> This is verifiably untrue......Slew induced distortion (before the onset of slew limiting) only occurs with any significance, in single-pole Miller compensated amplifiers..
This is only gross TIM, soft TIM is generated as soon as the input signal of the given stage is large enough to stress any nonlinearity mechanism. Sorry mike, you are really mixed everything, apples and carburettors.
This is only gross TIM, soft TIM is generated as soon as the input signal of the given stage is large enough to stress any nonlinearity mechanism. Sorry mike, you are really mixed everything, apples and carburettors.
Ouroboros said:
Indeed the vast majority of commercial designs use single-pole minor loop compensation...this is because it is easy to implement, linearises the TIS...and is often entirely trouble-free with regard to instability.....
..However it increases the current swing demanded of the input stage with increasing frequency, provoking a propotionate increase in input stage distortion,.....
.... while simultaneously leaving the output stage exposed due to the fall in feedback, as the compensation capacitor rolls off loop transmission...
dimitri said:
No Dimitri....
'gross TIM'=Outright slew limiting
'soft TIM'=Distortion provoked in input stage by increased loading of minor loop capacitor before it goes into outright current clip...
Ouroboros said:
Would avoid 'current' feedback in audio applications, as the input stage operates in class-AB at the feedback node.....
It's bad enough that one has to endure class-AB operation in the output stage....
...extending it to the input stage in audio frequency applications is near perverse....
C'mon, I know what a carburettor is, I was being ironic.......
I'm not sure I get you here; can you explain more? I would have thought the input stage operates in pure class A; as I see it, at no stage would either diff pair device switch off completely until clip, or on those rare occasions when the feedback loop loses control.
Cheers
Hugh
I'm not sure I get you here; can you explain more? I would have thought the input stage operates in pure class A; as I see it, at no stage would either diff pair device switch off completely until clip, or on those rare occasions when the feedback loop loses control.
Cheers
Hugh
Christer said:
we in the states have those beauties until late 1980s: my 1987 chevy nova has one.
mikeks said:
mikeks, you aren't suggesting that current feedback applies to class-ab at the input stage, are you?
the jlh1969 for example has a current feedback input stage which operates in class a (the input stage).
You can run a current-feedback input stage in class-A, but it does mean that the dissipation in the current-mirror stages is rather high.
Normally the input stage will run in class AB, to ensure there's plenty of current available to charge and discharge the compensation cap between the output of the current mirrors and ground.
Normally the input stage will run in class AB, to ensure there's plenty of current available to charge and discharge the compensation cap between the output of the current mirrors and ground.
janneman said:
This is particularly true at higher freqs., and for typical feedback resistor values....
Note the particularly low values commonly used......
Avoid current feedback in audio apps. in general, unless this is part of a composite arrangement in which the current feedback circuitry acts as a compound output stage....particularly in 'small signal' applications..
Otherwise stick to voltage feedback, where ultra high slew rate if desired, can be readily obtained by other means.....as i have shown elsewhere.....
AKSA said:
There is no diff. stage in 'current feedback' arrangement....
In voltage feedback circuit, diff. stage normally used...and this goes into current clip when the tail source current is switched entirely from one member of the diff. stage to the other...
This will will not necessarily occur when the amp. goes into voltage clip at the TIS output....
...as 1st stage current clip is entirely a function of frequency, and the size of your minor loop comp. cap....
......assuming of course, that you've adopted single-pole minor-loop comp. in the first place........
John,
Sorry if I was unclear. I know both you and Gilbert are native
speakers of english. By different languages I meant the kind
of differences in terminology that often arises even between
people in neighbouring fields of technology. What I meant
was that perhaps there were people like Gilbert who knew
about the phenomenon, as he claims, but that he didn't
realize TIM as described by audio people had anything to
do with this. Or perhaps that is not the case? Anyway, as
Upupa Epops said, it is not really that important who was
first. It was unnecessary of me to mention it in the first
place since I didn't mean to start a debate on whether
Gilbert was right or not on this.
