Quote
I have no idea, tried it as IV-converter after DAC some years ago when the part was just released.
Clearly, you are a :bs: engineer
did you check it out with a scope?
was your layout ok
was the supply ok
Was it oscillating
Were you driving a capacitive load
etc
Why dis a product in this excessively subjective manner without qualifying in some detail your statement? This is straight mis-information and pure pollution.
Its crap like this that gives op-amps a bad rap in audio, when most of the time they actually do a pretty damn fine job. Most of th e time, the root causes are that people don't know how to use them properly.
Now, apologize and commit to a decent, qualified comment/assessment.
I need another glass of red wine.
I have no idea, tried it as IV-converter after DAC some years ago when the part was just released.
Clearly, you are a :bs: engineer
did you check it out with a scope?
was your layout ok
was the supply ok
Was it oscillating
Were you driving a capacitive load
etc
Why dis a product in this excessively subjective manner without qualifying in some detail your statement? This is straight mis-information and pure pollution.
Its crap like this that gives op-amps a bad rap in audio, when most of the time they actually do a pretty damn fine job. Most of th e time, the root causes are that people don't know how to use them properly.
Now, apologize and commit to a decent, qualified comment/assessment.
I need another glass of red wine.
QSerraTico_Tico said:
.... the power supply a "modified" Jung regulator. No not with the AD825 opamp
While some enjoy jumping on you about a remark that I can't find anything wrong with, let me join you.
I've only tried an AD825 in a servo cct and did not like the sound also although it measured quite well, what do you use in the Jung regulator?
john curl said:
Most of us agree that it is wise to build the most linear open-loop amplifier we can before applying negative feedback. I think this is true for both low-feedback and high-feedback advocates.
The controversy about amount of negative feedback in some ways goes to the core of the objective vs subjective alternative views of assessing amplifier quality. If someone says that to their ears low feedback just sounds better, this is a subjective preference that no-one can really dispute. It is no different than a person expressing a preference for the sound of tubes vs solid state, or vinyl vs SACD, or JFET input stages vs BJT input stages, or MOSFET vs BJT output stages. That person’s experience and preference is what it is.
On the other hand, the debate on the objective side is a bit more tractable. If someone says that they prefer low-feedback designs because they produce less distortion, that is something that can be checked by measurement.
I believe that Otala, for example, was perhaps more of an objectivist than some give him credit for. He asserted that high feedback and low open-loop gain caused greater TIM. He also asserted that high feedback was responsible for PIM and IIM. These were measurable distortions, and indeed he proposed measurement techniques for those distortions. This aspect is plainly in the objective realm. Properly carrying out those very same measurements ended up showing that his assertions were wrong; that they were unjustified generalizations.
There is no measurement that I know of that leads to the general conclusion that increased negative feedback, when properly applied, increases distortion of any type. As long as the comparison is legitimately apple-to-apples.
Consider a given open-loop amplifier that has 20 dB of NFB at 20 kHz with a conventional 6 dB/octave rolloff. That will give it a closed loop bandwidth of about 200 kHz. A “low feedback” design would typically load the output of the VAS in some way so that the open loop gain would not be allowed to increase at frequencies below 20 kHz, thus resulting in a “wide” open loop bandwidth of 20 kHz. A “high feedback” design would allow the VAS to operate with a very light load, allowing the open loop gain of the amplifier to continue to increase at lower frequencies. Does this “allowance” of the VAS gain to continually increase as frequency decreases cause increased measurable distortion? No.
Does it, for some as yet unknown reason, cause the amplifier to sound subjectively better or worse? I don’t think we know.
Under the lightly-loaded conditions that result in higher open-loop gain at low frequencies, the VAS in fact usually does not have to work as hard (e.g., swing current) and so produces less input-referred distortion. Under these conditions, the input stage also does not work as hard. The signal amplitude that it must handle at lower frequencies is much smaller due to the larger VAS gain that follows it. Indeed, the input stage in amplifiers with no negative feedback face a big challenge in that they must handle the full amplitude of the input signal all the time over all frequencies.
Interestingly, “low feedback” designs FORCE a designer to do a better job on the input stage (e.g. signal-handling ability, degeneration, etc) if they are to achieve the same open-loop linearity. This is because the input stage in a low-feedback design must handle larger signals. If you compare amplifiers with the same well-designed input stage, you will usually find that the amplifier that is permitted to have higher open loop gain at low frequencies will have equal or less distortion by any objective measure – including TIM or PIM.
If you build a “low feedback” amplifier with a passively loaded VAS, and then remove that loading, virtually all measurable distortions will usually go down. This, in spite of the fact that you have reduced its open loop bandwidth and increased its low-frequency negative feedback.
It is ironically often the case, by the way, that if one wants to have a “low feedback” amplifier, it is best to limit the rise of VAS gain at low frequencies by the use of, you guessed it, negative feedback, rather than passively loading it.
Finally, it is not inconsistent to have a high feedback amplifier with wide open-loop bandwidth. If one goes with a 2 MHz gain crossover, one can have 40 dB of NFB from 20 kHz on down to low frequencies. I would think that most people would call this a high feedback amplifier. Will this amplifier perform better than an amplifier with only 20 dB of NFB and a similar 20 kHz open-loop bandwidth. Probably, if it is executed properly (e.g., don’t sacrifice stability in order to achieve the higher gain crossover frequency).
Cheers,
Bob
AndrewT said:
Hi Andrew,
There is a lot of truth to what you are saying, but bear in mind that simply dc coupling the power amp will not gurantee the absence of large input-bias-current-induced offsets. Indeed, in most cases it will not, since the preamp likely has a coupling capacitor at its output, and the preamp will not guarantee a low-dc-resistance return to ground.
If you really know your preamp, and it has a well-behaved dc coupled output, then of course this is a good way to go. Indeed, dc coupled preamps themselves can have built-in dc servos to guarantee that their own output will have minimal offset without resort to an output coupling capacitor.
The biggest reason for the servo, in my opinion, is to get rid of that potentially bad capacitor in the feedback shunt leg. And the best way to achieve this without other compromizes, in my opinion, is to use a JFET input stage.
Cheers,
Bob
I do not disagree with your premiss that BJTs can be bettered by FETs at the input. Many designers state and show the improvements to be had.Bob Cordell said:
It is the mixed AC & DC coupling that I don't like.
If the pre-amp has a DC blocking cap at it's output, then the power amp must have a DC blocking cap on the NFB stage to preserve the DC balance at the inputs that we refer to and thus ensure minimal output offset.
Misusing a DC servo to allow bad topology to be incorporated is bad design in my book, even when input stage FETs are used to ameliorate the design error.
and that's why a DC coupled amp and pre-amp output, need a DC servo and DC detect with output isolation.
PMA said:
So? This is not a new or outstanding feature. There are lots of similar products from AD, TI, Linear, etc...
FYI the AD825 has a pretty low OLG for an opamp.
What's really interesting is the phase which indeed shows high stability potential (also in capacitive loads) and explains in part the good time response that you measured.
But then the noise performance is typical for a low cost, JFET input, opamp. You would have to go with AD745 for low noise applications.
- Status
- Not open for further replies.