What is wrong with op-amps?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Disabled Account
Joined 2006
Why, is not the big problem,

John, with respect, "Why" is actually very important indeed. It leads to understanding and hence to true knowledge that can be applied for positive future results.

In post #1 the primary question was “What”, but the underlying question is “Why”. So far this thread has produced a wonderfully diverse set of replies!

We have many diverse opinions, with three seemingly prevalent, as follows:

1.
Some commercial companies take decisions to do discrete design based on marketing reasons (i.e. to look ‘expert’), and hence their ‘op-amp vs discrete design’ decisions are not always based on technical superiority.

2.
Op-amps work excellently, and if not then it is the application / support circuit details that are wrong.

3.
Op-amps for simple gain stages can be bettered by simple open-loop discrete design.


I do not know the absolute fact, but I think there may be some truth in each of these three opinions.
 
Hi Guys,

I love opams and use them extensively. What I want to say here is purely philosophical and has nothing to do with engineering.

Here it goes:

There is something odd conceptually to strive for HUGE open loop gain around six orders of magnitude for getting 1-10x amplification at the end.

Nature does not operate this way. I am not aware a of a single natural feedback system with such a high OLG (please correct me if I am wrong).
There are plenty of amplification systems in biology and most have a feedback built in. However, they all consist of cascades with moderate amplification.

You might already guessed where I am going with this. For me more elegant solution is few HIGH quality cascades with moderate OLGs and local feedbacks instead of opamp.
 
There are plenty of amplification systems in biology and most have a feedback built in. However, they all consist of cascades with moderate amplification.

I'd be interested to consider some examples you might offer here. As far as I'm aware there's nowhere in nature (as opposed to man-made artifacts) where representation is used. This is a step back from where you started but in philosophy its important to check every assumption.

In audio we use a signal (usually a voltage, sometimes a current) to refer to or represent sound pressure. An analog if you will. That's inherently unnatural - nature doesn't operate on representations - the ear does not produce information which represents the sound pressure at its tympanic membrane. The eye does not produce information which represents photons impinging on its retina. And so on - the whole representation thing is an artifact of humans.

So wherever your argument against high open loop gains is going, you're going to have to be a bit more thorough in your observations of the natural world if you want to do sound philosophy based on it.
 
AX tech editor
Joined 2002
Paid Member
Maybe we differ in our interpretation of what constitutes a "discrete" amplifier. I'm pretty sure (according to what I consider discrete) there are discrete microwave amplifiers that have GBWs that are massively larger than any "standard IC package" opamp. The package is, for example, the reason that you will never find a Ku-band (12GHz) amplifier with more than 40dB gain in a standard IC package. I consider cascading a couple of such amplifiers on a PCB a "discrete" solution.

Well, OK, then your definition of discrete is different from the rest of us audio diy-ers.
In audio, discrete solutions never get to the GBW that IC's reach.

jd
 
AX tech editor
Joined 2002
Paid Member
[snip]When equipment comes available with no distortion (neither measurable nor audible) at anything near affordable prices, we'll all stand in line for it like Apple fans buying the latest iPhone. [snip]

There's tons of equipment available that has inaudible distortion. The problem is that nobody accepts anything close to a controlled, objective test that shows the inaudibility.
It goes further than that. Nobody is really interested in inaudible distortion, although everybody says so. Nobody is interested in a test that shows that a good tube amp and a ss amp cannot be distinguished on sound alone as long as they not clip (which has been done). It would be the end of our hobby!

jd
 
AX tech editor
Joined 2002
Paid Member
I'd be interested to consider some examples you might offer here. As far as I'm aware there's nowhere in nature (as opposed to man-made artifacts) where representation is used. [snip].

There are many systems that do this. Example: riding a bike is the application of a stable negative feedback system using several actuators ('amps') and feedback paths (visual, tactile, orientation wrt the outside world).

Another example: spread your arm sideways. Look at a point straight ahead of you. Then try to move your arm as fast as possible forward so that it points to the point you are looking at. Really try to do it as fast as possible. You will see an overshoot. Why? This is a feedback system: the feedback from your eyes which continuously report the arm position tells your motor system when to stop moving, when the end point is reached (the difference between actual and commanded position becomes zero).
Trying to do it as fast as possible your brain motor system cranks up the gain so high that the edge of stability is reached (conditional stability) and overshoot results. An exquisite feedback system.

