's always a choice as long as it fits the application.