Does an Op Amp Voltage Follower Operate in Class A?

Since my sources are all high level, I plan on building a preamp using an op amp in voltage follower mode. I know that when operating such a device with a gain of more than 1, the output will be in Class AB and a device such as a pull down resistor is required to force the output into class A. However, I was told that in the voltage follower configuration the output remains in Class A. I haven't been able to verify this so am asking the question here. If it's relevant, the device in question is an Analog Devices AD845 with the output taken from pin 6.
 
I've used the AD845 for audio and found it excellent. I wouldn't recommend biasing to class just as a matter of course or because it seems fashionable.
The quiescent current is quite high as these things go (10mA typ), so I'd expect it to be operating in class A anyway, with any reasonable signal level and load.

On another note...
It's unity gain stable, but the stability margins look too close for comfort to me. I think I'd want to throw away some loop gain in the interests of better stability and transient response.

YMMV, but I hear Mooly's other thread calling. ;)
 

Attachments

  • ad845 stab1.GIF
    ad845 stab1.GIF
    11.9 KB · Views: 201
  • ad845 srab2.GIF
    ad845 srab2.GIF
    8.4 KB · Views: 199
Thanks very much for all your responses. I understand the point that the bias current in the output stage is indepenent of the gain setting and determines the class of operation. If I use this op amp I'll have to figure out a way to bias its output stage for class A operation.

The way to do this is to connect a high A.C. impedance current source, or sink, to the op-amp's output pin. A JFET serves nicely in such a capacity. So, if you want to bias the op-amp output stage at 10mA, then you would program the current source to conduct 10mA D.C. Be careful, however, not to exceed either the current output or the thermal limits of the op-amp.
 
I wouldn't recommend biasing to class just as a matter of course or because it seems fashionable.
Me neither.
In fact, I have a gut feeling that *forcing* the Op Amp stage to feed or source an important (and "useless") constant current will affect its performance , in a bad way, more than any imaginary "improvement".

I understand it being made in, say, an LM358, but in that case it's justified because the output stage has *zero* bias, but in any "normal" one, it strokes me the bad way.

What load must it drive anyway?
Please state load impedance and expected signal voltage .
 
Hi,

Since the issue already appeared, here some thoughts:
I'd use a OPamp as gain=1 follower, if I.....
- wanted to test a circuit generally on function with lowest effort
- were in a hurry
- wanted/needed to build a no-brainer
- needed to reduce the budget by a couple of cents
- had tight space and power efficiency constraints
- were to produce marketing figures of no sonic meaning
- would eventually find one that sounds natural and authentic to my ears

I'd decide against OPamps, if I....
- cared for a natural, authentic sound
- had the time, budget and space a discrete soultion requires

A OPamp as follower is like driving a Ferrari at fullthrottle and braking at the same to gain the speed of a slow walker! Does it wonder it sounds rather like the distress signals of a bunch of tiny tortured souls instead of music? :smash:
A single pair of JFETs achieves the same functionality by following their very nature.
The difference beeing musical performance. :p

jauu
Calvin
 
Me neither.
In fact, I have a gut feeling that *forcing* the Op Amp stage to feed or source an important (and "useless") constant current will affect its performance , in a bad way, more than any imaginary "improvement".

I understand it being made in, say, an LM358, but in that case it's justified because the output stage has *zero* bias, but in any "normal" one, it strokes me the bad way.

What load must it drive anyway?
Please state load impedance and expected signal voltage .

Thanks for all your responses as I appreciate any help I can get. I'm certainly no expert and most of my ideas for circuits come from being a long-time reader of Audio Amateur Magazine and its later evolutions as Glass Audio and Audio Express. Based on ideas presented in various issues of the magazines, some time ago I had built a unity gain preamp using an AD845 without increased bias on the output. It sounded really good to me and I would still have it today if not for the fact that it got lost in moving to another house. It clearly sounded better than the commercial preamp it replaced although I recognize that being a DIY project undoubtedly contributed psychologically to its sonic appeal. I want to rebuild it and thought that this time, insuring class A output might improve the sound still further, although the original sounded so good to me I don't know if I would even be able to hear the difference.