Nature is so full of it, we don't even notice it. If feedback wouldn't work, your breathing would stop within a minute due to lack of feedback.
We engineers are just scratching the rich world of feedback systems nature has perfected over the eons. Feedback is very, very natural.

jd
 
Last edited:
Nobody is interested in a test that shows that a good tube amp and a ss amp cannot be distinguished on sound alone as long as they not clip (which has been done).

That's because such a test would be a scientific test, boring boring boring. Don't be such a scientist Jan:p

Those who hear differences (I'm one of them - heard a difference between two otherwise identical ss amps that had slightly different grounding) aren't interested because we trust our ears. Such a person (who hears a difference) isn't going to give up preferring and buying what they like in the possibility that the difference disappears under blind listening conditions. That's because blind conditions are only for scientists doing tests. In real world conditions, we hear differences and that's what matters. What could be the purpose of submitting to a blind test when its so unlike the practical reality of listening to music for enjoyment?
 
There are many systems that do this. Example: riding a bike is the application of a stable negative feedback system using several actuators ('amps') and feedback paths (visual, tactile, orientation wrt the outside world).

Where's the representation? I'm not denying feedback systems but just show me where some 'quantity' is represented by some information in a biological system.

Another example: spread your arm sideways. Look at a point straight ahead of you. Then try to move your arm as fast as possible forward so that it points to the point you are looking at. Really try to do it as fast as possible. You will see an overshoot. Why? This is a feedback system: the feedback from your eyes which continuously report the arm position tells your motor system when to stop moving, when the end point is reached (the difference between actual and commanded position becomes zero).
Trying to do it as fast as possible your brain motor system cranks up the gain so high that the edge of stability is reached (conditional stability) and overshoot results. An exquisite feedback system.

Again, I'm by no means denying feedback systems. Yet I see no representation in what you say (which I accept). I can only assume you're adding the representation yourself.

Nature is so full of it, we don't even notice it. If feedback wouldn't work, your breathing would stop within a minute due to lack of feedback.
We engineers are just scratching the rich world of feedback systems nature has perfected over the eons. Feedback is very, very natural.

No disagreement there Jan. But where's the representation?

<edit> I've pondered more on your middle paragraph, Have you ever read any Maurice Merleau-Ponty ? I recommend that you study some of the writings of Hubert Dreyfus if this stuff about "commanded position" is something you'd like to understand better. Dreyfus explains how these systems operate without the notion of 'commanded position'.
 
Last edited:
AX tech editor
Joined 2002
Paid Member
Don't be such a scientist Jan:p

:rolleyes:

[snip] (I'm one of them - heard a difference between two otherwise identical ss amps that had slightly different grounding) [snip]

Which, of course, would be easily measureable and confirmed. Do you know we can measure the difference in feedback pick-off point to within a few mm on the amp output trace? You would not hear that though, not by a long shot.

We all know audible differences exists between equipment. There's so much equipment designed on 'faith' that the design flaws are easily heard and measured.
But audible differences between competently designed equipment? Don't think so.

jd
 
Which, of course, would be easily measureable and confirmed.

Would you like to propose how I might be able to measure it? Bearing in mind I don't have an AP but do have a nice scope, DMM and various other fairly cheap bits of kit. I agree in principle its going to be measureable, because its definitely not magic:D

Do you know we can measure the difference in feedback pick-off point to within a few mm on the amp output trace? You would not hear that though, not by a long shot.

Have done so myself when I had access to an AP, yes. I wouldn't be so sure its not audible though without a testable hypothesis for why.

We all know audible differences exists between equipment. There's so much equipment designed on 'faith' that the design flaws are easily heard and measured.
But audible differences between competently designed equipment? Don't think so.

Since you're the scientist here I'm going to ask for a watertight definition of 'competently designed' before I entertain this further. That's because I've heard the 'competently designed' argument so many times, I consider it to be a weasel phrase:D

Just to start the ball rolling, is any piece of amplification equipment which doesn't use star earthing consistently 'competently designed'? Just answer 'yes' or 'no' please, no prevarication:p
 
Last edited:
AX tech editor
Joined 2002
Paid Member
Would you like to propose how I might be able to measure it? Bearing in mind I don't have an AP but do have a nice scope, DMM and various other fairly cheap bits of kit. I agree in principle its going to be measureable, because its definitely not magic:D[snip]

Yes it's measureable. I accept that not everybody here has the experience or equipment to do it, but that's not important to the discussion. The important thing is that audible differences MUST be measurable because they rely on electrical differences of the output signal. The probelem is what and how to measure.