Sorry if I was unclear. I know both you and Gilbert are native
speakers of english. By different languages I meant the kind
of differences in terminology that often arises even between
people in neighbouring fields of technology. What I meant
was that perhaps there were people like Gilbert who knew
about the phenomenon, as he claims, but that he didn't
realize TIM as described by audio people had anything to
do with this. Or perhaps that is not the case? Anyway, as
Upupa Epops said, it is not really that important who was
first. It was unnecessary of me to mention it in the first
place since I didn't mean to start a debate on whether
Gilbert was right or not on this.
This is the situation, folks. There are many feedback related events that generate distortion. TIM (SID), PIM, load instability, and higher order conversion from lower order nonlinearities.
A sufficent slew rate, coupled with a linear input stage, eliminates just TIM.
Please remember, that 30 years ago, when we were wondering what was wrong with our audio designs, 0.5 V/us was considered by MANY COMPANIES as sufficent. This included tape recorders, studio boards, phono input stages, etc, etc. We had to FIGHT to get people to understand that simple slew rate was a worse case situation, AND that the distortion started building 5-10 times BELOW slew rate.
After 100's or even 1000's of measurements and giving papers at the AES, etc, people begrudgingly started to make faster IC's at low cost, as well as faster power amps, and the problem began to recede.
Then you get the 'academics' who want to take over, by criticizing previous work, and renaming the distortion mechanism.
And so it goes.
As far as I understand, two stage compensation is rarely used in most audio designs.
Once, 25 years ago, I developed a discrete 600 ohm driver for Sound Technology that could do VERY LOW DISTORTION at 100KHz and below. I used 2 stage compensation in order to get a 100KHz open loop bandwidth in order to have MAXIMUM feedback available at 100KHz. So far as I know, this circuit resides in the next generation ST distortion analyzer, after my own equipment.
Folks, let's give credit where credit is due, and ignore the small stuff that you don't completely know all about.
A sufficent slew rate, coupled with a linear input stage, eliminates just TIM.
Please remember, that 30 years ago, when we were wondering what was wrong with our audio designs, 0.5 V/us was considered by MANY COMPANIES as sufficent. This included tape recorders, studio boards, phono input stages, etc, etc. We had to FIGHT to get people to understand that simple slew rate was a worse case situation, AND that the distortion started building 5-10 times BELOW slew rate.
After 100's or even 1000's of measurements and giving papers at the AES, etc, people begrudgingly started to make faster IC's at low cost, as well as faster power amps, and the problem began to recede.
Then you get the 'academics' who want to take over, by criticizing previous work, and renaming the distortion mechanism.
And so it goes.
As far as I understand, two stage compensation is rarely used in most audio designs.
Once, 25 years ago, I developed a discrete 600 ohm driver for Sound Technology that could do VERY LOW DISTORTION at 100KHz and below. I used 2 stage compensation in order to get a 100KHz open loop bandwidth in order to have MAXIMUM feedback available at 100KHz. So far as I know, this circuit resides in the next generation ST distortion analyzer, after my own equipment.
Folks, let's give credit where credit is due, and ignore the small stuff that you don't completely know all about.
Charles Hansen said:
Neither can you....
Nobody...but nobody can hear the difference between a current feedback amp. , a voltage feedback amp, etc....
If all these units are calibrated to deliver exactly the same voltage swing across the same transducer.....below clip....and each possesses low THD....say below 100ppm across the audio band.......throughout the voltage swing.....NOBODY!
This is common sense.....scientifically proven common sense....
Why is it sooo hard to make simple factual statements in this field without having them disputed??
Subjectivism reminds me of the Vatican...Galileo....and the later's simple factual statement......the earth goes round the sun.....
This nearly got him killed by the 'infallible' pope....Same with this subjectivist nonsense.....Cannot make a statement of fact ...proven fact without somebody disputing it....because he can....
mikeks said:Why is it sooo hard to make simple factual statements in this field without having them disputed??
because if it were easy, a lot of people wouldn't be able to make a living off of selling otherwise average amps at extrodinary prices.
Why is it sooo hard to understand so simple of a fact? Why would you keep disputing those "infallible" claims of super human hearing?
millwood said:
No....'current feedback' obtain their high slew by running the 'small signal' stage in class AB....
This has nothing to do with the class of operation of the output stage.....
millwood said:
No.......it has a voltage feedback input stage.....
He merely chose a single common-emitter stage over the usual diff. stage....
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.