In answer to your questions, my new build will use a 50K ohm stepped attenuator at the input, the AD845 connected as a voltage follower, a 49.9 ohm 1% metal film resistor at the output (to define the output impedence of the op amp per a statement by Walt Jung in an old AA article) followed by a polypropylene capacitor to keep any DC out of the power amp. This is pretty much the way I built the original except I used a 50k ohm Alps volume control.

The load impedence will be the 50K ohm input impedence of the power amp and the maximum signal voltage will be 2.0V from the CD player and somewhat less from my FM tuner. Since the preamp is mounted close to the power amp in the audio cabinet, the length of the connecting cables is only 1.5 meters.

Since there seems to be a difference of opinion as to the value of forcing the output into class A, I think the best course is to build it without the increased output bias and then maybe experiment with adding it at a later date to see if a noticeable improvement exists.
 
Hi,

Since the issue already appeared, here some thoughts:
I'd use a OPamp as gain=1 follower, if I.....
- wanted to test a circuit generally on function with lowest effort
- were in a hurry
- wanted/needed to build a no-brainer
- needed to reduce the budget by a couple of cents
- had tight space and power efficiency constraints
- were to produce marketing figures of no sonic meaning
- would eventually find one that sounds natural and authentic to my ears

I'd decide against OPamps, if I....
- cared for a natural, authentic sound
- had the time, budget and space a discrete soultion requires

A OPamp as follower is like driving a Ferrari at fullthrottle and braking at the same to gain the speed of a slow walker! Does it wonder it sounds rather like the distress signals of a bunch of tiny tortured souls instead of music? :smash:
A single pair of JFETs achieves the same functionality by following their very nature.
The difference beeing musical performance. :p

jauu
Calvin

Thank you for your informative, and entertaining, response to my question. I especially liked your equating the sound of an opamp to "the distress signals of a bunch of tiny tortured souls instead of music." Fortunately, that hasn't been my experience but I can readily understand your antipathy towards opamps if your ears have been assaulted with such a horrifying cry.

To respond to some of your points, reducing my budget by a few cents was hardly my priority as the power supply kit I purchased for my original build (Preamp Power Supply by Gary A. Galo: AA 4/90, p.47) cost me substantially more than the signal circuit components. I readily admit to this design being a no-brainer, however, because with the power supply a given, after reading Walt Jung's book on opamps building a voltage follower circuit was simplicity itself even with my limited knowledge. I will be using a similar power supply design in my new build.

In terms of finding an opamp that sounded natural and authentically musical, I socket my opamps allowing easy comparison among various types. After trying various devices, including the AD711, I finally settled on the AD845 as my ultimate choice for musicality. In connection with this, I would like to recount a marked difference in sound quality I noted between my homebrewed design and the commercial preamp it replaced. One of my favorite works is Beethoven's Violin Concerto in D, Op. 61. One of the versions I own is a Philips with Arthur Grumiaux as soloist with the Concertgebouw Orchestra Amsterdam. On both the vinyl disc and later the ADD CD version I bought when I switched to digital, about 5:23 minutes into the third movement, when the soloist plucks the strings, the sound came across as a dull thud when played on my commercial preamp. On my opamp design, the sound was clear, and without question, a "pluck." You can imagine my delight in this discovery. Perhaps not as delighted as Shockley, Bardeen and Brattain when they invented the transistor, but close.

As to why I chose a follower design, since my input sources were all at high level now, I felt that gain was superfluous and my attenuator would have a more extended useful range of operation yet still have the desirable impedance transformation of high in, low out. Why such a simple design? I wanted to get my feet wet and thought, perhaps mistakenly, that the simplest audio circuit is the best and you can't get much simpler than one opamp per channel. I even went so far as to house the toroidal transformer, the rectifier diodes and the input caps in a separate cabinet down near the power outlet and ran the power supply wires up to my preamp to keep all AC away from the signal circuits. I was pretty proud of my first effort and thoroughly enjoyed it while it lasted.

Notwithstanding the above, I recognize there may be better, more musical, designs than are possible with opamps, using FETs as you suggest, although I'm not sure that I could hear the difference. Although I'm a music lover I claim no golden ear status. However, you have piqued my interest and I may search out, and even try a FET design.

Again thanks for your comments. :)