[snip]Since you're the scientist here I'm going to ask for a watertight definition of 'competently designed' before I entertain this further. That's because I've heard the 'competently designed' argument so many times, I consider it to be a weasel phrase:D[snip]

My personal definition of a competently designed amp is an amp that does not sound different from another cpompetently designed amp in a statistically relevant controlled test.

[snip]Just to start the ball rolling, is any piece of amplification equipment which doesn't use star earthing consistently 'competently designed'? Just answer 'yes' or 'no' please, no prevarication:p

Are you looking for a reasonable answer or are you trying to trick me into a corner by fabricating this question and then give two answers that you KNOW are not the only ones? Tsk tsk!

Try again.

jd
 
Administrator
Joined 2007
Paid Member
I would call a Bosendorfer, a Steinway and a Fazioli "competently" designed instruments.
No note on each produces anything like the same sound as the other... and it's personal preference that decides which you would choose.

Similarly with amplifiers... the text book perfect amps do all have a "sameness" about them on audition I find. Which isn't surprising. Do you like that sound ? and many do. If so look no further.
The big question is when you are faced with "perfection" and secure in the knowledge that your amp whatever it may be is "perfect", and then some little upstart of an amp comes along, cobbled together, that absolutely wipes the floor with your "reference" on audition.

The question now is what do you do. Do you accept that you actually prefer certain types of distortion, and that really what matters is the "enjoyment factor" of the amp.

Or do you stick to the idea of wrinkling out the nth degree of non linearity, the minutest distortion mechanisms etc, which still give an "unsatisfying" sound to many.

The perfect piano can be designed on a computer, built by a robot, and then listened to by us... and we would probably prefer the old "wooden contraption" full of springs, felt, glue etc with all it's overtones, it's unique blend of harmonics and so on.

Just my thoughts :)
 
AX tech editor
Joined 2002
Paid Member
I would call a Bosendorfer, a Steinway and a Fazioli "competently" designed instruments.
No note on each produces anything like the same sound as the other... and it's personal preference that decides which you would choose.

Similarly with amplifiers... the text book perfect amps do all have a "sameness" about them on audition I find. Which isn't surprising. Do you like that sound ? and many do. If so look no further.[snip]

It depends on your definition. My definition of competently designed is that they don't add to or substract from the input signal. Therefor, two competently designed amps WILL sound the same. Therefore, if they sound differently, one of them (or both) are not competently designed by this definition.

If you accept audible differences then maybe you are more looking for a 'pleasant' sound than for accurate reproduction? Which is of course your choice, and nobody should have any problem with it.

jd
 
Yes it's measureable. I accept that not everybody here has the experience or equipment to do it, but that's not important to the discussion.

Obviously not disagreeing that its definitely measurable but I think methods are important to the discussion. Until we have a reliable measurement which can pick up differences like this, and people can go out and buy a device that does it, the issue won't be settled. We're in fairly dire need of standardized measurements which go beyond THD+N and intermod.

The important thing is that audible differences MUST be measurable because they rely on electrical differences of the output signal. The probelem is what and how to measure.

Its important that its measurable yes, but no less important is that people must be able to measure it in practice. Which means having measurement kit which does it available to all.


My personal definition of a competently designed amp is an amp that does not sound different from another cpompetently designed amp in a statistically relevant controlled test.

Well I humbly submit that your definition is circular;). An amp that passes the test is competently designed, only amps which are competently designed pass the test.


Are you looking for a reasonable answer or are you trying to trick me into a corner by fabricating this question and then give two answers that you KNOW are not the only ones? Tsk tsk!

No, no kicking, I'm not a footballer or a kungfu master. Oh, sorry you said 'trick' my eyesight is worse than I thought! What other answers do you want? If you'd like to say that 'competently designed' is a matter of degree (rather than all or nothing) then I'd agree with you, but your test didn't afford degrees of passing. So looks like we're a bit stuck:eek:
 
Administrator
Joined 2007
Paid Member
Hi Jan,
I think I do prefer a more "pleasant" sound, and I like an amp to sound musically satisfying on the majority of music fed to it, not just on a few choice recordings... as you say that's a personal thing, my choice and I accept others may not see it that way.

Your definition is OK... but do you like the result on audition ;)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.